Content uploaded by Jerry D. Davis
Author content
All content in this area was uploaded by Jerry D. Davis on Sep 20, 2017
Content may be subject to copyright.
SOUThEASTERN
GEOLOGY
V.
36,
No.
2,
August
1996,
p.
47-64
EVIDENCE
FOR
PLIO-PLEI$TOCENE
POLYGENETIC
DEVELOPMENT
OF
OKEFENOKEE
ARCUATE RIDGES
JERRY
DAVIS
Department
of
Geography
and
Human
Environmental
Studies
San
Francisco
State
University
San
Francisco,
CA
94132
ABSTRACT
Arcuate
sand ridges
are
prominent
features
of
the
Okefenokee
Swamp
(southeast
Georgia!
northeast
Florida).
They
are
generally convex
to
the
south
or
southeast,
and
are
up
to 1680
m
in
width,
10
km
in
length, and
7
m
in
height
above
the
subpeat
sandy
surface.
Vibracores
collected
from
a
sample
of
ridges
display bed-
forms
characteristic
of
estuarine,
deltaic,
or
flu-
vial
environments.
Grain
sizes
range from
pure
clays
to
fine gravels, though
most
samples
are
moderately
sorted
fine
sands.
Okefenokee
sand
ridge
grain
size
distributions range
in
paramet
nc
character
from
clusters
similar
to
nearby flu-
vial
environments
to
those
associated with
beaches
and
coastal
dunes,
with
most
samples
intermediate
in
nature.
Post-depositional
solu
tional etching
dominates grain
surface textures,
though water transport
and
wind
modification
is
indicated
in
some
samples.
Heavy
mineral
suites
indicate
a
Pleistocene
or
greater
age
with
closest
correspondence
to
terrace
upland
sedi
ments
to
the
west.
Taken together,
the
form
and
sedimentology
of
arcuate
sand
ridges within the
Okefenokee
Swamp
(southeast
Georgia/north
east
Florida)
indicates
a
polygenetic
origin,
with
older
ridges
formed within
a
deltaic
sys
tem
dissimilar
to
modern
environments
in
the
region,
and
somewhat younger
ridges formed
as
Carolina
bay
rims.
ARCUATE
RIDGE
LOCATIONS
AND
MORPHOLOGY
Ridges
of
predominantly
sandy
composi
tion
form
some
of
the more
prominent
islands
and
non-paludial
upland features
of
the
Okefenokee
swamp-marsh complex.
Seen
in
aerial
photography
and
Landsat
imagery, these
ridges clearly form arcuate
shapes with
many
intriguing
spatial
patterns
and
trends
(Figure
1).
Most
of
these arcuate
features are
convex
towards
the
south
or
southeast
and
most
have
high
length/width
ratios
(ranging
from
4.5
to
7
1
.0).
The
apparent
surface
morphology
of
the
ridges
is
in
part
determined
by
their
relationship
to
the
present
swamp surface, which
has
Un-
doubtedly
buried much
of
their
extent.
The
maximum
vertical relief,
2.5
m,
occurs
on
Floyd’s
Island, while
many ridges
are
at
or
be-
low
the
swamp
surface
and can
be
seen
in
imag
ery
only
through
the
effect
of
shallow
sands
on
vegetation
patterns. Some
of
the
larger
islands
appear
to
be
composed
of
multiple
parallel
ridg
es;
this
tendency
is
most
apparent
on Floyd’s
Is-
land (Figure
2)
and
was
confirmed
by a
leveling
survey.
The distribution
of
arcuate ridges
and
their
orientations
do
not
appear
random.
Groups
of
ridges
with
related
shapes
and
orientations
can
be
identified
in
several
areas.
The
group
con-
sisting
of
The
Pocket,
Billy’s
Island,
Honey
Is-
land, and
perhaps Bugaboo
Island
appears
to
be
segments
of
related
arcs
that intersect
at
a
point
near
the
present
swamp
outlet
of
the
Suwannee
River.
Other
more
distant
islands
form
arc
trends
that
also
appear
to
converge
at
this
outlet
(e.g. Floyd’s
and
Minnie’s
islands).
Groups
of
nested
or
perhaps
concentric
ar
cuate features
are
apparent
on
aerial
photo
graphs (Figure
3)
of
the
northeast
quadrant
of
the
swamp (between
Floyd’s
Island
and
Cow
house
Island).
None
of
these
features
are
identi
fied
on
maps
as
islands; they
appear
as
arcs
of
distinct
vegetation
and
oriented
lakes.
Another
group
of
similar
arcs
includes
Floyd’s
Island
47
JERRY
DAVIS
and arcs
of
distinct
vegetation
to
the
north
of
this
ridge.
Islands (including
Craven’s
Island and
Hickory
Hammock)
to
the
north
and
northwest
of
The
Pocket
group
appear
to
be
parts
of
ovals
or
ellipses, which
are also
represented
by
arcu
ate
extensions
of
the swamp
margin
into
the
northwest terrace
uplands
(Figure
4).
Another
ellipse
can
be
traced
in
the
center
of
the
swamp
(Figure
5)
to
the
north
of
Bugaboo
Island
as
an
arc
of
aquatic
and
shrub/prairie
vegetation
(based
on
a
map
by
McCafftey
and
Hamilton,
1980).
It appears
to
be
bounded
on
the
south
by
a
small
arcuate
sand
ridge
extending from
the
northeast
corner
of
Honey
Island
and
probably
continues eastward
as
a
submerged ridge.
Fol
lowed
to
the
west,
the
arc
appears
to
truncate
Honey Island. The trace
continues
through
Bil
Figure
1.
Landsat
Multispectral
Scanner
(MSS)
image.
Source:
North
American
Landscape
Characterization
program,
Mission
to
Planet
Earth,
National Aeronautics
and
Space Administration,
1980’s
image.
48
OKEFENOKEE
ARCUATE RIDGES
ly’s
Island
and
can
be seen
here
as
a
string
of
trace
is
less
distinct
near
Floyd’s
Island,
but
can
sink
ponds. To the north
of
Billy’s
Island,
the
be
picked
up
again
to
the south
where
it
forms
arc
recurves
to
the
northeast
and
is
apparent
in
the
western
margin
of
Chase
Prairie.
This
el
the
forced orientation
of
Minnie’s
Lake.
The lipse
or
oval
is
such
an
intriguing feature
in
im
Figure
2.
Northeast
Margin
of
Floyd’s
Island,
showing subparallel submerged
ridges.
