Content uploaded by Eric Milbrandt
Author content
All content in this area was uploaded by Eric Milbrandt on Dec 03, 2018
Content may be subject to copyright.
A
multiple
habitat
restoration
strategy
in
a
semi-enclosed
Florida
embayment,
combining
hydrologic
restoration,
mangrove
propagule
plantings
and
oyster
substrate
additions
E.C.
Milbrandt
a,
*,
M.
Thompson
a
,
L.D.
Coen
b
,
R.E.
Grizzle
c
,
K.
Ward
c
a
Marine
Laboratory,
Sanibel-Captiva
Conservation
Foundation,
900A
Tarpon
Bay
Rd.,
Sanibel,
FL
33957,
United
States
b
Department
of
Biological
Sciences
and
Harbor
Branch
Oceanographic
Institute
at
Florida
Atlantic
University,
5700
U.S.
1
N,
Ft.
Pierce,
FL
34946,
United
States
c
Jackson
Estuarine
Laboratory,
University
of
New
Hampshire,
Durham,
NH
03824,
United
States
A
R
T
I
C
L
E
I
N
F
O
Article
history:
Received
28
April
2015
Received
in
revised
form
8
June
2015
Accepted
28
June
2015
Available
online
xxx
Keywords:
Clam
Bayou
SW
FL
Restoration
Habitats
Sanibel
Island
Monitoring
Seston
uptake
SAV
Oysters
Ecosystem
services
A
B
S
T
R
A
C
T
Habitat
loss
and
disturbance
are
ranked
globally
as
the
greatest
threats
to
biodiversity.
Development
and
coastal
population
growth
are
the
leading
causes
for
habitat
losses.
Recently,
the
restoration
of
marine
habitats
has
increased,
especially
with
the
goal
of
increasing
non-consumptive
ecosystem
services
derived
from
mangrove
and
submerged
aquatic
vegetation
(SAV)
along
with
biogenic
oyster
reefs.
Habitats
reside
in
landscapes
dominated
by
multiple
species.
Rather
than
focusing
on
a
single
habitat
such
as
oysters
or
mangroves
or
SAV,
we
took
an
approach
restoring
multiple
adjacent
habitats
to
accelerate
restoration
in
a
Florida
embayment
that
had
been
significantly
degraded
prior
to
the
restoration
of
natural
tidally
generated
flows.
After
a
multiple
habitat
die-off,
a
project
was
initiated
in
2006
to
reintroduce
tidal
flushing.
The
re-introduction
of
tidal
flushing,
however,
did
not
result
in
immediate
recovery
of
mangrove
shorelines
or
oyster-dominated
reefs.
There
was
a
lack
of
mangrove
propagule
production
and
significant
substrate
limitation
in
areas
with
appropriate
salinity,
sediment
and
tidal
flows.
From
2009–2012,
red
mangrove
(Rhizophora
mangle)
propagules
were
collected
(over
500,000)
and
planted
for
a
total
area
of
3.24
ha.
From
2009–2010,
five
intertidal
reefs
were
constructed
by
adding
bagged
and
fossil
shell
(54
MT)
for
Crassostrea
virginica
larvae
to
recruit
onto
totaling
over
779
m
2
.
Monitoring
of
planted
mangrove
versus
unplanted
shorelines
demonstrated
that
prop
root
and
drop
root
densities
were
higher
where
propagules
were
planted
(28
m
2
)
versus
unplanted
(2.3
m
2
).
Oyster
densities
and
mean
sizes
(multiple
year
classes)
at
new
and
natural
reefs
were
measured
after
8,
12,
and
24,
and
36
months.
An
initial
settlement
pulse
was
observed
in
the
first
8
months
followed
by
an
increase
in
the
density
of
greater
than
1-year
old
oysters.
Xanthid
crab
densities
(Eurypanopeus
depressus
and
Panopeus
spp.)
in
restored
reefs
and
natural
reefs
were
similar,
while
Petrolisthes
armatus
densities
were
lower
in
restored
reefs.
Whole
reef
seston
filtration
rates
over
restored
reefs
were
26
to
157
L
m
2
h
1
when
measured
at
4,15,
28,
and
40
months.
A
multiple
habitat
approach
may
be
useful
in
accelerating
the
natural
ecological
succession,
especially
if
the
project
site
has
reached
a
degraded,
alternate
ecological
state.
These
results
suggest
a
multiple
habitat
approach
can
be
useful
in
providing
non-provisioning
ecosystem
services
to
a
Florida
embayment.
ã
2015
Elsevier
B.V.
All
rights
reserved.
1.
Introduction
Habitat
loss
and
disturbance
are
ranked
worldwide
as
the
greatest
threats
to
biodiversity
(Roberts
et
al.,
2002;
Thrush
and
Dayton,
2002;
Hoekstra
et
al.,
2012;
Airoldi
et
al.,
2008).
Development
associated
with
coastal
population
growth
is
the
leading
contributor
to
habitat
loss
and
degradation
(e.g.,
Tilman,
1999;
Lotze
et
al.,
2006).
As
of
2003,
23
of
the
25
most
densely
populated
U.S.
counties
were
coastally-located
(Crossett,
2008),
with
Florida
recently
leading
the
nation
in
coastal
population
growth
(75%),
normalized
to
percent
change,
from
1980
to
2003.
Hydrological
restoration
of
tidal
flushing
offers
the
potential
to
re-connect
functional
linkages
in
mangrove
and
other
coastal
ecosystems
after
degradation
and
impairment.
