Mineral versus organic contribution to vertical accretion and elevation change in restored marshes (Ebro Delta, Spain)

Full text

Ecological
Engineering
61 (2013) 12–
22
Contents
lists
available
at
ScienceDirect
Ecological
Engineering
j
ourna
l
ho
me
pa
g
e:
www.elsevier.com/locate/ecoleng
Mineral
versus
organic
contribution
to
vertical
accretion
and
elevation
change
in
restored
marshes
(Ebro
Delta,
Spain)
Juan
Calvo-Cuberoa,∗,
Carles
Ibá˜
nezb,
Albert
Rovirab,
Peter
J.
Sharpec,
Enrique
Reyesa
aDepartment
of
Biology,
East
Carolina
University,
Greenville,
NC,
USA
bAquatic
Ecosystems
Program,
IRTA.
St.
Carles
de
la
Ràpita,
Catalonia,
Spain
cUS
National
Park
Service,
120
Chatham
Lane,
Fredricksburg,
VA
22405,
USA
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
2
May
2013
Received
in
revised
form
9
August
2013
Accepted
20
September
2013
Keywords:
Sea-level
rise
Marsh
restoration
Root
growth
Marsh
elevation
Sediment
inputs
Paspalum
distichum
a
b
s
t
r
a
c
t
The
Ebro
Delta
(Catalonia,
Spain)
is
one
of
the
most
valuable
coastal
zones
within
the
Mediterranean
Sea,
supporting
a
highly
productive
rice
agricultural
system,
as
well
as
a
myriad
of
coastal
marsh
habi-
tats.
However,
chronic
reductions
of
fluvial
sediments
coupled
with
accelerated
relative
sea-level
rise
have
created
an
environment
where
approximately
half
of
the
Ebro
Delta
is
now
vulnerable
to
flood-
ing
impacts.
To
assess
relative
sea-level
rise
(RSLR)
mitigation
options
through
marsh
restoration
within
abandoned
deltaic
rice
fields,
we
established
the
experimentally
restored
marshes
spanning
three
years.
We
used
two
freshwater
input
type
treatments
(riverine
irrigation
and
rice
field
drainage
water)
and
three
water
level
treatments
(10,
20
and
30
cm
deep).
Our
hypotheses
were
that:
(1)
vertical
accretion
and
elevation
change
in
oligohaline
restored
marshes
would
be
primarily
controlled
by
organic
contrib-
utions
under
sediment-deficit
conditions,
and
(2)
both
vertical
accretion
and
elevation
change
would
demonstrate
higher
rates
compared
with
predicted
RSLR
in
the
Ebro
Delta
(5–8
mm
yr−1).
Vertical
accre-
tion
had
higher
mean
values
in
both
water
type
treatments
(11.5
and
15.5
mm
yr−1)
than
elevation
change
(9.1
and
8.8
mm
yr−1).
Vertical
accretion
(but
not
elevation
change)
was
significantly
higher
in
drainage
water
treatment
receiving
greater
sediment
mineral
input,
which
caused
higher
surface
soil
mineral
con-
tent.
Conversely,
experimentally
restored
marshes
closer
to
rice
fields
in
both
water
type
treatments
had
greater
elevation
change
(11.3
and
17.8
mm
yr−1)
than
vertical
accretion
(8.3
and
15.1
mm
yr−1)
due
to
higher
belowground
biomass
because
of
high
weed
colonization
by
Paspalum
distichum
L.
These
results
showed
that
vertical
accretion
and
elevation
change
were
generally
controlled
by
mineral
contribution,
although
fast
growing,
ruderal
plant
species
such
as
P.
distichum
can
play
a
significant
role
in
marsh
ele-
vation
via
root
growth.
The
results
supported
the
hypothesis
that
restored
marshes
using
either
water
type
promote
marsh
elevation
gains
higher
than
predicted
RSLR
at
least
during
the
initial
marsh
devel-
opment
(3
years).
This
study
indicates
that
the
use
of
agricultural
runoff
as
a
primary
source
of
sediment,
nutrient,
and
freshwater
is
beneficial
for
marsh
restoration
projects
focused
primarily
on
mitigating
RSLR.
This
research
also
highlights
how
nuisance
species
such
as
P.
distichum
can
play
a
key
role
in
mitigating
RSLR
impacts
when
inexpensive
and
effective
measures
are
needed
to
promote
marsh
elevation
as
the
primary
restoration
goal.
Published by Elsevier B.V.
1.
Introduction
The
Ebro
Delta
is
a
vital
coastal
ecosystem
in
the
western
Mediterranean
extending
∼330
km2(Fig.
1)
and
is
the
second
most
important
special
protection
area
for
birds
(SPA)
in
Spain
(Seo/BirdLife,
1997).
The
Ebro
Delta
possesses
a
diverse
number
of
ecosystems
including
coastal
lagoons,
marshes
and
seagrasses,
which
comprise
the
Ebro
Delta
Natural
Park
and
are
part
of
∗Corresponding
author.
Tel.:
+1
252
328
5778.
E-mail
address:
calvocuberoj11@students.ecu.edu
(J.
Calvo-Cubero).
the
Natura
2000
network
of
the
European
Union
(EU).
Jux-
taposed
with
these
natural
areas
there
are
20,000
Ha
of
rice
fields.
Rice
agriculture
is
the
main
economic
activity
within
the
Delta
comprising
up
to
60%
of
the
land
surface
of
the
Ebro
Delta,
providing
an
annual
gross
income
of
about
D
60
million
and
a
total
rice
production
of
120,000
tons
per
year
(Cardoch
et
al.,
2002).