19
JERRY
DAVIS
agery
of
the
swamp
that
a
name
is
proposed:
“The
Okefenokee
0.”
INTERPRETATIONS
AND
OTHER
POSSIBLE
ANALOGS
Interpretation
of
these
ridges
must
be
made
in
the
context
of
the
regional
geologic history
and
interpretations
of
related
Okefenokee
fea
tures
such
as
Trail
Ridge.
Establishing
a
time
frame
is
the
most difficult
part of
the
interpreta
tion.
The
general lack
of
fossils
and
other
date
able
materials
in
these deposits, probably
the
result
of
acid
weathering
conditions,
makes
es
tablishment
of
ages
difficult. Fossil discoveries
have limited
the
maximum
age
of
Trail Ridge,
which forms
the
eastern
margin
of
the swamp,
to
the late
Pliocene
(Pirkie
and Czel,
1983),
with
an
early-
to
mid-Pleistocene
interglacial
age
most
likely.
Cohen
and
others (1984)
felt
that
the
smaller,
arcuate
ridges
in
the
swamp
formed
as
“shoreline
features” before the
devel
opment
of
Trail
Ridge1,
but
other
interpreta
tions
(e.g.
Wadsworth
and
others,
1984)
would
suggest
a
younger
age;
no
fossils
have been
re
covered
to
help sort this
out.
Cohen
and
others
(1984)
noted
that Chesser
Island
and
Waycross
Ridge
are
truncated
by
Trail
Ridge, which
1.
In
his
review
of
the
manuscript
for
this
article,
Art Cohen
noted
“Arcuate
ridges
also
appear
to
me
to
be
typical
of
a
‘mac
rotidal
shoreline’
(as
per Hayes
1979);
whereas,
Trail Ridge
is
‘microtidal.’
This
perhaps
suggests
some
major change
in
gra
dient
of
shoreline
between time
of
formation
of
the
arcuate ridges
and
time
of
formation
of
Trail Ridge.”
Figure
3.
Region
of
submerged concentric
ridges
northeast
of
Floyd’s
Island.
50
OKEFENOKEE
ARCUATE
RIDGES
clearly
places
these
features
(but not
necessarily
all
arcuate ridges)
as
older.
Though most
studies have
not
focused
on
the
arcuate ridges
themselves,
interpretations
have
most
often
described
them
as
bars or
dunes
formed
in
the
coastal
or
near-coastal
environ
ment.
Pirkle
(1972,
p.
60)
described
the
ridges
as
bar-like
features
“built
along
a
prograding
shore line
of
a
regressing
sea.”
Cohen
(1973)
assigned
the
development
of
the bars
to
marine
current
action.
Fluvial and/or
deltaic
bars
were
implied
by
Brooks
(1966)
in
his
proposal that
a
proto-Su
wanee River flowed
to
the
Atlantic
(the
Su
wanee presently
flows
to
the
Gulf
of
Mexico
from
the
Swamp) leaving
downstream sediment
deposits
in
the
vicinity
of
the
present
Okefeno
kee.
Deltaic
and
estuarmne
sediment
systems
ap
pear
to
hold the
most
promise for
producing
arcuate
ridges
of
similar
plan
morphology;
in
particular, deltaic
subaqueous
levees
described
by
Wright
(1977)
as
resulting
from
plane-jet
diffusion
are
similar
in
morphology
and
sedi
ment
characteristics.
Wind
is
another possibility.
Smedley
(196$) referred
to the
Okefenokee
ridges
as
“Pleistocene
sand dunes and
bars.” Parrish and
Rykiel
(1979),
in
their
review
of
swamp
ori
gins,
advanced
awind
hypothesis.
Wadsworth
and others (1984), who
studied
the
Okefenokee
ridges
in
most
detail,
noted
their morphological
similarity
to
raised rims
associated
with
Caroli
na
bays,
including
their tendency
to
“crosscut,”
and
proposed that
similar
lake
basins
may
have
existed
in
the
Okefenokee.
Unfortunately
the
question
of
the
origin
of
Carolina
bays
them
selves
has
engendered considerable
controver
sy,
though
the
most plausible cause
of
raised
rims
on
the
southeastern margins
of
these
fea
tures involves
seasonal
storm-driven
winds
dur
ing
full-glacial
low
lake stands
(Thom,
1967,
1970;
Kaczorowski,
1977).
Figure
4.
Narrow
arcuate
ridges
forming subelliptic
traces
in
northwest region
of
Okefenokee
Swamp.
Com
posed
of
Hickory Hammock,
Craven’s
Island,
and
Craven’s
Hammock.
51
JERRY
DAVIS
Other
analogs
might
be
considered,
includ
ing
various aeolian dune
systems.
The “pan
margin
dunes”
of
South
Africa
(Lancaster,
1978),
and
the
similar
lunettes
described
in
Texas,
North
Africa,
and
Australia
(Hills,
1940;
Campbell,
196$)
are,
like
Carolina
bay
ridges,
associated
with
relict
lake
margins. Parabolic
dunes
have
been
described
developed
on
the
northeast
flanks
of
NW-SE
trending rivers
on
the
coastal
plain
of
Georgia
(Asmussen,
1971;
Thames,
1982)
and
North
Carolina
(Miller,
1979;
Daniels
and
others,
1970),
and
appear
to
result
from blowout
of
sandy
channel sediments
during
dry
periods.
ANALYSIS
OF
THE
RIDGES
A
viable
interpretation
of
puzzling
features
such
as
the
Okefenokee arcuate ridges
cannot
rely upon
a
single layer
of
information.
In
this
study,
morphological
characteristics (described
above), bedding
characteristics, sediment
size
parameters,
grain
surface
texture
and heavy
mineral assemblages
were
used
in
conjunction
to
describe
and
interpret
the ridges and samples
Figure
5.
The
Okefenokee
‘0’,
a
subelliptic
trace
in
the
center
of
the
Okefenokee
Swamp.
52
OKEFENOKEE
ARCUATE
RIDGES
from
cores.
Sediments
were
analyzed from
21
vibra-cores
of
the
Okefenokee
arcuate ridges,
and
10
other
comparative
sites
from
various
up
land
and
ridge
sites
around
the
swamp,
includ
ing
Trail
Ridge;
two
Georgia coastal
plain
river
margin
sand
ridges,
on
the
Altamaha
and
Ohoopee
rivers;
and
a
Carolina
bay,
Alligator
Bay
in
South
Carolina. Information
on
other
an
alogs
was
derived
from
the
literature.
