Ecology, 96(11), 2015, pp. 2960–2972� 2015 by the Ecological Society of America
Nutrient enrichment intensifies hurricane impact in scrubmangrove ecosystems in the Indian River Lagoon, Florida, USA
ILKA C. FELLER,1,4 EMILY M. DANGREMOND,1 DONNA J. DEVLIN,2 CATHERINE E. LOVELOCK,3 C. EDWARD PROFFITT,2
AND WILFRID RODRIGUEZ1
1Smithsonian Environmental Research Center, Smithsonian Institution, 647 Contees Wharf Road, Edgewater, Maryland 21037 USA2Department of Biological Sciences, Charles E. Schmidt College of Science, Florida Atlantic University,c/o Harbor Branch Oceanographic Institution, 5775 Old Dixie Highway, Fort Pierce, Florida 34946 USA
3Centre for Marine Studies/School of Integrative Biology, University of Queensland, St Lucia, Queensland 4072 Australia
Abstract. Mangroves are an ecological assemblage of trees and shrubs adapted to grow inintertidal environments along tropical, subtropical, and warm temperate coasts. Despiterepeated demonstrations of their ecologic and economic value, multiple stressors includingnutrient over-enrichment threaten these and other coastal wetlands globally. These ecosystemswill be further stressed if tropical storm intensity and frequency increase in response to globalclimate changes. These stressors will likely interact, but the outcome of that interaction isuncertain. Here, we examined potential interaction between nutrient over-enrichment and theSeptember 2004 hurricanes. Hurricanes Frances and Jeanne made landfall along Florida’sIndian River Lagoon and caused extensive damage to a long-term fertilization experiment in amangrove forest, which previously revealed that productivity was nitrogen (N) limited acrossthe forest and, in particular, that N enrichment dramatically increased growth rates andaboveground biomass of stunted Avicennia germinans trees in the interior scrub zone. Duringthe hurricanes, these trees experienced significant defoliation with three to four times greaterreduction in leaf area index (LAI) than control trees. Over the long term, theþN scrub treestook four years to recover compared to two years for controls. In the adjacent fringe andtransition zones, LAI was reduced by .70%, but with no differences based on zone orfertilization treatment. Despite continued delayed mortality for at least five years after thestorms, LAI in the fringe and transition returned to pre-hurricane conditions in two years.Thus, nutrient over-enrichment of the coastal zone will increase the productivity of scrubmangroves, which dominate much of the mangrove landscape in Florida and the Caribbean;however, that benefit is offset by a decrease in their resistance and resilience to hurricanedamage that has the potential to destabilize the system.
Key words: coarse woody debris; delayed mortality; disturbance; Florida; growth rate; hurricane;Indian River Lagoon; leaf area index; mangrove forest; nutrient enrichment; resilience; resistance.
INTRODUCTION
Hurricanes are a frequent but aperiodic natural form
of disturbance along coastal regions around the world.
In Florida and throughout much of the Caribbean, these
episodic storms have direct and indirect effects on the
dynamics of mangrove forests, which dominate coastal
landscapes in this region (Lugo 2008, Smith et al. 2009,
Jiang et al. 2014). The initial damage to mangroves and
subsequent recovery time depend not only on hurricane
intensity (e.g., wind speed, forward speed, and storm
surge), but also on forest structure, including stand
density, tree height, and biomass accumulation (Roth
1992, Doyle and Girod 1997, Piou et al. 2006, Ross et al.
2006). In mangroves, such structural characteristics also
vary in undisturbed forests along environmental gradi-
ents in tidal elevation, salinity, and nutrient availability
(Lugo and Snedaker 1974, Sherman et al. 2001, Ball
2002, Feller et al. 2003a, b, Castaneda-Moya et al. 2013).
During recovery, these characteristics can be influenced
by huge allochthonous subsidies of nutrients delivered to
mangrove forests by hurricanes (Scatena et al. 1996,
Silver et al. 1996, Davis et al. 2004, Lovelock et al.
2011).
The mangrove forests that line the Atlantic coast of
Florida were hit by more than 100 hurricanes between
1851 and 2004 (Blake et al. 2005), but multiple storms
hitting the same area of coast in a given season are
extremely rare events (Liu 2007). However, two of the
four hurricanes that hit Florida in 2004 made landfall
within ;3 km of each other along the Indian River
Lagoon (IRL) and followed nearly identical paths across
the state. Hurricane Frances, a Category 2 storm (Saffir-
Simpson Hurricane Wind Scale [NHC 2012]), came
ashore near Stuart on 5 September (Bevan 2014) with
167 km/h maximum sustained winds. It was among the
Manuscript received 1 October 2014; revised 17 April 2015;accepted 21 April 2015; final version received 13 May 2015.Corresponding Editor: R. A. Dahlgren.