Although
rice
agriculture
development
has
trans-
formed
much
of
the
Ebro
Delta
over
the
last
two
centuries
(Cardoch
et
al.,
2002),
rice
fields
provide
significant
ecosystem
services
such
as
seasonal
habitat
for
migratory
birds,
preven-
tion
of
saline
intrusion
and
nutrient
removal
(Martinez-Vilalta,
1995).
0925-8574/$
–
see
front
matter.
Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.ecoleng.2013.09.047

J.
Calvo-Cubero
et
al.
/
Ecological
Engineering
61 (2013) 12–
22 13
Fig.
1.
The
top
map
shows
the
location
of
the
experimentally
restored
marshes
in
the
Ebro
Delta
(Catalonia,
Spain).
The
bottom
map
shows
its
detailed
location
between
an
active
organic
rice
field
and
an
old
restored
marsh
dominated
by
Phrag-
mites
australis
(Cav.)
Steudel
and
Typha
latifolia
L.
The
marshes
and
rice
fields
within
the
Delta
receive
irrigation
water
from
the
Ebro
River,
which
is
the
largest
river
of
the
Iberian
Peninsula
(flow
ca.
400
m3s−1).
A
series
of
dams
(∼170)
were
built
along
the
watercourse
during
the
1960s
to
support
a
vari-
ety
of
intensive
water
uses
(Ibá˜
nez
and
Prat,
2003).
These
dams
retain
an
estimated
99%
of
the
sediment
that
would
normally
be
deposited
within
the
Ebro
Delta,
thus
creating
a
severe
sediment
deficit
(Ibá˜
nez
et
al.,
1996).
Global
eustatic
sea-level
rise
(ESLR)
has
increased
at
a
rate
of
1–2
mm
yr−1over
the
last
century
and
it
is
now
higher
than
3
mm
yr−1(FitzGerald
et
al.,
2008).
However,
pre-
dicted
relative
sea-level
rise
(RSLR)
in
Ebro
Delta
may
range
from
5
to
8
mm
yr−1at
the
end
of
the
present
century,
due
to
ESLR
and
land
subsidence
(Ibá˜
nez
et
al.,
2010).
Both
sediment
reduction
and
RSLR
have
created
an
environment
where
∼40%
of
the
emerged
Ebro
Delta
plain
has
an
elevation
lower
than
50
cm
and
∼10%
of
the
delta
is
below
sea
level
(Ibá˜
nez
et
al.,
1997).
Thus,
50%
of
the
Ebro
Delta
is
vulnerable
to
flooding
impacts
and
permanent
submergence
of
both
marshes
and
rice
fields
(DMAiH,
2008;
Alvarado-Aguilar
et
al.,
2012).
This
is
not
a
problem
unique
to
the
Ebro
Delta
as
similar
systems
such
as
the
Ganges,
Mississippi,
Nile,
Rhone
and
Po
Deltas,
all
suffer
from
similar
sediment
deficits
(Syvitski
et
al.,
2009).
RSLR
impacts
on
worldwide
deltaic
rice
agriculture
would
have
effects
on
the
global
market
by
reduced
production
and
subsequent
increases
in
rice
prices,
which
may
have
important
implications
for
food
security
(Chen
et
al.,
2012).
Several
studies
in
the
Mediterranean
and
Asian
Deltas
(e.g.
Ebro,
Nile,
Ganges
and
Mekong
Deltas)
also
suggest
potential
population
displacements,
loss
of
biodiversity
and
cultural
heritage
(Syvitski
et
al.,
2009;
Day
et
al.,
2011).
One
proposed
measure
to
mitigate
deltaic
impacts
is
the
intro-
duction
of
riverine
sediments
into
marshes
as
a
means
of
correcting
the
sediment
deficit
(e.g.
Mississippi
Delta,
Rhone
Delta
and
Ebro
Delta)
(Ibá˜
nez
et
al.,
1997;
DeLaune
et
al.,
2003;
Day
et
al.,
2007).
The
reintroduced
river
water
would
also
provide
a
source
of
nutri-
ent
input,
which
theoretically
would
increase
marsh
elevation
by
stimulating
autogenic
organic
contribution
via
plant
growth
(McKee
and
Mendelssohn,
1989;
Day
et
al.,
2008).
Freshwater
inputs
also
reduce
soil
stressors
such
as
hyper-salinity,
anoxia
and
toxins
that
typically
inhibit
plant
growth
(Day
et
al.,
2011).
Several
studies
also
emphasize
that
the
organic
contributions
to
marsh
elevation
may
be
more
relevant
than
mineral
contributions
in
sediment-deficient
deltas
and
estuaries
(DeLaune
and
Pezeshki,
2003;
Blum
and
Christian,
2004;
Nyman
et
al.,
2006).
However,
more
data
are
required
to
directly
link
organic
contribution
to
marsh
elevation
(Cahoon
et
al.,
2006;
Day
et
al.,
2011;
Fagherazzi
et
al.,
2012).
In
the
Ebro
Delta
only
marshes
that
maintain
signifi-
cant
freshwater
and
sediment
inputs
will
likely
survive
predicted
RSLR
(Ibá˜
nez
et
al.,
2010).
So
it
is
important
to
understand
under
which
conditions
the
mineral
and
organic
contributions
to
marsh
elevation
can
be
optimized
as
a
key
restoration
objective.