Core
Descriptions
and
Sediment
Size
Characteristics
ments,
with
significant bioturbation
in
many
sections.
Significant clay-rich
layers
halted
penetration
of
the
vibracoring
system
at
depths
ranging
from
1
.72
to
4.82
m;
these layers
typi
cally
exhibited
mud cracks
indicating
periods
of
subaerial
dessication. Cores from The
Pocket
exhibited
grass-like root
traces.
Such
character-
istics
are
typical
of
deltaic/estuarine
sedimenta
tion
patterns
of
fluctuating
channels, with
alternating
deposition
of
sands, muds, and
or-
ganic matter (Reineck
and
Singh,
1980;
Cole-
man,
1981;
Howard
and
Scott,
1983;
Frey
and
Howard,
1986).
In
near-surface
samples
from
The Pocket,
mean
grain
size
decreased
from
Transect
B
(moment
mean
=
2.52
)
to
Transect
A
(2.77
),
and
from seaward
to
landward margins
(Transect
A:
2.63 to 2.95
;
Transect
B:
2.38
to
2.47
).
Both
patterns
are
also typical
of
sub
aqueous
levees
described
by
Wright
(1977).
Sequences
of
relatively coarse
sands and
fine
gravels
(up to
4
mm),
suggestive
of
fluvial
deposition,
were
noted
near
the
base
of
cores
from
Pine
Island
(mean
of
sample
8
from
core
P11
=
0.91)
and the
terrace
uplands
just
to
the
west
of
the
swamp
(mean
of
sample
8
from
core
Ui
=
0.13).
Fining-upward
sequences,
typical
of
migrating
point
bars
(Reineck
and
Singh,
1980)
and
some
progradational
sequences
Transects
from
The
Pocket
and
Floyd’s
Is
land
provided
the
most
detailed information
on
sediment
patterns
along
and
across
the
ridges
(Figures
6
&
7).
In
each
of
the
16
cores
arranged
in
4
cross-ridge
transects, visible
inspection
of
cut
cores
revealed
alternating
layers
of
relative
ly
clean fine-to-medium
sands,
less
well-sorted
sands
with
finer
sediments,
and
clay-rich
sedi
Figure
7.
Transect
and
core locations
on
Floyd’s
Island.
Figure
6.
Transect
and core
locations
on
The
Pocket.
53
JERRY
DAVIS
(Klein,
1970)
were observed
in
the
Pine Island
core
(3.0 to
3.5
m
depth)
and
in
one
core
from
The
Pocket
(P2
above
2
mdepth).
Coarsening-
upward sequences,
typical
of
delta
front
progra
dation (Tucker,
1981),
were
apparent
in
the
Pine
Island
core
and the
terrace uplands
core.
These
observations
are
consistent
with
Brooks’s
(1966)
hypothesized
outlet
of
a
proto-Suwan
nee
River
in
the
Okefenokee
area.
Nearly
all
samples
exhibited
a
positive
skew;
exceptions
were
primarily
in
deeper
sam-
ples
of
cores
from
Floyd’s
and
Chesser
islands,
which also
exhibited better
sorting than
most
other samples
from
the ridges.
In
general,
ridge
samples
ranged
from poorly-sorted
to
well-sort-
ed
fine
sands, with
somewhat
better
sorting
overall
on
Floyd’s
Island
due
largely
to
lower
mud
percentages
--
8.2%
in
Floyd’s
Island
sam-
ples
vs.
21
.3%
in The
Pocket
samples
(2.5%
vs.
9.7%
in
near-surface
samples).
Cluster
Analysis
Based Upon
Size
Parameters
Sedimentological
classification
of
samples
based upon
size
parameters
is
a
problematic
procedure. Errors
due
to
inconsistent sediment
source
generally
indicate
the need
to
compare
samples
only from
similar
or at
least
regional
environments,
and
many features
contain
sedi
ments that
have likely
been
transported
over
short distances
and
thus are
little
modified
by
the
depositional
system being
studied.
Individ
ual
samples
were
thus
compared
only
with
sam
ples
from
depositional
environments close
to
the Okefenokee. Data
derived from
the
litera
ture
were
from
two
suitable
sources,
both
de
scribing features
on
the
Georgia coast
and
coastal
plain:
Hails
and
Hoyt
(1969),
which
in
cluded
samples from
barrier
islands,
fluviatile
dune
sands,
and
lagoon-salt
marsh sediments;
and
Thames
(1982),
which
included
samples
from river-margin
sand ridges,
river
channel,
coastal
dune,
and
beach
sediments. Since
other
depositional
environments
were
also
consid
ered
but
were not
represented
in
the region,
cluster
analysis
was
selected
a
the
most
suit
able
method
of
classification; cluster analysis
delineates
sample groups without
the
need
for
known
model categories.
Three
variables (moment
mean,
moment
standard
deviation,
moment
skew)
grouped
the
samples
clearly into seven
clusters,
four
of
which
represented
known
environments
(groups
2,
4,
5
and
6),
and
three
represented
un
known environments
(groups
1, 3
and
7).
Each
group
included some samples
from various
depths
in
Okefenokee arcuate
ridge
cores:
Group
1
:
somewhat poorly sorted
(1.07
c1),
positive
skew
(1.38
),
fine
sands
(2.63
p).
Includes
many samples
from
Okefenokee
ridges,
especially
from
The
Pocket
cores.
Group
2:
river-margin
sand
ridges
--
moderately
sorted (0.79
),
positive
skew
(0.47
p),
medium
sands
(1.74
p
).
Represented
only
in
a
few
deeper
samples
from
Pine Island,
Floyd’s
Is-
land,
and
Chesser
Island.
Group
3:
moderately
sorted
(0.79
),
very
positive
skew
(1
.58
),
fine
sand
(2.38
).
Includes
several
sam-
ples
from
Floyd’s,
Pine
and
Chesser
islands, especially
near
the surface;
very
few
from The Pocket;
perhaps
wind-modified
group-
1
samples.
Group
4:
river channel
--
moderately
sorted
(0.79
p),
negatively skewed
(-
0.31
0),
coarse
sands
(0.65
).
Only
represented
in
deepest Pine Island
sample
and near
base
of
core
from
the
nearby
terrace uplands.
These
samples
include
fine gravels.
Group
5:
beach
--
well
sorted
(0.48
0),
negatively
skewed
(-0.43
),
fine
sands
(2.68
).
Deeper
samples
from
the
seaward edge
of
Floyd’s
and
Chesser
islands.