Once
the
tidal
hydrology
is
restored,
the
resulting
improved
environmental
conditions
allow
ecological
successional
processes
to
occur,
*
Corresponding
author.
E-mail
address:
emilbran@sccf.org
(E.C.
Milbrandt).
http://dx.doi.org/10.1016/j.ecoleng.2015.06.043
0925-8574/ã
2015
Elsevier
B.V.
All
rights
reserved.
Ecological
Engineering
83
(2015)
394–404
Contents
lists
available
at
ScienceDirect
Ecological
Engineering
journal
homepage:
www.elsevier.com/locate/ecoleng
resulting
in
long-term
restoration
of
habitats
(,
2010).
This
approach
focus
on
levees,
dikes,
causeways,
and
roads
built
around
and
through
estuarine
habitats
from
either
dredge
and
fill
or
impoundment
activities
in
1940
to
present.
Barriers
to
natural
colonization
of
mangroves
such
as
propagule-limitation
(Lewis,
2005)
or
limitations
to
larval
recruitment
(Brumbaugh
and
Coen,
2009)
such
as
substrate-limitation
(Coen
et
al.,
2007)
can
benefit
from
actions
to
‘jump
start’
natural
ecological
processes
such
as
competition,
recruitment,
and
succession.
Restoration
activities
to
remedy
barriers
to
natural
succession
have
included
planting
mangrove
propagules
or
herbaceous
marsh
plants
to
improve
mangrove
seedling
establishment
(Milbrandt
and
Tinsley,
2006),
and/or
by
adding
oyster
cultch
or
live
animals
for
reef
development
(Coen
and
Luckenbach,
2000).
The
goal
of
this
project
was
to
determine
whether
such
activities,
when
combined
with
restoration
of
tidal
flushing,
could
enhance
and
restore
ecosystem
function
in
an
embayment
with
multiple
degraded
habitats.
Several
factors
affect
the
population
dynamics,
community
structure,
and
succession
of
mangroves
in
disturbed,
natural,
and
restored
mangrove
wetlands
(Smith
et
al.,
1994).
Large-scale
mangrove
die-offs
may
be
caused
by
the
collapse
of
the
underlying
peat
and
soil
structure
accompanying
subsidence
(Cahoon
et
al.,
2003)
or
alterations
in
the
tidal
hydroperiod
can
result
in
an
inundation
time
above
the
critical
level
tolerated
by
live
mangrove
trees
and
seedlings
(Lewis,
2005).
Lack
of
propagule
production
occurs
in
areas
where
there
are
no
mature
trees,
such
as
in
an
area
recently
experiencing
a
large
scale
disturbance.
While
the
planting
of
multiple
mangrove
species
consistent
with
nearby
community
structure
is
recommended
(Lewis,
2014)
as
a
restoration
strategy,
the
goal
was
to
increase
habitat
complexity
along
the
edge
of
the
shoreline.
The
shoreline
edge
in
southwest
Florida
is
composed
primarily
of
red
mangrove
(Rhizophora
mangle)
which
creates
a
complex
structural
habitat
for
fish
and
other
species.
Oyster
reefs
are
one
of
the
most
threatened
estuarine
habitats
worldwide
(Beck
et
al.,
2011).
Overharvesting,
sedimentation
due
to
human
development,
oyster
disease,
and
dredging
for
channeli-
zation
or
other
human
uses
have
reduced
oyster
reefs
by
an
estimated
85%
overall.
When
conditions
are
suitable,
native
oysters
(Crassostrea
virginica)
form
reefs
in
Southwest
Florida
(Volety
et
al.,
2009)
and
create
structurally
complex
habitat.
Oyster
reefs
filter
water
and
serve
as
prey
and
habitat
for
many
other
estuarine
animals
(Coen
et
al.,
1999;
Coen
et
al.,
1999).
Oysters
are
often
referred
to
as
‘ecosystem
engineers’
(Jones
et
al.,
1994)
because
of
their
ability
to
form
biogenic
3-dimensional
reef
structure
that
attracts
fishes
and
invertebrates.
Other
ecosystem
services
are
performed
by
oyster
reefs
and
include
providing
habitat
for
other
organisms,
water
filtration,
and
shoreline
stabilization.
Oysters
and
other
filter
feeding
invertebrates
that
naturally
occur
on
oyster
reefs,
remove
suspended
solids
from
the
water
and
some
have
suggested
that
this
can
clarify
the
water
to
enable
seagrass
growth
and
reduce
the
likelihood
of
harmful
algal
blooms
(Newell
and
Koch,
2004;
Newell
et
al.,
2007).
A
restoration
approach
is
described
here
which
focused
on
the
restoration
of
multiple
habitats
simultaneously
(hydrology,
mangrove,
oyster
reefs,
and
submerged
aquatic
vegetation).
Historically,
Clam
Bayou
and
Blind
Pass
are
part
of
a
southwest
Florida
barrier
island
and
were
biologically
diverse
‘back
bay’
habitats
connected
to
both
Pine
Island
Sound
through
mangrove
wetlands
and
to
the
Gulf
of
Mexico.
The
unique
combination
of
protected
mangrove
shorelines
and
regional
larval
recruitment
from
the
mouth
of
the
Caloosahatchee
Estuary
created
a
landscape-level
continuum
of
habitats
from
a
mangrove-vegetated
shoreline
connected
to
oyster
reefs
with
subtidal
channels
and
submerged
aquatic
vegetation
(mostly
Halodule
wrightii)
surrounding
the
reefs.
However,
the
dynamic
nature
of
barrier
islands
combined
with
human-caused
road
construction
and
beach
re-nourishment
led
to
a
large-scale
(50
hectares)
multiple
habitat
die-off
caused,
in
part,
by
hydrological
impairment.