For
exam-
ple,
the
use
of
Paspalum
Distichum
L.
during
the
establishment
of
restored
marshes
may
be
a
viable
restoration
practice
to
mitigate
RSLR
due
its
ability
to
capture
sediments,
it’s
tolerances
to
salinity,
water
logging,
and
dry
conditions,
fast
growth,
and
ability
to
repro-
duce
from
rhizomes,
stolons,
or
seeds
(Anderson
and
Ehringer,
2000;
Carr,
2010;
Wanyama
et
al.,
2012).
The
practice
of
converting
rice
fields
into
marshes
in
areas
of
low
elevation
has
been
proposed
in
the
Ebro
Delta
as
a
way
to
mitigate
the
effects
of
climate
change
(e.g.
by
elevation
gain
and
carbon
sequestration)
and
improve
the
water
quality
of
agricultural
runoff
(Ibá˜
nez
et
al.,
1997).
Recently,
several
public
efforts
have
restored
rice
field
land
to
freshwater
marshes
in
the
Ebro
Delta
using
river
irrigation
water
or
rice
field
drainage
water
to
feed
them
(MARM,
2006;
Ibá˜
nez
and
Bertolero,
2009).
However,
no
previous
experimental
studies
in
the
Ebro
Delta
and
other
Mediterranean
deltas
have
assessed
restoration
initiatives
regarding
abiotic
and
biotic
factors
controlling
vertical
accretion
and
elevation
change
to
keep
pace
with
sediment
deficit
and
RSLR.
In
this
study
we
hypothesized
that:
(1)
vertical
accretion
and
elevation
change
in
oligohaline
restored
marshes
under
low
sediment
availability
conditions
are
controlled
by
organic
con-
tributions
as
a
function
of
water
type
and
water
level,
and
(2)
oligohaline
restored
marshes
can
have
rates
of
vertical
accre-
tion
and
elevation
gain
higher
than
RSLR
under
low
sediment
availability
conditions.
To
test
these
hypotheses,
we
conducted
an
experimental
study
during
three
years
in
a
newly
established
experimentally
restored
marshes
consisting
of
72
experimental
units
(100
m2each).
Two
different
freshwater
water
input
types
(riverine
irrigation
water
and
rice
field
drainage
water)
and
three
water
levels
(10,
20
and
30
cm
deep)
were
used.
2.
Methods
2.1.
Experimental
design
We
carried
out
the
field
experiment
at
an
organic
rice
farm
located
in
the
southeast
of
the
Ebro
Delta
(Catalonia,
Spain)
(Fig.
1,

14 J.
Calvo-Cubero
et
al.
/
Ecological
Engineering
61 (2013) 12–
22
Fig.
2.
Design
of
the
experimentally
restored
marshes.
The
two
water
treatments
consisted
of
36
experimental
units
receiving
either
irrigation
or
drainage
water.
Each
treatment
included
the
water
level
factor
(10,
20
and
30
cm
depth)
divided
in
3
blocks
via
a
complete
randomized
block
design.
top).
We
established
the
experiment
between
an
active
organic
rice
field
to
the
West
and
an
old
restored
marsh
to
the
East
sep-
arated
by
a
bare
soil
strip
(Fig.
1,
bottom).
We
used
a
partly
nested
experimental
design
to
compare
plant
biomass,
vertical
accre-
tion
and
elevation
change
in
response
to
water
type
and
water
level
treatments.
Water
type
comprised
2
treatments:
riverine
irrigation
water
(IW)
and
rice
field
drainage
water
(DW)
applied
to
36
experimental
units
(EUs)
for
each
treatment.
The
water
level
factor
consisted
of
3
water
level
treatments
of
10,
20,
and
30
cm
in
depth.
These
water
level
treatments
were
applied
inside
each
water
type
using
a
complete
randomized
block
design
with
three
blocks.
Therefore,
each
water
type
included
3
blocks
and
where
each
block
included
12
EUs;
4
replicate
EUs
for
each
three
water
levels
(Fig.
2).
We
included
a
block
design
to
account
for
the
varia-
tion
in
our
experimental
units
from
plant
colonization
effects
from
the
active
rice
field
on
the
West
side
of
the
experiment.
The
cho-
sen
water
types
and
levels
were
based
on
readily
available
water
sources
and
a
realistic
range
of
potential
water
levels
found
in
rice
fields
and
marshes
of
the
Ebro
Delta.
Freshwater
from
both
IW
and
DW
favors
the
development
of
a
helophytic
marsh
dominated
by
a
Phragmites
australis
(Cav.)
Steudel,
Scirpus
maritimus
L.
and
Scirpus
litoralis
Schrad.,
since
the
study
area
is
located
in
an
old
freshwater
marsh
area
(abandoned
distributary),
which
was
transformed
to
rice
cultivation
in
the
previous
century
(Curcó
et
al.,
1995).
We
initiated
the
experiment
in
August
2009
and
ran
it
for
3
years
(Appendix
1).
The
hydroperiod
for
the
experiment
(seasonal
flood-
ing
and
draw
down
periods)
was
coincident
with
the
regional
rice
harvest/planting
regime.
During
the
first
year,
the
EUs
were
fully
flooded
from
August
to
December.
In
the
second
and
third
years,
the
EUs
were
flooded
from
June
to
December.
Targeted
water
levels
were
maintained
using
an
average
water
in-flow
rate
of
4.5
L
s−1.