Group
6:
coastal
dune
--
well
sorted
(0.43
),
positive
skew
(0.61
p),
fine
sands
(2.65
).
Samples at
various
depths
and
locations
on
Floyd’s
Is
land; also
in
deeper
samples
from
Chesser
Island
and
The
Pocket
Transect
A.
Group
7:
silty
to
clayey,
poorly sorted
(1.54
),
positive
skew (0.74
),
fine
sand
(2.79
).
These
are
from
the
thin
clay-rich
zones
at
various
depths
of
most
cores.
54
OKEFENOKEE
ARCUATE
RIDGES
Grain
Surface
Texture
A
limited
set
of
samples
from
Okefenokee
and
other
sites
was
studied
using
a
Philips
505
Scanning Electron Microscope
(SEM)
at
the
Center
for Advanced
Ultrastructural
Research,
University
of
Georgia.
Table
1
lists
observed
Table
1:
Grain
surface textural characteristics
observed
with
SEM.
d
th
observed grain
Location
core
surface
characteristics*
Floyd’s
L
F3
2.0 MR, ABV,
WR,
SE
Pocket
P1
2.0 SE
P1
2.7 SE,
MR?
Chesser
I
Cl
2.8
ALV,
SE,
UP
Cl
2.0
ALV,SE
TrailR.
TRY
1.5
SE
Ohoopee
D.
01
2.0
ALV,
UP
*
Abbreviations
for grain surface characteristics:
MR:
meandering
ridges
—
WE: well-rounded
grains
—
SE:
solutional
etching
—
UP:
upturned
plates
—
ALV:
aligned
V-shaped
indentations
—
ABV:
abraded
V
shaped
indentations
grain surface patterns
using
terminology
from
Krinsley
and
Margolis
(1971) and
Manker
and
Ponder
(1979).
The
most
prevalent
characteris
tic
noted
in
all
samples
was
chemical
etching
(Figure
8A-B).
This
appears
to
have
obscured
pre-existing
surface
textures, making more
de
tailed
analysis impossible. V-shaped
indenta
tions
(a
characteristic
of
either
beach
or
river
sediments)
were observed
on
grains
from
Chesser
Island
(Figure
8C)
and
Floyd’s
Island,
as
well
as
from
the
Ohoopee
Dunes. One
sam
ple
from
Floyd’s
Island exhibited characteristics
associated
with
wind
transport
--
meandering
ridges
and
well-rounded
grains (Figure
8D).
Extensive
solutional
modification
of
Okefenokee
sands
has
obliterated
many
surface
patterns, making
electron
microscopy
tech
niques
inadequate
for
detailed
interpretations,
but
some
general
statements
can
be
made.
Dep
osition
of
sediment
from
an
aqueous
medium
is
a
part
of
the
history
of
all
of
the
ridges.
Long-
term
exposure
to
chemically
aggressive
water
is
also
evident
from
the
extensive
etching
of
sand
grains
in all
samples.
The
limited etching
on
Chesser
Island
grains may
indicate
a
shorter
ex
posure
to
acidic
waters.
Finally,
evidence
exists
for
sediments
from
Floyd’s
Island having
been
modified
and
abraded
through
wind
transport.
Heavy
Mineral
Analysis
Heavy
minerals separated
from
samples
collected
between depths
of
1-2
meters
in
each
core
suggest
a
terrace
uplands
sediment source
for
the
main group
of
arcuate
ridges
in
the
swamp.
Samples
from Transect
B
of
The
Pock-
et were
most
similar
in
both
mineral
suite
and
relative
concentrations
to
terrace uplands
sam-
ples,
both having
15-20%
sillimanite
and
less
than
5%
zircon.
Samples from
Transect
A
of
The
Pocket
were,
like
samples
from
Floyd’s
and
Pine
islands,
somewhat
more
enriched
in
silli
manite (generally
over
20%)
but
comparable
in
zircon.
Trail
Ridge
appears
to
be
less
closely
relat
ed
to
the
Okefenokee
ridges.
Little
if
any
epi
dote
was
noted from
any
other
Okefenokee
sample,
as
predicted
by
Pirkie
(1975); and
in
this
regard,
the
ridges within the
Okefenokee
are
similar
to
Trail
Ridge.
However,
Trail
Ridge
samples
differed
significantly
in
silli
manite
(1.6%) and zircon
(22.8%)
concentra
tions.
Chesser
Island sediments were
somewhat
closer
to
Trail Ridge
in
these
concentrations
(sillimanite
<
10%;
zircon
>
5%),
but
there
is
little
evidence
to
assign them
a
common
source.
CONSIDERATION OF
ANALOGS
Several statements
can
be
made
concerning
possible
analogs
for
the
Okefenokee arcuate
ridges
in
light
of
the
results
described
above:
1.
Coastal
beach
and
dune
environments
may have
pre-dated ridge construction.
Coastal
affinities
(beaches
and
coastal
dunes)
are
most
common
in
deeper
samples
especially
on the
seaward
sides
of
Chesser
and
Floyd’s
islands.
2.
Some
sedimentological
similarities exist
55
JERRY
DAVIS
Eq
oo
Cl)
C-
Cl)
L)
c,5’CD
C
C.)
C) C)
.E
C)
C
C)
•
Cl)
Cl)
C)CI)
C)’—,
C4
C)
o
c
C.)
•C)
E
Co
Cl)
coo
00)0)
0
00
•—c
•
.E
C)
••-C)
between upper
samples
of
The
Pocket
and
those
of
river-margin
sand
ridges,
but
differences
suggest
a
unique
depositional
environment.
3.
River
deposition
is
clearly
a
part
of
the
history
of
the
area where
Pine Island
was
built,
and was also
significant
in
the
terrace
uplands.
The western
upland watersheds presently
drained
by
small
creeks that
feed
the
Okefeno
kee
do
not
include
sources
of
gravels
similar
to
those
found
in
deeper
samples
from
Pine Island;
a
much
larger
fluvial
system
with
a
distant
source
region
is
indicated
(as
per
Brooks
1966).
4.
An
aqueous
medium
transporting
sedi
ments from
Transect
B
to
Transect
A
of
The
Pocket
is
indicated
by a
B-to-A
decrease
in
mean
grain
size.
5.
The source
of
sediments
in
The Pocket
likely
was
from
the west,
as
indicated
by
heavy
mineral
similarities.
6.
A
prograding
deltaic
system
is
a
likely
interpretation
for
most
of
the
arcuate
ridges.