A
blockage
of
tidal
flow
beginning
in
2000
caused
seasonal
flooding,
poor
water
quality,
fish
kills,
algae
blooms,
and
hypersalinity.
The
hydrology
was
restored
in
2006
by
the
installation
of
a
box
culvert
allowing
tidal
flushing
(NOAA
Restoration
Center
and
NOAA
Coastal
Services
Center,
2010).
Prior
to
2006,
a
disturbed
state
existed
in
Clam
Bayou
that
was
characterized
by
abundant
woody
debris
along
the
shorelines
as
a
result
of
the
mangrove
die-off
with
low
abundance
of
saltwort,
Batis
maritima,
and
white
mangrove
(Laguncularia
racemosa)-
dominated
shorelines.
There
were
no
observed
reproducing
red
mangroves
(R.
mangle),
the
typically
dominant
shoreline
species
in
southwest
Florida
(Odum
et
al.,
1982).
The
pre-culvert
water
levels
were
too
high
during
the
wet
period
and
too
low
during
the
dry
period
and
likely
caused
many
of
the
existing
eastern
oyster
(C.
virginica)
reefs
to
be
at
a
non-natural
elevation,
higher
than
the
intertidal
zone
and
thus
unavailable
as
substrate
for
new
oyster
recruitment.
There
were
also
areas
of
low
density,
scattered,
loose
shell
that
may
have
been
part
of
a
reef
at
one
time,
but
had
now
degraded
into
an
occasional
oyster
clump
in
sandy
mud.
These
characteristics
were
atypical
of
a
southwest
Florida
embayment
and
might
require
decades
to
achieve
a
functional
and
structural
equilibrium.
It
was
hypothesized
that
a
multiple
habitat
restoration
strategy
could
accelerate
the
natural
colonization
and
successional
processes
in
this
substrate
and
propagule-limited
embayment.
However,
as
with
many
restoration
projects,
it
was
not
possible
to
have
a
control
embayment
that
remained
hydrologically
impounded
to
compare
our
post-hydrological
restoration
ap-
proach.
The
goal
was
to
increase
non-provisioning
ecosystem
services
(Costanza
et
al.,
1997),
such
as
habitat
complexity,
diversity,
and
whole
reef
filtration,
in
addition
to
increasing
mangrove
and
oyster
abundances.
Propagule
collection
and
planting
of
the
shoreline
co-occurred
with
oyster
reef
construction
and
restoration.
Increases
in
oyster
size
and
density
(Luckenbach
et
al.,
2005;
Rodney
and
Paynter,
2006;
Hadley
et
al.,
2010)
were
expected
to
increase
the
water
filtering
capacity
and
increase
reef
resident
invertebrate
diversity.
Stabilization
of
the
shoreline
with
red
mangrove
propagules
was
expected
to
provide
increases
in
shoreline
structure
from
prop
and
drop
roots.
The
benefits
of
restoring
multiple
habitats
simultaneously
were
evaluated.
2.
Methods
2.1.
Site
description
Clam
Bayou
is
a
400-acre,
shallow,
mangrove-lined
embayment
located
on
Sanibel
Island,
Lee
County,
FL.
It
is
adjacent
to
one
of
the
most
visited
natural
areas
in
the
United
States,
the
J.N.
“Ding”
Darling
National
Wildlife
Refuge
and
to
the
Gulf
of
Mexico
beaches.
Land
use
in
the
area
is
a
mix
of
artificially
created
waterways,
natural
back
bays,
resource
protection
areas,
undeveloped
land,
and
residential
or
commercial
developments.
The
region
is
dynamic,
especially
on
the
Gulf
of
Mexico
side,
as
evidenced
from
the
position
of
channels
and
various
geomorphic
features
in
historical
aerial
photographs
(Wilson,
2003).
Clam
Bayou
suffered
from
a
multiple
habitat
die-off
caused,
in
part,
by
hydrological
impairment
(2000–2006)
when
the
sole
tidal
connection
was
blocked.
The
blockage
of
tidal
flow
caused
seasonal
flooding,
poor
water
quality,
fish
kills,
algae
blooms,
and
hypersalinity.
The
hydrology
was
restored
in
2006
by
the
installation
of
a
box
culvert
allowing
tidal
flushing
(Figs.
1
and
2;
NOAA
Restoration
Center
and
NOAA
Coastal
Services
Center,
2010).
E.C.
Milbrandt
et
al.
/
Ecological
Engineering
83
(2015)
394–404
395
2.2.
Re-establishing
the
red
mangrove
fringe
Shorelines
of
the
southwest
Florida
bays
are
typically
domi-
nated
by
a
red
mangrove
(R.
mangle)
edge
with
abundant
prop
roots
and
drop
roots
providing
fish
habitat.
Red
mangrove
propagules
were
collected
from
several
source
areas
on
Sanibel
Island
from
2009
to
2012.
Mature
propagules
were
available
from
September
through
December
and
were
collected
an
planted
at
15
different
sites
in
Clam
Bayou
(over
3
km
of
shoreline).
Propagules
were
planted
20
cm
apart
in
a
grid
from
the
existing
vegetation
(L.
racemosa)
from
1
to
6
m
depending
on
the
slope
of
the
shoreline.
The
approach
was
designed
to
‘saturate’
the
available
upper
intertidal
and
allow
natural
processes
(density-dependent
effects)
and
the
tidal
range
(post
culvert)
select
for
propagules
to
survive
within
the
existing
conditions.