The
EUs
were
individually
isolated
using
plastic
lined
wooden
walls
to
prevent/limit
water
loss.
2.2.
Water
and
soil
analysis
Physical
and
chemical
parameters
were
analyzed
monthly
from
2009
to
2011
for
both
water
type
inflows
(3
samples
per
month
and
water
type).
Dissolved
oxygen
(DO2),
temperature,
conduc-
tivity
and
pH
were
measured
using
an
YSI
556
multiprobe
(YSI

J.
Calvo-Cubero
et
al.
/
Ecological
Engineering
61 (2013) 12–
22 15
Incorporated,
Yellow
Springs,
OH,
USA).
Three
water
samples
per
month
were
collected
for
both
water
type
inflows
during
2010
and
2011
to
measure
total
suspended
solids
(TSS),
and
total
inor-
ganic
suspended
solids
(TISS)
according
to
the
UNE-EN
872
norm
(AENOR,
1996).
TSS
and
TISS
analysis
quantifies
both
total
and
mineral
sediment
inputs
in
both
water
types
that
may
cause
ver-
tical
accretion
and
elevation
change
response.
In
addition,
three
water
samples
per
month
for
both
water
type
inflows
during
2010
were
analyzed
for
the
following
nutrients:
total
phosphorus
(TP)
and
total
nitrogen
(TN);
inorganic
dissolved
nutrients;
phosphate
(P-PO43−),
nitrate
(N-NO3−)
and
ammonium
(N-NH4+),
following
standard
methods
(Grasshoff
et
al.,
1999).
Water
nutrient
analy-
sis
quantifies
nutrient
inputs
in
both
water
types
that
may
cause
a
plant
growth
response.
Furthermore,
the
same
physical
and
chemical
parameters
were
analyzed
from
a
subset
of
26
randomly
selected
EUs
(Fig.
2)
from
the
surface
and
sub-soil
(0.5
m
depth)
from
2009
to
2011
to
monitoring
water
characteristics
that
may
influence
plant
growth.
Superfi-
cial
soil
core
samples
(above
marker
horizons)
were
collected
in
May
2011
within
a
36
EUs
subset
to
analyze
soil
parameters
(i.e.
mineral
matter
content,
bulk
density
and
mineral
particle
size)
that
may
impact
vertical
accretion.
Samples
of
known
vol-
ume
were
weighted
to
determine
wet
weight
and
dried
to
a
constant
weight
at
60 ◦C.
Soil
bulk
density
was
calculated
from
these
data
(Page
et
al.,
1982).
Organic
matter
content
was
mea-
sured
by
loss-on-ignition
at
500 ◦C
during
12
h
and
mineral
matter
was
derived
from
organic
matter
percentage
following
Pont
et
al.
(2002).
The
determination
of
particle
size
distribution
in
mineral
soil
material
was
performed
using
wet
sieving
and
sedimenta-
tion
technique
(ISO11277,
2002).
The
following
particles
sizes
classes
were
measured:
clay
(d
<
2
␮m),
fine
silt
(2
␮m
<
d
<
16
␮m),
medium
silt
(16
␮m
<
d
<
45
␮m),
coarse
silt
(45
␮m
<
d
<
63
␮m)
and
sand
(63
␮m
<
d
<
2000
␮m).
2.3.
Aboveground
and
belowground
plant
biomass
Maximum
aboveground
biomass
(MAB)
or
peak
standing
crop
was
measured
to
obtain
an
estimate
of
plant
growth
(Cronk
and
Siobhan
Fennessy,
2001)
that
may
affect
vertical
accretion
and
ele-
vation
change.
Accordingly,
three
random
subsamples
of
0.25
m2
were
sampled
from
the
previously
identified
36
EUs
subset
in
the
last
growing
season
of
the
experiment
(August
2011)
following
a
direct
method
from
Schubauer
and
Hopkinson
(1984).
Plants
were
separated
by
species
and
dried
to
constant
weight
(at
60 ◦C).
Maximum
belowground
biomass
(MBB)
was
analyzed
to
obtain
information
of
root
growth
contribution
to
vertical
accretion
and
elevation
change.
MBB
was
sampled
in
the
last
dormant
season
of
the
experiment
(May
2012)
by
collecting
two
soil
core
subsamples
from
the
36
EUs
subset
(50
cm
long,
11.8
cm
internal
diameter).
To
improve
the
efficiency
of
core
extraction,
we
sealed
the
top
with
a
screw
top
cap
before
extracting
the
core
from
the
sed-
iment
(Schurman
and
Goedewaagen,
1971).
Each
soil
core
was
sectioned
and
washed
with
tap
water
through
1
mm
mesh
sieve
to
remove
inorganic
sediments.
The
plant
material
was
then
sealed
into
labeled
plastic
bags
and
stored
at
2–5 ◦C.
Each
thawed
sam-
ple
was
sorted
into
live
and
dead
roots
and
litter.
All
samples
were
dried
to
constant
mass
at
60 ◦C
and
weighed
(Curcó
et
al.,
2002).
2.4.
Vertical
accretion
and
elevation
change
Vertical
accretion
and
elevation
change
were
determined
using
marker
horizons
(Kaolinite)
(Cahoon
and
Turner,
1989)
and
surface
elevation
tables
(SETs),
respectively
(Cahoon
et
al.,
2002).
Marker
horizons
were
laid
upon
a
1
m2marsh
surface
in
the
center
of
the
36
EUs
subset
in
August
2009
and
two
random
subsamples
were
Table
1
Mean
and
standard
error
of
water
characteristics
of
both
water
type
inflows.