Coarsening-upward
sequences
were noted
in
56
OKEFENOKEE
ARCUATE RIDGES
cores
of
the
terrace
uplands,
Pine
Island
and
Floyd’s
Island,
and
fluctuating
sandy and
finer
layers were
noted
in all
cores. The
plan
mor
phologies
of
The
Pocket,
Billy’s
Island,
Honey
Island,
and
New
Island
are
similar
to
the
spreading
subaqueous
levees
described
by
Wright
(1977)
and
Coleman
(1981)
as
the
result
of
plane-jet
diffusion
of
an
otherwise
inertial-
dominated
river
delta
as
it
encountered
friction
from
shoaling
at
its
mouth.
7.
Tidal marsh
deposition
may have been
significant
in
some areas,
especially
The Pock-
et,
as
indicated
by
mud
cracks
and
root
traces
in
split
cores.
8.
A
progression
from coastal
(beach,
coastal
dune)
sediments
to
estuarine
(mud
lay-
ers,
mud-clasts, root
traces)
or
deltaic
(fluctuat
ing
channel/marsh)
sediments
is
seen
in
some
cores from
all
ridges.
Cores
that
do
not
exhibit
this
progression
(e.g.
Transect
B
of
The
Pocket)
may have
been
too
shallow
to
provide
a
com
plete data
set.
Progradation
is
again indicated.
9.
A
history
of
wind
deposition
is
suggest-
ed
not
only by
deeper seaward
samples
from
Chesser
and
Floyd’s
islands, but
also
in
surface
samples from
Floyd’s
and
Pine
islands,
based
upon
sediment
size
characteristics
and
grain
surface
texture. One
interpretation
of
these
ten
dencies that
agrees
with
observed
morphologies
is
that
all
of
the
ridges may have been
deposited
through
the same
mechanism
(probably
delta
ic).
However,
The Pocket,
Billy’s
Island,
Honey
Island,
and
New
Island
are
the
most recent
for
mations;
whereas,
the
other
islands
formed
ear
lier
and
have
since
been modified
by
wind.
10.
Oval
traces
similar
to
Carolina
bays
in
clude arcuate
ridge
segments. Since
these
ap
pear
to
truncate other
Okefenokee
ridges, they
likely
represent
a
system
that postdates other
ridge-forming
processes
within
the
swamp.
The
wind
modification
of
sediments
described
above
may
be
associated
with
Carolina
bay
ridge
development
in
pre-existing
sediments.
THE ORIGIN
OF
THE
ARCUATE
SAND
RIDGES:
A
PROPOSED SCENARIO
No
single
interpretation
ofthe
arcuate
ridg
es
can
explain
all
of
the observations.
An
early
assumption
in
this
study
--
that
such
a
single
in-
terpretation
must
exist
--
has
been abandoned
in
favor
of
a
polygenetic model
invoking
deltaic
and
Carolina
bay
systems.
The modern
land-
scape (Figure
9)
is
a
palimpsest
resulting
from
the
superimposed impacts
of
these
and
other
geomorphic
systems.
The
earliest depositional
systems
observed
in
the cores were
coastal beaches
and dunes.
The depth
of
these samples, coupled
with
the
low
areal
sampling
density
of
ridge
cores,
al
lows
little
interpretation
of
these
systems.
If
a
statement
can
be
made
about
these
samples,
it
is
that
shore-zone
depositional
systems
pre-dated
overlying environments
in
this
area.
The first major geomorphic system
signifi
cant
to
arcuate
ridge
morphology
in
the
area
is
that
of
a
prograding
delta
with
distributary
channels bounded
by
subaerial
and
subaqueous
levees.
The central
major
arcuate ridges,
includ
ing
The
Pocket,
and
Minnie’s,
Billy’s
and
Hon
ey
islands,
can
be
interpreted
as
remnants
of
levees
from
a
major
deltaic
channel. One
possi
ble
scenario
to
explain
the spatial
ridge
pattern
is
illustrated
in
figures
10-12,
which
depict
both
a
lateral (southward)
shift
and a
progradational
trend
typical
of
many deltas.
We
can
only
speculate
about
the
source
re
gion
of
the
river forming
the
delta
here,
though
subsequent river captures
are
probably
signifi
cant.
Both
the
Ocmulgee
and
Alapaha
systems
have
morphologies
suggestive
of
capture,
and
the
Suwannee
River
may
have changed
direc
tions,
as
suggested
by
Brooks
(1966).
The
Oc
mulgee,
now
part
of
the
Altamaha
basin,
presently
drains
a
portion
(9300
km2)
of
the
Piedmont
and
Coastal Plain
provinces
of
Geor
gia.
The
Alapaha
is
presently
an
insignificant
part
(35J(j
km2)
of
the
Suwannee
basin,
but
may
have
been
much
larger
in
the
past:
much
of
its
catchment
(now
limited
to
the
coastal
plain
Tifton Uplands)
has
apparently
been
captured
57
JERRY
DAVIS
by
streams
to
the west
of
the
Peiham
escarp
ment,
now
draining
via
the
Flint
and
Chatta
hoochee
Rivers
to
the
Appalachicola
River.
The
size
of
this
former
drainage
basin
is
difficult
to
estimate; though
approximately
50,000 km2
would have
been
drained
if
the basin
included
portions
of
the
Piedmont
and
Appalachian
Highlands
Provinces presently drained
by
the
Chattahoochee
and
Flint
Rivers. The
Alapaha
in
turn
was
likely captured
by
the
Suwannee
af
ter
its
flow
was
reversed
to
the
Gulf,
as
Upper
Eocene
limestones
(Ocala Group)
were
ex
posed
at
the
crest
of
the
Ocala
Arch.
Either
the
Ocmulgee
or
the (expanded former)
Alapaha
system could
have
provided
the
coarser
sedi
ments found
in
cores
of
Pine
Island
and
The
Pocket.
Isostatic
uplift resulting
from
karstic
re
moval
of
underlying limestones
may
have
been
a
factor
in
both
Alapaha
captures.
Such
uplift
was
interpreted
by
Opdyke
and
others (1984)
for
northern
Florida.
Between
the
Alapaha
Riv
er
and
the
Okefenokee drainage
basin
is a
major
area
of
Carolina
bays
believed
to
have formed
on
sands
underlain
by
carbonates. Seepage
from
these
sinks
into the
Principal Artesian
Aquifer,
which
flows
beneath
the swamp,
fur-
ther
reduced
the
sources
of
surface
runoff
into
the
swamp.
Only
with
a
diversion
of
major river
inputs
could
Trail Ridge have been
constructed
with
no
significant
breaks from
the
St.