Permanent
plots
(n
=
27)
were
established
to
determine
the
initial
mortality
of
the
propagules.
Shoreline
surveys
in
2015
were
conducted
to
compare
planted
to
unplanted
shorelines.
2.2.1.
Surveys
of
shorelines
planted
with
R.
mangle
versus
shorelines
with
no
planting
Permanent
transects
were
established
shortly
after
the
conclusion
of
the
planting
efforts
in
2010.
GPS
coordinates
were
used
to
permanently
designate
the
sites
and
a
30
m
transect
tape
was
extended
across
planted
shorelines
immediately
adjacent
to
shorelines
where
no
restoration
action
occurred.
Approximately
15
m
of
planted
and
15
m
of
adjacent
unplanted
shoreline
were
chosen
for
the
4
permanent
transects.
The
number
of
prop
roots
per
m
2
in
planted
shorelines
versus
unplanted
shorelines
was
recorded
in
January
2015.
The
sampling
event
occurred
5
years
post-restoration.
A
random
number
table
was
used
to
choose
5
quadrats
in
the
planted
area
and
5
in
the
unplanted
area.
The
number
of
drop
roots
or
prop
roots
per
square
meter
was
recorded.
Data
were
analyzed
with
a
t-test
to
compare
means
in
the
planted
versus
unplanted
areas.
2.3.
Pre-restoration
oyster
site
evaluation
In
order
to
assess
the
potential
for
natural
recruitment
to
the
constructed
oyster
reefs,
standard-sized
(30
cm
45
cm)
recruit-
ment
monitoring
trays
filled
with
fossil
oyster
shell
(Rehrig
Pacific;
Fig.
4)
were
deployed
in
August
2008
and
retrieved
in
May
2009.
High
settlement
of
oyster
spat
occurs
from
June
to
October
in
southwest
Florida
(Volety
et
al.,
2009).
Clam
Bayou
may
be
considered
a
“trap
estuary”
(Pritchard,
1952)
and
has
a
high
degree
of
water
retention
that
potentially
promotes
retention
of
shellfish
larvae
and
colonizing
species.
The
oyster
reef
construction
sites
were
chosen
in
areas
adjacent
to
planned
shoreline
mangrove
restoration
and
with
relatively
high
tidal
flows.
Sites
without
significant
living
oysters
and
no
seagrass
were
chosen
to
avoid
replacing
existing
habitat.
Other
consider-
ations
included
minimizing
the
distance
to
existing
natural
reefs,
within
the
intertidal
zone,
and
firm
substrate.
The
restoration
sites
typically
had
degraded,
low
density
live
oyster
clusters
from
pre-
culvert
conditions.
2.3.1.
Oyster
reef
construction
Five
sites
(TNC-1,
TNC-2,
TNC-3,
NACo-1,
NACo-2)
were
established
for
restoration.
Approximately
54
MT
(54,000
kg)
of
fossilized
large,
washed
shell
from
SMR
Aggregates
(Sarasota,
FL)
were
delivered
to
a
vacant
area
in
Bowman’s
Beach
County
Park.
Three
truckloads
(40.5
MT)
were
delivered
in
2009
and
one
truckload
was
delivered
in
2011.
The
shell
was
shoveled
and
raked
into
5-gallon
buckets.
The
buckets
were
used
to
fill
PVC
tubes
(20
cm
diameter,
183
cm
length;
schedule
40)
covered
with
Delstar
#1142
mesh,
cut
from
rolls
into
150
cm
pieces.
One
end
was
tied
with
a
knot
and
the
bag
was
filled
with
fossil
shell
and
knotted
to
form
a
mesh
shell
bag.
The
shell
bags
were
stacked
on
pallets
at
Bowman’s
Beach
and
remained
there
until
being
loaded
on
a
flatbed
trailer
for
transport
to
Clam
Bayou.
A
22-foot
pontoon
boat
to
transport
shell
bags
to
the
reef
construction
sites
was
donated
(Jensen’s
Marina,
Captiva,
FL).
In
May
and
June
2010,
approximately
4200
bags
of
shell
were
loaded
by
volunteers
and
transferred
to
the
pontoon
boat.
The
bags
of
shell
were
unloaded
into
a
previously
marked
reef
site,
one
layer
in
height.
The
bags
were
placed
end
to
end
with
no
spaces.
Adjustments
were
made
to
fill
holes
or
move
misplaced
bags
during
low
tides
in
July
and
August
2010.
A
second
phase
of
reef-
construction
was
accomplished
in
May
2011.
2.3.2.
Oyster
reef
monitoring
All
four
“universal
metrics”
recommended
by
(Baggett
et
al.,
2014)
were
measured
to
determine
restoration
success
of
the
constructed
oyster
reefs.
Additionally,
the
associated
invertebrate
taxa
that
colonized
the
reefs
were
also
assessed.
Recruitment
tray
Fig.
1.
Cross
section
of
the
concrete
box
culvert
structure
installed
to
restore
tidal
hydrology
in
2006.
Fig.
2.
Aerial
image
(2013)
of
the
study
area.
The
white
arrow
indicates
the
box
culvert
installed
to
reintroduce
tidal
flushing
in
2006.
396
E.C.
Milbrandt
et
al.
/
Ecological
Engineering
83
(2015)
394–404
samples
provided
oysterrecruitmentand
growth
data
and
the
extent
of
habitat
use
by
reef-associated
species.
Each
tray
at
the
restoration
sites
were
lined
with
a
fine
mesh
(0.5
cm,
Industrial
Netting),
filled
with
fossil
shell,
and
then
covered
with
larger
mesh
(2
cm,
Industrial
Netting).