4 E-mail: [email protected]
2960
10 costliest storms that have hit the United States, with
much of the damage attributed to its slow movement
(data available online).5 Three weeks later, on 25
September, Hurricane Jeanne, a Category 3 storm,made landfall near Stuart with maximum sustained
winds near 195 km/h. In the year following these storms,
several studies examined the damage to Florida’s coastal
ecosystems (e.g., Lapointe et al. 2006, Paperno et al.
2006, Ridler et al. 2006, Sallenger et al. 2006, Steward etal. 2006, Tomasko et al. 2006, Wang and Horwitz 2007,
Dix et al. 2008, Carlson et al. 2010). These studies found
spatial variation in the amount of physical damage
sustained by coastal environments and the resident flora
and fauna (Greening et al. 2006). In mangrove forests,the physical damage and level of defoliation varied with
distance from the eye walls of the hurricanes (Milbrandt
et al. 2006, Tomasko et al. 2006). Mangrove stands with
larger mangrove trees and lower densities sustained
greater damage from hurricane winds (Vogt et al. 2011).In addition, Proffitt et al. (2006) found that, in the first
year following the storms, mangrove reproduction was
reduced by an order of magnitude. However, no studies
have subsequently reported on mangrove recovery over
the longer term after hurricanes Frances and Jeanne.
Our specific objective was to investigate how differ-ences in nutrient availability affected damage levels and
long-term recovery in a mangrove forest along the IRL
following hurricanes Frances and Jeanne. These storms
passed directly over an established fertilization experi-
ment in which we had measured forest structure andtracked mangrove response to nutrient enrichment. The
initial purpose of this experiment was to examine what
types of changes occur within mangrove ecosystems in
response to nutrient over-enrichment, which remains
one of the major global threats to coastal wetlands (e.g.,Cloern 2001, Boesch 2002, Deegan et al. 2012). Results
from our experiments have previously shown that
primary production in mangrove forests along the IRL
was nitrogen (N) limited and that N enrichment
significantly altered plant growth, photosynthesis rates,tree height, leaf area index, foliar nutrient dynamics,
herbivory, decomposition, and competitive interactions
among mangrove species (Table 1; Feller et al. 2003b,
2010, 2013, Lovelock and Feller 2003). However, these
studies also showed that the size of the response to Nenrichment varied by species and zone along a tidal
elevation gradient. In the present study, we used that
fertilization experiment to determine how increased
nutrient availability and forest zone affected the
susceptibility of mangrove to hurricane damage (resis-tance) and their rate of recovery (resilience) after the
storm.
The 2004 hurricanes struck our study site eight years
after we had initiated the fertilization experiment,
providing an additional opportunity to assess the role
of disturbance in modulating growth, canopy dynamics,
and recovery. Because N was found to be the elementlimiting mangrove growth at our site, we hypothesized
that the N-fertilized (þN) trees would sustain greaterdamage than control trees as a result of decreasedresistance to hurricane winds due to larger, denser
canopies, but that they would recover faster as a resultof increased resilience associated with more rapid
growth rates (Herbert et al. 1999). We further hypoth-esized that damage and recovery would also vary byspecies and zone. Although some studies found that
susceptibility to hurricane damage varies with speciesand size of mangrove (Baldwin et al. 1995, Imbert et al.
1996, McCoy et al. 1996, Sherman et al. 2001, Vogt et al.2011), others found similar levels of damage amongmangrove species and among size classes within species
(Odum et al. 1982, Gresham et al. 1991, Smith et al.1994, Sherman et al. 2001). Because mangrove forests
are often characterized by low floristic diversity, severalstudies have predicted that the recovery trajectory isrelatively simple such that biological diversity, forest
structure, ecosystem processes, and ecological functionswill completely regain their former states followinghurricane disturbance (Lugo et al. 1976, Ogden 1992,
Smith et al. 1994). Although mangrove species diversityis low in the Atlantic-Caribbean region, the physiogno-
my of mangrove forest types is diverse, ranging fromscrub stands ,1.5 m tall to .30 m tall in riverinesystems as a function of hydrogeomorphology (Rivera-
Monroy et al. 2004). Long-term studies followingcumulative damage and recovery from multiple storms
across south Florida found that hurricane damage wasmost significantly related to forest type such that basinmangroves suffered more damage than riverine or island
mangroves while other studies observed that scrubforests showed little if any damage following even
catastrophic hurricanes (Lugo 1997, Ross et al. 2006,Smith et al. 2009). Ross et al. (2006) proposed that suchdifferences in mangrove forest type and productivity
along hydrogeomorphic gradients may arise throughcompetitive interactions that develop during recovery
following hurricane damage, suggesting a long-term
TABLE 1. Summary of growth responses of fertilized Avicenniagerminans (scrub and transition zones) and Rhizophoramangle (fringe zone) trees at Mosquito Impoundment 23(MI 23), Avalon State Park, North Hutchinson Island,Florida, USA, prior to the September 2004 hurricanes.