Marys
River
Figure
9.
Major drainage
networks
in
the vicinity
of
the
Okefenokee
Swamp.
58
OKEFENOKEE
ARCUATE
RIDGES
to
the
Satilla
River.
The
Suwannee
River’s
out
let may
have
already been
established
to
the
southwest.
Alternatively,
the
remaining
inputs
to
the
swamp basin
may
have
reached
the
sea
through
one
or more
outlets:
the
present
outlet
of
the
St.
Marys
River,
or
a
low
point
in
Trail
Ridge
just
south
of
Chesser
Island.
One
factor
that
may
have
been significant
(and
may
contin
ue
to
be
so)
is
isostatic
subsidence
due
to
deltaic
sediment
accumulation.
This
may
be
the
cause
of
the
present
lower elevations
on
the
western
margins
of
the
swamp, where
delta sediments
would
have been
thickest.
The
developments
presented
to
this
point
are
thus seen
as
pre-dating
the
development
of
Trail Ridge, which
is
assumed
to
have
been
constructed during
an
interglacial
transgression
of
the early-
to
mid-Pleistocene.
No
precise
age
has
yet
been
established
for
this
major
barrier
ridge,
though
we
know
that several
major
trans
gression-regression
cycles have
occurred
since
this
time
(along
with
the
general
progradational
trend
of
the
southeastern coastal plain)
and that
fossil evidence
places
Trail
Ridge
as
no
older
than
the
late
Pliocene
and
most likely
within
the
Pleistocene
(Pirkle and Czel,
1983).
The
amount
of
Trail
Ridge structural
deformation
(an
admittedly unreliable dating
method)
would
indicate
an
age
in
the
vicinity
of
0.4
-
1.0
mil
lion years.
What
has
happened
in
the
Okefenokee
ba
sin
since
this
time?
We
know
that
late
Pleis
Figure
10.
Deltaic
construction
of
Floyd’s
and
Minnie’s
Islands.
59
JERRY
DAVIS
figure
11.
Deltaic construction
of
Pine Island
and
Mixon’s
Hammock.
tocene interglacial
conditions
in
this
region
were
warmer and
wetter
than
the
glacials;
this
is
based
upon
palynological
studies
of
the
south
eastern coastal plain
(Watts,
1971;
Watts
and
Stuiver,
1980).
By
this
time,
the
Suwannee
had
captured the
Okefenokee
basin’s
drainage,
as
well
as
that
of
Suwannoochee
Creek.
As
they
do today,
fluviatile
swamp
developments
ex
panded
up
river
valleys where
stream
gradients
were
limited.
In
areas
of
structural subsidence,
such
as
the
Okefenokee
basin,
gradients
in
out
ward-draining
streams were
particularly
re
sthcted.
Lakes developed preferentially
in
depressions
located
at
groundwater seepage
points.
The region
of
sinks
to
the
west
of
the
swamp developed
in
one
of
these
areas,
as
dis
cussed
above. Ponds
within
the
swamp may
also
have
developed
at
seepage
points,
prefer-
entially developed
(1) to
the
northwest
of
The
Pocket,
(2)
between
Billy’s
and
Floyd’s
Island
(the
Okefenokee
0),
and
(3)
to
the
northwest
of
Floyd’s
Island.
Pre-existing deltaic
arcs
certain
ly
affected
the
drainage
patterns,
and
thus
par
tially
controlled
lake
development.
Drier
and
cooler
glacial
conditions
are
shown
in
Figure
13.
Regression
of
sea
level,
an
important
regional
climatic
factor, had
in
creased
stream gradients
and thus
reduced
swamp formations, probably aided
by
peat
fires.
Wet-season
(summer) lake
expansions
may have
been
oriented
into
typical
Carolina
bay
forms,
the
result
of
current
action;
this pro-
Deltaic
construction
of
Pine
Island,
Mixon’s
Hammock,
etc.
levees
levees
&
bars
0
15
km
60
OKEFENOKEE
ARCUATE RIDGES
figure
12.
Deltaic
construction
of
The
Pocket,
Billy’s
Island,
Honey
Island,
Chesser
Island,
and
other
ridges.
cess was
demonstrated
by
Kaczorowski
(1977).
It
was
at this
time
that pre-existing
ridges
might
have been susceptible
to
erosion
by
lake
margin
current
action;
the truncation
of
Honey Island,
forming
part
of
the
Okefenokee
0,
is
an
exam
ple.
During
the
dry
winter
season,
fire-exposed
sands
were susceptible
to
wind
modification.
Arcuate
ridges
were
constructed
as
parabolic
dunes
on the
southeastern
and
eastern
margins
of
shrinking
Carolina
bay lakes, from
strong
northwesterly
to
westerly
winds.
Craven’s
Is
land,
Hickory
Hammock,
parts
of
Minnie’s
Is
land,
and
many other subswamp ridges
(especially
to
the
north
and
northeast
of
Floyd’s
Island)
may
represent
receding
lake-margin
dunes dating
from
one
of
these
periods.
Floyd’s
Island may
have
been
heightened
at
this
time,
and
the
multiple
ridges
on
its
surface
may
have
migrated
as
dunes across
its
crest;
similar
mod-
ifications
may
have affected Pine Island
and
Mixon’s
Hammock.
We are
presently
in
an
interglacial
cycle,
the
Holocene,
with
a
return
to
wetter
conditions
initiated
with
the
recession
of
Wisconsinan
gla
ciers
prior
to 10,000
years
B.P.
The
Okefenokee
has
again
expanded
as
amultiple
riverine
swamp
of
the
Suwannee
River
and
Cypress
Creek,
as
outlined
by
Parrish
and
Rykiel
(1979).
The headward development
of
the
St.
Marys River
has
renewed
a
limited
amount
of
drainage
to
the
Atlantic
through a
gap in
Trail
Ridge.
Wind
is
no
longer
an
important factor
Deltaic
formation
of
The
Pocket,
<
Billy
s,
Honey,
and
Chesser
I
—
/
/_
levees
marsh
4/
J
Honey
I.
:baymouth
smaller
stream
.:-‘
iP)
)
bar?
or
alternate
.
Strange
(Chesser
I)
4
jack
I
-‘MttchII
I.
4.
15
/
0
km
61
JERRY
DAVIS
here
as
vegetation
has
become
more
extensive
and
peat
has
accumulated.