An
asterisk
indicates
a
significance
difference
(˛
=
0.05)
on
repeated
measures
ANOVA
results
between
water
type
inflows
for
water
nutrient
content
(TP,
P-PO4−,
TN,
N-
NO3−and
N-NH4+)
and
sediment
inputs
(TSS
and
TISS).
Parameter
Water
type
inflows
Irrigation
water
Drainage
water
Temperature
(◦C) 20.67
±
1.74 20.56
±
2.72
pH
8.59
±
0.20
7.73
±
0.25
DO2(mg
L−1)
7.75
±
0.94
5.55
±
0.75
Conductivity
(mS
cm−1)
1.22
±
0.10
1.48
±
0.18
TP
(mg
L−1)
0.08
±
0.01
0.18
±
0.02*
P-PO4−(mg
L−1)
0.10
±
0.04
0.13
±
0.04
TN
(mg
L−1) 1.18
±
0.04 1.05
±
0.21
N-NO3−(mg
L−1) 1.40
±
0.17* 0.54
±
0.13
N-NH4+(mg
L−1)
0.05
±
0.01
0.22
±
0.09*
TSS
(mg
L−1)
4.38
±
0.67
26.66
±
3.83*
TISS
(mg
L−1) 2.49
±
0.51 23.16
±
3.36*
sampled
at
the
end
of
the
experiment
(May
2012).
SET
stations
were
established
adjacent
to
the
marker
horizons
in
26
randomly
selected
EUs
(Fig.
2)
and
SETs
readings
were
taken
every
three
months
since
August
2009
until
May
2012.
In
this
study
we
use
ver-
tical
accretion
to
refer
to
surface
accretion
processes
due
to
mineral
sedimentation,
leaf
litter
deposition
and
root
growth.
We
refer
to
elevation
change
as
the
change
of
marsh
elevation
due
to
vertical
accretion,
subsurface
soil
expansion
by
root
growth
and
shallow
subsidence
by
compaction
and
decomposition.
2.5.
Statistical
analysis
A
repeated
measures
analysis
of
variance
(ANOVA)
of
fixed
effects
(Type
III
test)
was
performed
to
test
differences
in
sedi-
ment
input
(TSS
and
TOSS)
and
nutrient
content
(total
nutrients
and
dissolved
inorganic
nutrients)
on
both
water
type
inflows,
which
may
cause
a
different
response
of
vertical
accretion
and
elevation
change
in
the
experimentally
restored
marshes.
Multivariate
Prin-
cipal
Component
Analysis
(PCA)
was
carried
out
to
explore
the
relationships
between
abiotic
and
biotic
variables
with
vertical
accretion
and
elevation
change
and
identify
underlying
factors
that
control
them.
Kaiser–Meyer–Olkin’s
(KMO)
measure
of
sampling
and
Barlett’s
test
of
sphericity
were
used
to
assess
the
appropri-
ateness
and
adequacy
of
the
PCA
(McGarigal
et
al.,
2000).
A
partly
nested
ANOVA
of
fixed
effects
(Type
III
test)
was
used
to
test
differ-
ences
on
the
main
response
variables
(plant
biomass,
surface
soil
properties,
vertical
accretion
and
elevation
change)
among
water
types
and
water
levels
(Quinn
and
Keough,
2002).
Differences
in
response
variables
among
block
effects
were
also
analyzed
to
test
vertical
accretion
and
elevation
change
response
to
vegetation
col-
onization.
In
the
presence
of
significant
differences
(˛
<
0.05)
in
ANOVA
results,
pairwise
comparisons
were
made
with
the
Tukey
test.
All
statistical
analyses
were
performed
using
SPSS
software
version
20
(IBM,
2010).
3.
Results
3.1.
Water
characteristics
Both
water
type
inflows
were
characterized
as
oligohaline
and
oxygenated
waters
(Table
1).
The
ANOVA
indicated
DW
inflow
had
significantly
higher
TSS
(F1,4 =
59,247,
P
<
0.001),
TISS
(F1,4 =
58,285,
P
<
0.001),
N-NH4+(F1,4 =
73.8,
P
=
0.001)
and
TP
(F1,4 =
29.81,
P
=
0.05)
compared
to
IW
(Table
1).
IW
inflow
had
significantly
higher
N-NO3−(F1,4 =
51.67,
P
=
0.002)
with
no
signifi-
cant
differences
for
TN
(F1,4 =
0.32,
P
=
0.6)
and
P-PO43−(F1,4 =
5.26,
P
=
0.084)
(Table
1).
Experimentally
restored
marshes
had

16 J.
Calvo-Cubero
et
al.
/
Ecological
Engineering
61 (2013) 12–
22
Table
2
Mean
and
standard
error
of
surface
and
soil
water
characteristics
in
the
experimentally
restored
marshes
as
a
function
of
water
type
and
water
level
treatments.
Parameter
Irrigation
water
Drainage
water
10
cm
20
cm
30
cm
10
cm
20
cm
30
cm
Surface
water
temperature
(◦C)
Mean
20.00
19.56
19.51
20.26
19.98
19.65
S.E.
0.27
0.16
0.09
0.20
0.25
0.12
Soil
water
temperature
(◦C)
Mean
20.55
20.47
20.32
20.31
20.44
20.30
S.E.
0.05
0.07
0.17
0.23
0.20
0.15
Surface
water
pH Mean
7.59 7.78 7.72 7.71 7.74
7.65
S.E.