NutrientTreatment
Growth (cm/month)
Scrub Transition Fringe
Control 0.13 6 0.03 0.42 6 0.08 0.48 6 0.11þN 2.60 6 0.22 4.18 6 2.34 2.06 6 0.80þP 0.41 6 0.11 0.57 6 0.19 0.32 6 0.08
Notes: Growth is measured as shoot elongation in responseto nutrient enrichment with nitrogen (þN) as urea, phosphorus(þP) as triple superphosphate, or controls (no fertilizer added).Measurements are based on annual increases to individualshoots after two years of treatment (data from Feller et al.[2003b]).
5 http://www.hurricanediscovery.org/history/storms/2000s/frances/
November 2015 2961N ENRICHMENT INCREASES HURRICANE IMPACT
imprint of hurricanes on mangrove communities of the
region.
MATERIALS AND METHODS
Study site
Our study was conducted at Mosquito Impoundment
23 (MI 23), a 122-ha stand of coastal mangroves located
in the Avalon State Recreation Area on the lagoonal
side of North Hutchinson Island, St. Lucie County,
Florida, USA (278330 N, 808200 W; see Plate 1). This
impoundment was constructed in 1966 to control
populations of nuisance insects, and was maintained
until 1974 when its dike was breached (Carlson et al.
1983, Rey et al. 1986, 1990, 1992, Rey and Kain 1991).
Hydrological connection between MI 23 and the IRL is
maintained through the breach and three culverts.
Vegetation at this site is characterized by a tree-height
gradient, perpendicular to the shoreline (Fig. 1).
Rhizophora mangle L. (red mangrove) trees, 3–5 m tall,
dominate to form a dense fringe along the canal at the
seaward edge of MI 23. Landward of the fringe, the
transition zone is a mix of slender Avicennia germinans
L. (black mangrove) and Laguncularia racemosa
Gaertn.f. (white mangrove) trees of decreasing height,
from ,3.0 to .1.5 m tall. Landward of the transition
zone, the scrub zone in the interior of MI 23, is
dominated by stunted, ,1.5 m tall, old-growth A.
germinans trees characterized by low stature, infrequent
branching, and small internodal lengths (Lugo and
Snedaker 1974, Lugo 1997), which distinguishes them
from saplings, which also have low stature but are
relatively young, vigorously growing plants as indicatedby long internodes and frequent axial branching (Feller
and Mathis 1997). At MI 23, scrub A. germinans cover
over 60% of the mangrove landscape (Feller et al.
2003b).
Experimental design
The experimental design was a randomized complete
block with a factorial treatment arrangement. Transects
along the tree-height gradient were replicated in threeblocks, 100–150 m apart, along the western side of MI
23. In each block, three transects, 25–50 m long and ;10
m apart, were oriented perpendicular to the shoreline
and subdivided into fringe, transition, and scrub zones
(Feller et al. 2003b). Two species were targeted for
fertilizer treatment: R. mangle in the fringe zone and A.
germinans in the transition and scrub zones. In each
block, three replicate trees were selected within each
zone. Trees were fertilized with 300 g of N fertilizer asurea (45:0:0), or P fertilizer as P2O5 (0:45:0), as described
in Feller (1995). Nutrient treatment for each transect
within each block was assigned randomly. Nine trees per
nutrient treatment per zone in each block for a total of
81 trees (3 nutrient treatments 3 3 zones 3 3 blocks 3 3
replicate trees per zone) were treated and measured
before and after the hurricanes at six-month intervals
from 1997 to 2011. Doses (150 g) of fertilizer were
placed in small holes (3 cm diameter 3 30 cm deep),
FIG. 1. Profile of the tree-height gradient across the mangrove forest at Mosquito Impoundment 23 (MI 23), Avalon StatePark, North Hutchinson Island, St. Lucie County, Florida, USA based on point-centered quarter transects (N ¼ 4). Data fromFeller et al. (2003b).
ILKA C. FELLER ET AL.2962 Ecology, Vol. 96, No. 11
cored into the substrate beneath the drip line on
opposing sides of the canopy of each tree, and sealed.
We used this method rather than surface broadcasting to
minimize our nutrient footprint and to assure that the
fertilizers were available to tree roots rather than lost in
tidal flushing. For controls, holes were cored and sealed,
but no fertilizer was added.