Fires
still
play
a
part
in
retaining
the
marsh “prairie” components
of
the
Okefenokee
complex
(Izlar,
1984;
Hamil
ton,
1984),
but
not
until climates
become
much
drier
can
these significantly
reduce
the
peat
cover.
The
arcuate ridges
are
well
preserved
in
this
environment
of
limited
erosion.
CONCLUSION
This
study
has
demonstrated
that
the
arcu
ate
sand
ridges
of
the
Okefenokee
Swamp are
a
result
of
multiple
causes
with
at
least
two
sig
nificant developments
occurring
within
the
Quaternary
Period.
Sedimentological,
morpho
logical,
and
spatial
evidence
indicate
that
a
flu-
vial-dominant
deltaic
environment
may have
been
responsible
for
many
of
the
sand
ridges,
especially
those
in
the
vicinity
of
The
Pocket
(including
Billy’s
Island
and
Honey
Island).
The
river
system
responsible
for this
delta
drained
large areas
to
the
west
of
the
swamp
in
the
Pliocene
and
continued
to
drain
into
the
ba
sin
until some
time
before
the
development
of
Trail
Ridge
in
the
early-
to
mid-Pleistocene.
This
is
in
agreement
with
Brooks (1966),
who
proposed
an
Atlantic-draining late-Pliocene
Su
wannee
River, with
a
“strath”
in
the
Okefeno
kee
basin.
Similar kinds
of
evidence support
the
late-Pleistocene
development
of
Carolina
bays,
Figure
13.
Development
of
large Carolina
bays in
the
Okefenokee
Swamp.
62
OKEFENOKEE
ARCUATE
RIDGES
responsible
for
additional
arcuate
ridges that
truncated earlier
deltaic deposits
or
reflect
re
ceding
lake shorelines
(Craven’s
Island,
Hicko
*
ry
Hammock,
and
many
submerged
ridges).
Wind
modification
of
pre-existing
ridges,
espe
cially
Floyd’s
and
Pine Islands,
occurred
as
part
of
a
glacial
dry
cycle
at
this
time.
Swamp
developments
and
arcuate ridge
modifications
are
seen
as
cyclic.
At
least
over
the
last
few
glacial
cycles,
the
wet
interglacials
have
been times
of
swamp expansion
and
the
dry
glacials times
of
peat
reduction
by
fires.
The arcuate
sand
ridges
are
only
significantly
modified
during
glacials,
when
the peat
cover
is
reduced
and
the sands
exposed
to
current
and
wind
erosion.
Despite
the
acidic
weathering
en-
vironment
of
swamp
conditions, effective
in
re
ducing
the
sediments
to
a
quartz-dominant
composition,
the ridges are
well-preserved
dur
ing
these
periods.
Only
after
the
St.
Marys
Riv
er
recaptures
the basin
to
the
Atlantic
will
these
curious
arcuate
sand
ridges
be removed,
and
no
one will
wonder
about
them
again.
ACKNOWLEDGEMENTS
To
George
Brook,
Albert
Parker
and
Philip
Suckling,
University
of
Georgia Geography
Department;
and
Ray
Freeman-Lynde
and
Rob
ert
Carver,
University
of
Georgia
Geology
De
partment,
for
consultation
on
the
original
research.
To
the many field assistants
I
talked
into
a
weekend
of
coring
in
the swamp.
To the
reviewers
of
this
article,
Art Cohen
and
Fre
drick
Rich,
for their helpful
suggestions. This
research
was
supported
by
National Science
Foundation
grants BSR-8
114823
and
BSR
8215587.
REFERENCES
CITED
Asmussen,
L.
E.,
1971,
Hydrologic effects
of
Quaternary
sediments
above
the
marine terraces
in
the
Georgia
coastal
plain:
Southeastern
Geology,
v.
12,
p.
189-201.
Brooks,
H.
K., 1966,
Geologic
history
of
the
Suwanee
River,
in
Olson,
N.K., ed.,
Geology
of
the
Miocene
and
Pliocene
series
in
the
north
florida-south
Georgia
area:
Atlantic Coastal
Plain
Geological Association,
7th
field
Conference,
Southeastern Geological
Society
12th
Field Conference,
Guidebook,
p.
37-45.
Campbell,
E.
M., 1968,
LuncHes
in
southern south
Austra
ha: Transactions, Royal Society
of
South
Australia,
v.
92,
p.
85-113.
Cohen,
A.
D., 1973,
Possible influences
ofsubpeat
topogra
phy
and
sediment type
upon
the
development
of
the
Okefenokee
swamp-marsh
complex
of
Georgia: South-
eastern
Geology,
v.
15, p.
141-51.
Cohen,
A.
D.,
M.
J.
Andrejko,
W.
Spackman,
and
D.
A.
Cor
vinus,
1984,
Peat
deposits
of
the Okefenokee
Swamp,
in
Cohen,
A.
D.,
D.
J.
Casagrande,
M.
J.
Andrejko,
and
G.
R.
Best,
Eds., The
Okefenokee
Swamp: Its
natural
history,
geology, and
geochemistry:
Los
Alamos,
NM,
Wetlands
Surveys,
p.
493-553.
Coleman,
3.
M,
1981,
Deltas:
Processes
of
Deposition
and
Models
for
Exploration,
2d
Ed.:
Minneapolis,
Burgess,
124
p.
Daniels,
R.
B.,
E.
E.
Gamble,
and
S.
W.
Boul,
1970,
Eolian
sands
associated
with
coastal
plain
river
valleys
- -
Some problems
in
their
age and source:
Southeastern
Geology,
v.
11,
p.
97-110.
Frey,
R.
W.,
and
I.
D.
Howard,
1986,
Mesotidal estuarine
sequences:
an
overview
of
the
Georgia
coastal
zone:
(unpublished manuscript).
Halls,
I.
R.,
and
J.
H.
Hoyt,
1969,
The significance
and
lim
itations
of
statistical parameters
for
distinguishing
ancient
and
modern sedimentary
environments
of
the
lower Georgia coastal
plain:
Journal
of
Sedimentary
Petrology,
v.
39,
p.
559-580.
Hamilton,
D.
B., 1984,
Plant
succession
and
the
influence
of
disturbance
in
Okefenokee
Swamp:
jCohen,
A.
D.,
D.
J.
Casagrande,
M.
J.
Andrejko,
and
G.
R.
Best,
Eds.,
The
Okefenokee
Swamp:
Its
natural
history,
geology,
and
geochemistry:
Los
Alamos,
NM,
Wetlands
Sur
veys,
p.
86-111.
Hayes,
M.