The
trays
were
numbered
with
aluminum
tags
(Forestry
Suppliers)
and
covers
were
secured
with
galvanized
wire
or
cable
ties.
Trays
were
also
deployed
in
natural
reefs
located
in
the
region
(Fig.
4)
by
and
embedding
trays
of
natural
shell
clusters
within
the
reefs
and
enclosing
trays
with
similar
mesh.
The
trays
were
deployed
in
late
July
2010
shortly
after
the
reefs
were
constructed
and
held
in
place
with
20
cm
j-shaped
rebar
stakes.
They
were
positioned
at
approximately
mid
reef
elevation
equally
spaced
along
the
length
of
each
reef.
Trays
were
deployed
in
multiples
of
3
to
provide
replication
for
statistical
analysis
(see
below).
The
first
set
of
monitoring
tray
samples
(n
=
3
per
reef)
were
retrieved
in
May
2011
(after
8
months)
to
capture
one
high
spat
settlement
period.
Sets
of
three
trays
were
collected
again
in
early
Aug
2011
to
evaluate
conditions
1
year
post-construction.
Trays
were
collected
in
August
of
2012
and
2013
for
determining
reef
development.
The
contents
of
the
trays
were
sorted
in
a
large
stainless
steel
sieve
with
2
mm
stainless
mesh
after
they
were
retrieved
from
the
field.
Live
oysters,
fish
and
invertebrates
were
separated
from
fossil
shell
and
dead
shell
fragments.
Shell
height
was
measured
for
all
live
oysters
with
60
mm
calipers
(Fowler,
MA)
and
logged
in
an
Microsoft
Excel
spreadsheet
(Starrett,
MA).
Fish
and
invertebrate
fauna
was
fixed
with
5%
formalin,
enumerated,
and
identified
to
species
level
where
possible.
For
the
genus
Panopeus,
there
is
genetic
evidence
of
several
species
in
southwest
Florida
(Felder
et
al.,
2010).
Voucher
specimens
were
sent
to
the
Florida
DEP
(Walton,
pers.
comm.)
and
to
the
Bailey-Matthew
Shell
Museum
(Leal,
pers.
comm.)
for
verification.
The
footprints
of
the
constructed
reefs
were
mapped
using
a
Trimble
GeoXT
GPS
to
determine
the
total
area
of
each
reef.
Reef
elevations
were
estimated
using
a
laser
level
(Lasermark
Wizard,
CST
Berger,
USA)
and
graduated
surveying
rod.
Elevation
above
sediment
was
calculated
from
at
least
three
measurements
per
reef
taken
on
sediment
surface
near
bottom
edge
of
reef
subtracted
from
elevation
to
top
of
reef.
2.3.3.
Oyster
reef
filtration
The
removal
of
seston
(using
chlorophyll
a
as
a
proxy
at
the
oyster
restoration
sites
was
investigated
using
in
situ
fluorometry
following
methods
described
by
Grizzle
et
al.
(2006,
2008).
Sites
TNC-1
and
TNC-2
and
on
two
occasions
a
nearby
natural
reef
were
sampled
using
an
upstream/downstream
approach
reviewed
in
Dame
(1996)
relying
on
repeated
in
situ
measurements
of
chlorophyll
a
(chl
a)
concentration
with
field
fluorometers
(Turner
Designs,
CA).
Discrete
water
samples
were
taken
during
each
measurement
period
and
analyzed
in
the
laboratory
for
chl
a
following
an
acetone
extraction
(EPA
protocol
445.0);
the
resulting
data
were
used
to
calibrate
the
in
situ
fluorometers.
In
addition
to
fluorometry
measurements,
water
flow
speed
(mid-depth)
was
measured
with
a
Marsh-McBirney
(Hach,
CO)
electromagnetic
flow
meter.
Water
depth,
salinity,
temperature
and
dissolved
oxygen
were
measured
with
a
YSI
6920V2
(YSI,
OH)
during
each
sampling
event.
Water
soluble
dye
was
released
sporadically
during
the
measurement
period
at
the
upstream
position
and
observed
to
determine
water
direction
and
flow
relative
to
the
axis
of
the
reef.
The
dye
releases
also
provided
information
on
horizontal
mixing
of
the
water
flow
field.
Filtration
measurements
were
made
on
four
occasions:
4,
15,
28
and
40
months.
post-
restoration.
Whole-reef
water
filtration
based
on
differences
between
average
upstream
and
downstream
chl
a
concentrations
were
calculated
and
expressed
in
two
different
ways
using
the
following
equations.
Chl
a
%
removal
=
(C
1
–
C
2
)/C
1
100
(1)
Filtration
(clearance)
<
rate
(L
m
2
h
1
)
=
([A
xsec
v
1000]/A
reef
)
(%
CHL
Removal/100)
(2)
Fig.
3.
Map
of
Clam
Bayou
and
the
areas
where
mangrove
shorelines
were
planted
with
Rhizophora
mangle
propagules
(approximately
500,000
planted).
Fig.
4.
Map
of
oyster
reef
restoration
sites
in
Clam
Bayou.
Table
1
Mean
prop
root
density
(m
2
)
in
planted
shorelines
versus
unplanted
shorelines.
Shoreline
type
n
Mean
St.
Dev.
Planted
21
28.19
8
Unplanted
23
2.391
3.201
Table
2
Constructed
oyster
reef
metrics.