0.08 0.06 0.09
0.06
0.04
0.09
Soil
water
pH Mean
6.68
6.65
6.77
6.65
6.76
6.80
S.E.
0.07
0.05
0.12
0.08
0.04
0.03
Surface
water
DO2
(mg
L−1)
Mean
5.23
5.14
5.09
5.61
4.84
4.35
S.E.
0.59
0.35
0.43
0.54
0.26
0.77
Soil
water
DO2(mg
L−1)Mean
0.31
0.52
0.49
0.51
0.37
0.36
S.E.
0.03
0.09
0.08
0.11
0.07
0.04
Surface
water
conductivity
(mS
cm−1)
Mean
1.98
1.60
1.30
1.91
1.98
1.64
S.E.
0.44 0.12 0.06 0.15 0.16
0.04
Soil
water
conductivity
(mS
cm−1)
Mean
70.96
73.40
53.17
76.42
73.51
50.60
S.E.
1.39
3.49
8.20
2.54
1.54
4.26
oligohaline
surface
waters
and
hyperhaline
soil
waters
(0.5
m
depth)
due
to
saline
intrusion
characteristic
of
current
conditions
within
the
Ebro
Delta
(Table
2).
Experimentally
restored
marshes
had
oxygenated
surface
waters
and
sub-oxygenated
soil
waters
(Table
2).
3.2.
Relationships
between
forcing
and
response
variables
Principal
Component
Analysis
results
indicated
a
strong
rela-
tionship
of
vertical
accretion
with
surface
soil
properties
along
component
1
(Fig.
3,
top),
which
is
explained
by
water
type
variation
(Fig.
3,
bottom).
A
secondary
relationship
of
elevation
change
with
plant
biomass
appears
along
component
2
(Fig.
3,
top),
which
is
explained
by
block
variation
(Fig.
3,
bottom).
KMO’s
measure
of
sampling
adequacy
(0.60)
and
Bartlett’s
test
of
spheric-
ity
(P
<
0.0001)
indicated
the
appropriateness
of
the
PCA.
The
first
two
components
explained
a
total
of
43.13%
of
the
variance,
with
component
1
contributing
26.89%
and
component
2
contributing
16.24%.
An
oblimin
rotation
was
performed
to
aid
in
the
inter-
pretation
of
these
two
components
(McGarigal
et
al.,
2000).
The
PCA
rotation
solution
showed
a
number
of
medium
(±0.40)
and
high
(±0.60)
loadings
(correlations)
indicating
the
relationships
of
abiotic
and
biotic
variables
with
vertical
accretion
and
eleva-
tion
change
in
both
components
(Fig.
3,
top).
Along
component
1
appears
a
direct
relationship
between
vertical
accretion
(−0.56)
and
finer
grain
sizes;
namely
clay
(−0.75)
and
fine
silt
(−0.6).
These
variables
were
all
opposed
to
soil
bulk
density
(0.42)
and
coarser
grain
sizes;
medium
silt
(0.93),
coarse
silt
(0.94)
and
sand
(0.94).
Component
2
shows
a
direct
relationship
between
elevation
change
(−0.81),
MAB
(−0.65)
and
MBB
(−0.52).
These
variables
were
all
opposed
to
surface
water
pH
(0.75)
and
DO2(0.55),
soil
water
pH
(0.60)
and
soil
bulk
density
(0.57).
Scores
of
EUs
explained
two
main
underlying
factors
controlling
forcing
and
response
vari-
ables
relationships
(Fig.
3,
bottom).
The
first
factor
was
the
water
type
variation
along
component
1,
which
explained
vertical
accre-
tion
and
surface
soil
properties
relationships.
The
second
factor
was
the
block
variation
along
component
2,
which
explained
elevation
change,
plant
biomass
and
water
parameters
relationships.
3.3.
Plant
biomass
response
among
treatments
MAB
and
MBB
were
significantly
higher
in
blocks
closer
to
the
rice
field
due
to
P.
distichum
L
colonization,
and
MBB
also
responded
negatively
to
higher
water
levels
in
DW
treatment
(Fig.
4).
Plant
composition
of
the
EUs
corresponded
to
a
oligohaline
marsh
veg-
etation
dominated
by
S.
maritimus
L.
and
S.
litoralis
Schrad.,
with
high
colonization
by
the
weed
P.
distichum
as
a
consequence
of
the
proximity
of
a
rice
field
(Curcó,
2000).
There
were
no
signifi-
cant
differences
in
MAB
between
water
types
(F1,26 =
1.1,
P
=
0.309)
and
water
levels
(F2,26 =
2.1,
P
=
0.145)
nor
was
there
a
significant
block
effect
(F4,26 =
1.6,
P
=
0.192).
It
should
be
noted,
however,
that
the
MAB
data
displayed
an
increasing
trend
(higher
biomass)
from
block
1
to
block
3
(area
closest
to
an
existing
rice
field)
under
both
water
types
(Fig.
4,
top).
This
observed
trend
closely
tracked
the
biomass
of
P.
distichum
(Fig.
4,
center).
P.
distichum
was
the
most
significant
contributor
to
MAB
of
any
other
species
and
was
found
in
EUs
independent
of
the
IW
or
the
DW
treatments
(F1,26 =
4.1,
P
=
0.057)
water
level
(F2,26 =
0.57,
P
=
0.571)
or
block
(F4,26 =
2.1,
P
=
0.111).