0.,
1979,
Banierisland
morphology
as
a
function
of
tidal
and wave
regime:
in
Leatherman,
S.
P.
(ed),
Barrier
Islands,
Academic
Press,
p.
1-27.
Hills,
E.
5,
1940,
Thelunette,
anewland
formofaeolian
on-
gin:
Australian Geographer,
v.
3(7),
p.
15-2
1.
Howard,
J.
D.,
and
R.
M.
Scott,
1983,
Comparison
of
Pleis
tocene
and
Holocene
barrier
island
beach-to-offshore
sequences, Georgia
and
northeast Florida
coasts,
U.S.A.:
Sedimentary Geology,
v. 34,
p.
167-183.
Izlar,
R.
L.,
1984,
Some
comments
on
fire and
climate
in
the
Okefenokee swamp-marsh complex:
in
Cohen,
A.
D.,
D.
J.
Casagrande,
M.
J.
Andrejko,
and
G.
R.
Best,
Eds.,
The
Okefenokee
Swamp: Its
natural
history, geology,
and
geochemistry:
Los
Alamos,
NM,
Wetlands
Sur
veys,
p.
70-85.
Lancaster,
I.
N.,
1978,
Composition
and
formation
of
south
ern
Kalahari
pan
margin
dunes:
Zeitschrift
fur
Geomor
phologie,
v.
22,
p.
148-169.
Kaczonowski,
R.
T.,
1977,
The
Carolina
bays,
a
comparison
with
modern oriented
lakes:
Technical
Report
13-CRD,
University
of
South Carolina,
124
p.
Klein,
G.
deV.,
1970,
Depositional
and
dispersal dynamics
63
JERRY
DAVIS
of
intertidal
sand
bars:
Journal
of
Sedimentary Petrol-
ogy,
v.
40,
p.
1095-1127.
Krinsley,
D.
H.,
and
S.
V.
Margolis,
1971,
Grain surface
tex
hire,
Chapter
8
in
Carver,
R.
E.,
Ed., Procedures
in
Sed
imentary
Petrology:
New
York,
Wiley,
p.
151-180.
Manker,
J. P.,
and
R.
D.
Ponder,
1978,
Quartz grain
surface
features
from
ftuvial environments
of
northeastern
Georgia:
Journal
of
Sedimentary Petrology,
v.
48,
p.
1227-1232.
McCaffrey,
C.
A.,
and
D. B.
Hamilton,
1980, A
vegetation
map
of
the
Okefenokee
Swamp.
Appended
to
Cohen,
A.
D.,
Casagrande,
M.
J.
Andrejko,
and
G.
R.,
Eds.,
1984,
The
Okefenokee
Swamp,
Its
natural history,
geology,
and
geochemistry: Los
Alamos,
NM,
Wetland
Surveys.
Miller, W.,
1979,
Stratigraphic framework
of
the
Wharton
Station dune
field,
eastemmost
Beaufort
County,
North
Carolina:
Southeastern
Geology,
v.
20,
p.
261-273.
Opdyke,
N.
D.,
D.
P.
Spangler,
D.
L.
Smith,
D.
S.
Jones,
and
R.
C.
Lindquist,
1984,
Origin
of
the
epeirogenic uplift
of
Pliocene-Pleistocene
beach ridges
in
Florida
and
development
of
the
Florida
karst: Geology,
v. 12, p.
226-228.
Parrish,
P.
K.,
and
Rykiel,
E.
J.,
Jr.,
1979,
Okefenokee
Swamp
origin, review
and
reconsideration:
Elisha
Mitchell
Science
Society
I.,
v.
95,
p.
17-31.
Pirkle,
W.
A, 1972,
Trail
Ridge,
arelic shoreline
feature
of
Honda
and Georgia:
Ph.D.
Dissertation, University
of
North Carolina, Chapel
Hill,
85
p.
Pirkle,
F.
L.,
andL.
J.
Czel,
1983,
Marine
fossils
from
region
ofTrail
Ridge,
a
Georgia-Ploridalandform:
Southeast-
em
Geology,
v.
24,
p.
31-38.
Reineck, H.-E.,
and
I.
B.
Singh,
1980,
Depositional
sedi
mentary
environments
with
reference
to
terrigenous
clastics: Berlin,
Springer-Verlag,
549
p.
Smedley,
J.
E., 1968,
Summary
report
on
the
geology
and
mineral
resources
of
Okefenokee
National
Wildlife
Refuge:
United
States
Geological
Survey
Bulletin
1260-N,
10
p.
Thames,
B. J, 1982,
Origin
ofsand
ridges
along streams
in
southeastern
Georgia: Unpubl.
MS.
thesis,
Emory
Uni
versity,
Atlanta,
Ga.,
152
p.
Thom,
B.
G, 1967,
Coastal
and
Fluvial
Landforms, Harry
and
Marion Counties,
South Carolina:
Tech.
Rept.
No.
44,
Coastal
Studies Institute,
Louisiana
State
Univer
sity,
Baton
Rouge.
Thom,
B.
G,
1970,
Carolina
bays
in
Horry
and
Marion
Counties, South
Carolina:
Geological
Society
of
Amer
ica
Bulletin,
v.
81,
p.
783-814.
Tucker,
M.
E,
1981,
Sedimentary petrology,
an
introduc
tion:
New
York,
Halsted-Wiley,
252
p.
Wadsworth,
J.
R.,
Jr.,
E.
R.
Blood,
and
D.
B.
Hamilton,
1984,
Applications
of
remote
sensing
to
ecosystem
studies
of
Okefenokee
Swamp,
in
Cohen,
A.
D.,
D. J.
Casagrande,
M. J.
Andrejko,
and
G.
R.
Best, Eds.,
The
Okefenokee
Swamp:
Its
natural
history,
geology,
and
geochemistry: Los
Alamos,
NM,
Wetland
Surveys,
p.
189-200.
Watts,
W.
A.,
1971,
Postglacial
and
interglacial
vegetation
history
of
southern
Georgia
and
central Florida:
Ecol
ogy,
v.
52,
p.
676-690.
Watts,
W.
A.,
and
M.
Stuiver,
1980,
Late
Wisconsin
climate
of
northern
Florida
and
the
origin
of
species-rich
decid
uous forest: Science,
v.
210(4467),
p.
325-327.
Wright,
L.
D,
1977,
Sediment
transport
and
deposition
at
river
mouths,
a
synthesis:
Geological
Society
of
Amer
ica Bulletin,
v.
88,
p.
857-868.
64