Reef
site
Area
Elevation
(m,
MSL)
Relief
(m)
Distance
to
natural
reef
(m)
TNC-1
135
0.339
0.057
150
TNC-2
242
0.391
0.116
50
TNC-3
260
0.419
0.092
150
NACo-1
86
0.435
0.039
122
NACo-2
56
0.480
0.076
10
E.C.
Milbrandt
et
al.
/
Ecological
Engineering
83
(2015)
394–404
397
where
C
1
=
upstream
chl
a
concentration
(mg
L
1
),
C
2
=
down-
stream
chl
a
concentration
(mg
L
1
),
A
xsec
=
cross-sectional
area
of
water
column
(m
2
),
v
=
water
flow
speed
(m
h
1
),
and
A
reef
=
reef
surface
area
(m
2
).
2.3.4.
Statistical
analyses
Data
from
five
reefs
constructed
within
the
same
period
(TNC-1,
TNC-2,
TNC-3,
NACo-1,
NACo-2)
were
used
in
the
analyses.
All
constructed
reefs
were
located
within
Clam
Bayou
and
this
data
was
compared
to
results
from
natural
reefs
monitored
in
Clam
Bayou,
Tarpon
Bay
and
Pine
Island
Sound.
The
same
tray
method
as
described
for
constructed
reefs
was
used
to
sample
natural
reef
locations.
Trays
were
placed
at
6
natural
reef
sites
in
Clam
Bayou,
4
sites
in
Tarpon
Bay,
and
3
sites
in
Pine
Island
Sound.
One
tray
from
each
location
was
collected
after
8,12,
and
24
months.
In
addition,
3
natural
reef
sites
in
Clam
Bayou
were
sampled
36
months
after
reef
construction
Live
oyster
size
distribution
for
three
constructed
reef
sites
and
three
natural
reefs
was
examined
by
comparing
mean
oyster
density
at
constructed
and
natural
sites
(treatment)
for
the
8,
12,
24,
and
36
month
post-construction
sampling
events
(time).
Mean
values
were
calculated
from
the
number
(n)
of
monitoring
trays
deployed
in
each
group
or
site
for
a
reported
time
period.
Data
transformations
were
conducted
to
satisfy
normality
assumptions
of
statistical
tests
used
and
homogeneity
of
variance
was
tested
using
Levene’s
test.
Oyster
densities
were
log-transformed
comparisons
of
the
means
were
made
using
a
two-factor
general
linear
model
(time,
treatment)
with
Tukey’s
pairwise
comparisons
(Minitab
v.
13).
For
all
statistical
comparisons,
a
p-value
of
less
than
0.05
was
considered
significant.
Fig.
5.
Time
series
boxplots
of
live
oyster
densities
at
restored
and
natural
reefs.
0
50
100
150
200
250
2.5
7.5 12
.5 17
.5 22
.5 27
.5 32
.5 37
.5
42.5 47
.5 52
.5 57
.5 62
.5 67
.5 72
.5 77
.5 82
.5
87.5
Oyster Density/m
2
Size
Class
Midpoi
nt
(mm
SH)
Reef TNC
1
8 months
12
mon
ths
24
mon
ths
36
mon
ths
0
50
100
150
200
250
2.5
7.5
12.5
17.5
22.5
27.5
32.5
37.5
42.5
47.5
52.5
57.5
62.5
67.5
72.5
77.5
82.5
87.5
Oyster Density/m2
Size Class Midpoint
(mm SH
)
Reef TNC
2
8
months
12
months
24
months
36
months
Fig.
6.
Size
frequency
distribution
of
live
oysters
at
restored
sites
in
Clam
Bayou.
398
E.C.
Milbrandt
et
al.
/
Ecological
Engineering
83
(2015)
394–404
Reef
resident
samples
from
the
trays
were
compared
in
restored
and
natural
sites
using
multivariate
approaches.
Fishes
were
excluded
from
the
analysis
because
of
their
ability
to
escape
the
trays
easily
during
retrieval.
Other
gear
types,
such
as
hook
and
line
and
lift
nets
are
preferred
for
highly
mobile
species
(Tolley
et
al.,
2005).
Differences
in
resident
invertebrate
community
composition
between
treatment
and
time
were
analyzed
using
a
Bray–Curtis
similarity
index
of
square
root
transformed
data
(Primer
v.
6.1,
ANOSIM).
For
all
statistical
comparisons,
significance
values
(p)
of
less
than
0.05
were
considered
significant.
MDS
ordinations
were
also
derived
from
similarity
matrices
(Clarke
and
Warwick,
2001).
Diversity
metrics
including
total
abundance,
species
richness,
and
Shannon
diversity
index
were
calculated
using
the
Primer
software
DIVERSE
function.
Association
between
environmental
data
and
resident
invertebrate
community
structure
was
evaluated
using
the
Primer
BIOENV
function
which
identifies
the
subset
of
environmental
data
that
maximizes
the
rank
correlation
with
biotic
data
similarity
matrices.
When
using
the
BIOENV
function
the
following
environmental
variables
were
used:
recruit
density,
large
oyster
density,
total
oyster
density,
salinity,
chl
a,
turbidity,
reef
size,
and
distance
from
nearest
natural
reef.
When
evaluating
only
constructed
reefs,
the
variables
reef
elevation,
and
reef
height
above
sediment
were
added
to
those
listed
above.
The
Primer
RELATE
function
was
used
to
test
correlation
between
similarity
matrices
for
environmental
variables
and
biotic
variables
by
reef.
Relationships
between
densities
of
xanthid
crabs,
porcelain
crabs
and
mussel
and
oyster
density
were
investigated.
RELATE
was
also
used
to
investigate
associ-
ations
between
water
quality
variables
and
oyster
density.