MBB
was
significantly
different
among
water
levels
(F2,26 =
8.0,
P
=
0.002)
and
blocks
(F2,26 =
3.2,
P
=
0.028),
but
not
among
water
types
(F1,26 =
3.5,
P
=
0.072).
MBB
only
showed
sig-
nificant
differences
between
water
levels
in
the
DW
treatment
for
pairwise
comparisons
(Tukey),
where
MBB
was
significantly
higher
in
the
10
cm
water
level
than
in
the
30
cm
water
level
(P
=
0.014)
and
higher
than
the
20
cm
water
level
but
not
significant
(P
=
0.059)
(Fig.
4,
bottom).
A
significant
block
effect
was
observed
when
look-
ing
specifically
at
MBB
in
the
IW
treatment,
where
block
3
had
significantly
higher
MBB
than
blocks
2
and
1
(P
=
0.01
and
P
=
0.001,
respectively).
In
the
DW
treatment,
block
3
had
significantly
higher
MBB
than
block
1
(P
=
0.042)
for
pairwise
comparisons
(Tukey)
(Fig.
4,
bottom).
3.4.
Surface
soil
properties
response
among
treatments
Surface
soil
mineral
content
was
greater
in
DW
treatment,
as
well
as
bulk
density,
sand
and
silt
content.
Mineral
con-
tent
also
increased
in
both
water
type
treatments
with
the
higher
water
levels,
and
decreased
in
blocks
closer
to
rice
field
cultivation.
Mineral
content
responded
in
all
treatments
signif-
icantly:
among
water
types
(F1,28 =
4.7,
P
=
0.039),
water
levels
(F2,28 =
4.6,
P
=
0.019)
and
blocks
(F4,28 =
4.8,
P
=
0.004)
(Fig.
5,
top).
Pairwise
comparisons
between
water
levels
only
showed
sig-
nificant
differences
in
mineral
content
in
DW
treatment:
10
cm
with
20
cm
(P
=
0.04)
and
30
cm
(P
=
0.005).
Pairwise
compar-
isons
between
blocks
showed
significant
differences
in
mineral

J.
Calvo-Cubero
et
al.
/
Ecological
Engineering
61 (2013) 12–
22 17
Fig.
3.
Principal
Component
Analysis
of
the
forcing
and
response
variables.
Factor
loadings
of
the
variables
(top)
and
scores
of
experimental
units
classified
by
water
type
treatments
and
block
effects
(bottom).
content
in
both
water
type
treatments.
Block
3
had
significantly
lower
mineral
content
than
block
1
(P
=
0.02)
in
the
IW
treat-
ment.
Block
2
had
significantly
lower
mineral
content
than
block
1
(P
=
0.04).
There
were
significant
differences
in
bulk
density
between
water
types
(F1,28 =
12.9,
P
=
0.001)
but
not
between
water
levels
(F2,28 =
1.1,
P
=
0.365)
and
blocks
(F4,28 =
0.1,
P
=
0.979)
(Fig.
5,
center).
There
were
no
significant
differences
for
any
par-
ticle
size
content
between
water
types,
water
levels
and
blocks
(Fig.
5,
bottom):
sand
(F1,28 =
1.7,
P
=
0.205;
F2,28 =
1.5,
P
=
0.249;
and
F4,28 =
2.4,
P
=
0.071),
silt
(F1,28 =
2.4,
P
=
0.129;
F2,28 =
1.1,
P
=
0.338;
and
F4,28 =
2.9,
P
=
0.061)
and
clay
(F1,28 =
3.8,
P
=
0.062;
F2,28 =
0.8,
P
=
0.462;
and
F4,28 =
2.5,
P
=
0.082).
Two
general
trends
of
particle
size
content
appeared
among
treatments.
First,
both
sand
and
silt
increased
in
DW
treatment
inversely
to
clay
content.
Secondly,
silt
and
clay
content
varied
inversely
among
water
levels
and
blocks
in
both
water
type
treatments.
3.5.
Vertical
accretion
and
elevation
change
response
among
treatments
Vertical
accretion
showed
a
significant
increase
in
DW
treat-
ment
compared
to
IW
treatment
due
to
significantly
higher
mineral
sediment
input,
which
promoted
higher
surface
soil
mineral
con-
tent.
Elevation
change
appeared
to
be
affected
by
block
effects
due
to
P.
distichum
colonization
and
all
three
parameters
(ele-
vation
change,
MAB
of
P.
distichum
L
and
MBB)
increased
in
block
3
closest
to
rice
cultivation.
In
both
water
type
treatments,
mean
values
(±standard
error)
of
vertical
accretion
(11.5
±
0.8
and
15.5
±
0.6
mm
yr−1)
were
higher
than
those
for
elevation
change
(9.1
±
1.4
and
8.8
±
2.8
mm
yr−1).
ANOVA
results
showed
vertical
accretion
was
significantly
higher
in
the
DW
treatment
than
the
IW
treatment
(F1,21 =
9.5,
P
=
0.006)
(Fig.
6,
top).
Verti-
cal
accretion
did
not
show
significant
differences
among
water
levels
(F2,21 =
0.4,
P
=
0.681)
but
showed
significant
differences
among
blocks
(F4,21 =
2.8,
P
=
0.05).
Vertical
accretion
significantly
increased
on
blocks
closer
to
rice
field
cultivation
(block
3)
in
the
DW
treatment,
possibly
in
response
to
sediment
capture
by
the
combination
of
higher
sediment
input
and
substantially
higher
aboveground
biomass
of
P.
distichum.
Block
3
displayed
substan-
tially
higher
vertical
accretion
than
block
1
(P
=
0.075)
and
block
2
had
significantly
higher
vertical
accretion
rates
than
block
1
(P
=
0.05)
for
pairwise
comparisons
(Fig.
6,
top).
In
the
IW
treat-
ment,
vertical
accretion
significantly
changed
among
blocks,
where
block
3
had
significantly
lower
accretion
than
block
1
(P
=
0.013)
and
block
2
(P
=
0.034)
for
pairwise
comparisons
(Fig.
6,
top).
Elevation
change
did
not
show
significant
differences
among
water
types
(F1,16 =
0.1,
P
=
0.797)
and
water
levels
(F2,16 =
0.1,
P
=
0.868)
but
it
showed
significant
differences
among
blocks
(F4,16 =
4.2,
P
=
0.016)
(Fig.
6,
bottom).
Particularly,
block
3
in
both
water
type
treatments
had
higher
elevation
change
(11.3
±
2.3
and
17.8
±
2.9
mm
yr−1)
than
vertical
accretion
(8.3
±
1.3
and
15.1
±
1.3
mm
yr−1).
This
positive
difference
when
subtracting
ver-
tical
accretion
rates
from
elevation
change
rates
points
out
that
significantly
higher
belowground
biomass
in
block
3
also
con-
tributed
to
elevation
change
by
root
growth
due
to
high
weed
colonization
by
P.
distichum
from
rice
field
edge.
Block
3
of
the
DW
treatment
had
significantly
higher
rates
of
elevation
change
than
either
block
2
or
1
(P
=
0.043
and
P
=
0.042,
respectively)
for
pair-
wise
comparisons
(Fig.
6,
bottom).
In
the
IW
treatment,
elevation
change
did
not
show
any
significant
differences
between
blocks
although
it
displayed
an
increasing
trend
from
the
block
1
to
the
block
3
(Fig.
6,
bottom).
4.
Discussion
Marsh
development
and
stability
are
dependent
on
allogenic
(e.g.
flooding,
sediment
inputs)
and
autogenic
factors
(e.g.
plant
growth)
(Singer
et
al.,
1996;
Waller
et
al.,
1999).
Initial
stages
of
marsh
succession
are
usually
controlled
by
mineral
contributions
that
control
the
optimal
elevation
to
favor
vegetation
coloniza-
tion
and
development
(Morris
et
al.,
2002).
Mature
marshes
may
have
greater
organic
contribution
(e.g.
leaf
litter
deposition
and
root
growth)
because
a
more
extensive
plant
community
cover
is
developed
and
marsh
elevation
gain
causes
higher
isolation
from
marine
and
riverine
sediment
sources
(Brinson
et
al.,
1995;
Mitsch
and
Gosselink,
2007).
Under
a
sediment-deficit
scenario,
autogenic
organic
contributions
may
be
critical
to
marsh
stability
and
survival
to
support
the
balance
between
elevation
gain
and
RSLR
(DeLaune

18 J.
Calvo-Cubero
et
al.
/
Ecological
Engineering
61 (2013) 12–
22
Fig.
4.
Partly
nested
ANOVA
results
of
mean
(±SE)
plant
biomass
response
among
water
types,
water
levels
and
blocks.
Where
there
were
significant
main
effects
(˛
=
0.05)
on
ANOVA
results
among
water
levels
and
block
effects,
Tukey
pairwise
comparisons
were
tested
within
each
water
type
treatment
and
different
letters
denotes
significant
differences.
and
Pezeshki,
2003;
Blum
and
Christian,
2004;
Nyman
et
al.,
2006;
Langley
et
al.,
2009;
Kirwan
and
Guntenspergen,
2012).
In
our
study,
mineral
contributions
controlled
vertical
accretion
and
ele-
vation
change
in
the
establishment
of
oligohaline
restored
marshes
even
under
sediment-deficit
conditions.
Higher
mean
vertical
accretion
rates
(11.5
±
0.8
and
15.5
±
0.6
mm
yr−1)
than
elevation
change
(9.1
±
1.4
and
8.8
±
2.8
mm
yr−1)
in
both
water
type
treat-
ments
suggests
that
the
surface
accretion
processes
via
mineral
sedimentation
overall
controlled
marsh
elevation.
However,
sub-
soil
expansion
via
root
growth
also
caused
a
positive
elevation
gain,
especially
when
the
weed
P.
distichum
colonized
rapidly
and
densely
the
experimentally
restored
marshes.
The
observed
overall
mean
rates
of
vertical
accretion
and
elevation
change
also
support
the
hypothesis
that
oligohaline
restored
marshes
often
have
eleva-
tion
gains
higher
than
RSLR
in
the
Ebro
Delta,
at
least
during
the
initial
phase
of
marsh
establishment.
Vertical
accretion
responded
positively
to
higher
mineral
input
from
the
rice
field
drainage
water
but
vertical
accretion
was
not
affected
by
plant
colonization
among
blocks.
Our
results
show
a
significant
higher
surface
soil
mineral
content
in
DW
treatment
due

J.
Calvo-Cubero
et
al.
/
Ecological
Engineering
61 (2013) 12–
22 19
Fig.
5.
Partly
nested
ANOVA
results
of
mean
(±SE)
surface
soil
properties
response
among
water
types,
water
levels
and
blocks.
An
asterisk
indicates
significant
increase
(˛
=
0.05)
between
water
type
treatments.