When
using
RELATE,
a
significance
level
of
0.05
and
a
rho
(r)
value
greater
than
0.5
was
considered
a
significant
relationship.
3.
Results
3.1.
Red
mangrove
shoreline
restoration
A
total
3.24
ha
and
3.53
km
of
shoreline
was
planted
with
red
mangrove
(R.
mangle)
propagules
(Fig.
3).
The
footprint
of
each
location
was
variable
due
to
the
location
and
geomorphology
of
the
site.
Some
sites
were
large
broad
areas
and
others
were
steeper
sloped.
The
largest
area
was
0.80
hectares
and
the
smallest
was
0.012
ha.
The
density
of
prop
roots
in
permanent
transects
at
4
locations
indicated
significantly
higher
densities
in
planted
shorelines
versus
unplanted
shorelines
(MINITAB;
ANOVA;
F
=
203.94;
p
<
0.00).
Prop
root
density
in
planted
shorelines
was
28.19
m
2
while
prop
root
density
in
adjacent
unplanted
areas
was
2.391
m
2
(Table
1).
3.2.
Oyster
reef
construction
A
total
of
5
reefs
were
constructed
for
a
total
area
of
850
m
2
(Fig.
2,
Table
2).
The
restored
reef
site
names
were
TNC-1
(135
m
2
),
TNC-2
(232
m
2
),
TNC-3
(260
m
2
),
NACo-1
(86
m
2
),
NACo-2
(56
m
2
).
Elevations
relative
to
mean
sea
level
(MSL),
the
reef
relief
(m),
and
estimated
distance
of
the
restored
reefs
to
a
natural
reef
is
provided
in
Table
2.
3.3.
Oyster
metrics
Live
oyster
densities
were
compared
by
treatment
(con-
structed
versus
natural)
and
age
(8,
12,
24,
36
months).
Densities
at
constructed
sites
(669
m
2
)
were
lower
than
natural
sites
(2063
m
2
)
initially
(8
months),
but
over
time
the
natural
densities
decreased
while
the
constructed
sites
stayed
relatively
consistent
(Fig.
5).
The
size
frequency
distribution
of
the
natural
0
100
200
300
400
500
2.5 7.
512
.5 17
.5
22.5 27
.5 32
.5 37
.5
42.5
47.5 52
.5 57
.5 62
.5
67.5 72
.5 77
.5 82
.5
87.5
Oyster Density/m
2
Size
Class Mi
dpoint
(mm SH
)
Natu
ral
Reef
Clam
Bayou
8
months
12
month
s
24
month
s
36
month
s
Fig.
7.
Size
frequency
of
live
oysters
at
a
nearby
natural
oyster
reef
in
Clam
Bayou.
Table
3
Comparision
of
abundance
and
diversity
of
invertebrate
species.
Reef
site
Shannon
diversity
Species
richness
Total
abundance
(m
2
)
Xanthidae
Petrolisthes
armatus
TNC-1
1.941
15
2,554
18
0
TNC-2
1.561
15
4,232
6
0
TNC-3
1.849
11
1,633
45
18
NACo-1
2.079
19
4,696
12
0
NACo-2
2.078
21
5,146
57
8
CBNat
a
1.54
22
9,349
87
40
TBNat
a
1.54
18
14,13 4
222
1,024
PISNat
a
1.08
15
19,002
269
2,715
a
Natural
oyster
reefs.
E.C.
Milbrandt
et
al.
/
Ecological
Engineering
83
(2015)
394–404
399
sites
shows
high
settlement
throughout
the
study
area
during
the
initial
8
months
of
the
reef
construction
(Fig.
6).
A
similar
finding
at
TNC-2
suggests
that
for
a
subset
of
constructed
sites,
settlement
on
natural
reefs
was
comparable
to
natural
reefs
(Fig.
7).
However,
high
settlement
was
not
observed
at
the
other
constructed
sites
where
samples
were
collected
(TNC-1,
TNC-3).
Large
densities
of
juvenile
oysters
(<60
mm)
were
present
for
the
first
two
sampling
events
(8
and
12
months).
Settlement
during
subsequent
sampling
events
at
constructed
and
natural
sites
was
lower
and
there
were
fewer
spat
and
juvenile
oysters.
Overall,
these
data
indicate
strong
initial
recruitment
to
both
constructed
and
natural
reefs
but
greater
densities
on
the
natural
reefs,
then
declines
in
total
oyster
densities
on
both.
By
the
24-month
sampling,
there
were
no
differences
between
natural
and
constructed
reefs
in
total
oyster
densities
and
both
had
similar
size
frequency
distributions.
Thus,
the
expected
trend
of
increasing
oyster
densities
on
the
constructed
reefs
was
not
observed
over
the
duration
of
the
study.
3.4.
Reef
residents
The
abundance
and
diversity
of
invertebrate
species
using
the
constructed
reefs
versus
the
natural
reefs
was
compared
after
12
months
(Table
3).
The
Shannon
diversity
index
and
species
richness
at
the
construction
site
was
similar
to
the
natural
sites
after
12
months.
The
abundance
of
reef
residents,
however,
was
significantly
higher
at
the
natural
sites
than
the
constructed
sites.
The
abundance
of
all
reef
invertebrates
in
the
constructed
reefs
was
between
1633
m
2
(TNC-3)
and
5146
m
2
(NACo-1)
while
abundances
at
natural
reefs
was
between
9340
m
2
(CBNat)
and
19,002
m
2
(PISNat).
The
number
of
xanthid
crabs
and
Petrolisthes
also
reflected
these
differences.
3.5.
Oyster
reef
filtration
There
was
wide
variability
in
filtration
by
both
the
constructed
and
natural
reefs,
and
no
clear
trends
over
time
(Table
4).
In
particular,
the
expected
trend
of
increasing
filtration
as
the
Table
4
Clam
Bayou
summary
of
environmental
and
oyster
reef
filtration
data
2010–2013.
Reef
Sampling
date
Months
post-
contruction
Sampling
duration
(min)
Water
depth
(cm)
Flow
length
(m)
Flow
speed
(cm/s)
Salinity
(ppt)
Temp
(
C)
Dissolved
oxygen
(mg/L)
%
Chl
a
removal
Clearance
rate
(L/
m
2
/hr)
TNC
1
10/16/
2010
4
100
13
15.0
1.5
35.8
24.4
6.6
21.4
132
TNC
2
10/19/
2010
4
35
23
16.5
2.4
35.7
25.6
6.1
10.4
67
TNC
2
10/19/
2010
4
25
28
16.5
1.0
35.6
25.6
5.9
15.9
102
TNC
2
10/19/
2010
4
65
25
16.5
1.5
35.6
25.6
5.9
22.9
148
TNC
1
9/27/2011
15
20
8
16.9
2.5
34.4
28.3
4.1
15.2
91
TNC
2
9/27/2011
15
27
20
14.5
2.8
33.4
28.7
4.9
21.2
107
TNC
2
9/28/2011
15
8
12
17.0
3.4
34.2
28.7
3.4
4.4
26
TNC
1
10/24/
2012
28
50
21
8.7
3.0
35.2
25.6
4.5
3.0
10
TNC
2
10/24/
2012
28
51
29
15.0
2.2
35.4
26.5
6.3
11. 5
65
TNC
1
10/24/
2012
28
22
24
16.8
1.0
36.2
27.1
6.8
25.3
157
TNC
2
10/25/
2012
28
140
16
11.7
3.7
36.0
24.7
3.9
0.9
4
CBNat
10/25/
2012
28
48
22
10.5
5.2
35.6
25.6
4.3
10.7
44
TNC
1
10/18/
2013
40
30
15
10.0
1.5
32.1
29.2
4.8
11. 0
38
TNC
2
10/16/
2013
40
69
45
12.3
1.0
32.8
28.0
5.0
7.0
31
TNC
2
10/17/
2013
40
58
25
9.8
3.2
32.4
28.1
5.9
6.7
23
CBNat
10/16/
2013
40
153
26
14.2
(n/a)
32.9
28.1
5.1
21.8
(n/a)
CBNat
10/16/
2013
40
351
13
14.2
(n/a)
32.4
28.8
6.2
31.4
(n/a)
CBNat
10/16/
2013
40
115
19
14.2
(n/a)
32.7
28.1
4.8
7.7
(n/a)
CBNat
10/17/
2013
40
375
14
11.6
(n/a)
32.6
27.5
3.9
40.8
(n/a)
Fig.
8.
Multiple
habitat
restoration
(oyster
cultch
addition
to
provide
substrate
and
mangrove
propagule
planting)
in
Clam
Bayou.
400
E.C.
Milbrandt
et
al.
/
Ecological
Engineering
83
(2015)
394–404
constructed
reefs
developed
was
not
observed.
The
Clam
Bayou
reefs
(constructed
and
natural)
had
filter
feeding
taxa
other
than
oysters
at
substantial
densities.
Thus,
our
in
situ
filtration
measurements
represent
more
than
that
of
oysters.
In
order
to
assess
the
relationship
between
the
measured
filtration
rates,
scatterplots
and
correlations
analyses
were
performed
on
various
combinations
of
oyster
and
other
filter
feeder
densities
with
respect
to
filtration
rates.
Although
there
was
a
weak
positive
(though
non-significant)
correlation
between
filter
feeder
density
(oysters
and
other
species
of
filter
feeders
combined)
and
%
chl
a
removal,
no
correlations
were
significant.
On
two
occasions
chl
a
concentrations
averaged
higher
downstream
than
upstream,
indicating
export
from
the
reef
(Table
4).
On
one
of
these
occasions,
wind
speed
increased
producing
waves
and
turbidity
clouds
in
the
water
column
above
the
reef
during
the
measurement
period.
Turbidity
clouds
also
were
observed
on
the
second
occasion,
perhaps
caused
by
striped
mullet
feeding
on
the
reef.
Thus,
these
two
occasions
likely
represent
physical
disturbance
events
to
the
reef
surface
causing
re-suspension
of
sediments
and
presumably
associated
chl
a.
It
should
be
noted
that
these
data
do
not
necessarily
indicate
that
the
suspension
feeders
on
the
reef
were
not
feeding,
but
rather
that
any
feeding
‘signal’
was
masked
by
export
of
chl
a
during
the
disturbance.
Excluding
these
two
datasets,
the
constructed
reefs
averaged
14.1%
chl
a
removal
and
80.4
L
m
2
h
1
filtration
for
the
duration
of
the
study.
Fewer
measurements
were
made
on
the
natural
reefs
but
their
mean
filtration
rates
were
higher,
averaging
25%
chl
a
removal.
4.
Discussion
Simultaneous
habitat
restoration
projects
co-occurred
to
stabilize
the
shoreline
and
construct
oyster
reefs
in
order
to
increase
mangrove
and
oyster
abundances
and
improve
ecosystem
services
(Figs.
8
and
9).
The
goal
was
to
determine
whether
a
multiple
habitat
approach
could
‘jump
start’
a
propagule
and
substrate-limited
area.