Plant growth.—As a bioassay of the effects of storm
damage and nutrient treatment on plant growth, we
tracked the responses of five, initially unbranched,
shoots (first order) in sunlit positions in the outer part
of the canopy of each tree for two years following the
September 2004 hurricanes. In June 2007, about three
years after Hurricanes Frances and Jeanne, we harvested
the tagged shoots and measured shoot length and
biomass, number of nodes, and leaf area and biomass.
These data were used to determine growth rates in the
fertilization experiment.
Damage and recovery.—Damage and recovery of the
mangrove canopy structure were quantified with leaf
area index (LAI) using a gap fraction method.
Hemispherical photographs were taken with a Nikon
Coolpix digital camera (model 995; Nikon, Tokyo,
Japan) fitted with a fisheye lens and in the same
position under the canopy of each of the fertilized trees
in the experiment (Fig. 2). Images were processed using
Hemiview Canopy Analysis Software (version 2.1;
Delta-T Devices, Cambridge, UK). Pre-hurricane
LAI values were measured based on hemispherical
photographs taken in April 2003. To quantify the
damage caused by each storm, additional hemispher-
ical photographs were taken one week after Hurricane
Frances and one week after Hurricane Jeanne. To
track recovery, this process was repeated at ;6–12
month intervals thereafter for six years. At each time,
we also measured tree height and porewater salinity.
Porewater was collected from a depth of 15 cm at the
base of each tree with a probe attached to a suction
device.
Woody debris.—The input of biomass of downed and
standing woody debris as a result of the September 2004
hurricanes was quantified using 4 3 4 m plots in fringe,
transition, and scrub (N ¼ 3) zones at MI 23. The nine
plots were initially sampled in April 2005, approximately
seven months after hurricanes Frances and Jeanne. To
track changes in the amount of dead wood over the
subsequent five-year period, these plots were resampled
in June 2007 and November 2010. Each time, all pieces
of dead wood on the ground (downed) and in the
canopy (standing) were removed and sorted. After
sorting to species, samples were sorted by size class into
fine woody debris (FWD), which included twigs and
small branches (,1 cm diameter), pneumatophores,
seedlings, and coarse woody debris (CWD), which
included larger branches, boles, and prop roots (.1
cm). These samples were further sorted into three decay
classes (i.e., sound, intermediate, rotten), following the
methods described in Krauss et al. (2005). The ‘‘sound’’
class was recently killed wood with intact bark and was
indicative of dead wood caused by the September 2004
hurricanes. Both the intermediate and rotten classes had
been dead longer than one year. For the intermediate
class, the wood was still firm but was missing its bark.
For the rotten class, the wood was soft and disintegrat-
ing. All samples were then weighed, and a subsample
from each sorted component was oven-dried to deter-
mine a conversion factor for estimating dry weight per
sample, which was used for estimating biomass of
woody debris (g/m2) by species and size/decay classes.
Statistics.—Our data from the fertilization experiment
were grouped by nutrient treatment (Control, þN, þP)and zone (fringe, scrub) at three replicate sites at MI 23.
We used the GLM procedure repeated-measures anal-
ysis of variance (ANOVA) univariate tests of hypotheses
for within-subject effects on LAI measurements repeated
at 6–12 month intervals to look for possible differences
in recovery patterns over a seven-year period. We used
two-way ANOVAs for each response variable. Tukey’s
FIG. 2. Examples of the hurricane damage to the mangrove canopy at the same position in the transition zone one week afterHurricane Frances and Hurricane Jeanne and after six years of recovery.
November 2015 2963N ENRICHMENT INCREASES HURRICANE IMPACT
Honestly Significant Difference (HSD) tests were
applied to examine pairwise differences within and
among the treatment levels. Response variables were
either log- or square-root transformed prior to ANOVA
analyses to respect variance homogeneity and normality.
Repeated-measures ANOVAs were done with R soft-
ware 2.9.0 (R Development Core Team 2009); all other
analyses were done in SYSTAT 11 (Systat 2004).
RESULTS
Plant growth.—Shoot elongation rates (cm/month),
which were measured four times between June 2000 and
January 2007, were used to compare plant growth before
and after the September 2004 hurricanes. The hurricanes
had a significant effect on shoot elongation over that
time period (repeated-measures ANOVA, F4,48 ¼ 33.56,
P , 0.001; Fig. 3). Before the hurricanes, the þNtreatment caused a significant increase in growth relative
to both control andþP treatment in each zone (repeated-
measures ANOVA, F2,51¼ 20.291, P , 0.001), with the
highest rates in þN trees in the transition zone trees
(repeated-measures ANOVA, F4,51¼ 4.403, P , 0.005).
For almost three years after the hurricanes, growth rates
in all zones were significantly lower than they were
before the storms, with no significant difference among
nutrient treatment levels (repeated-measures ANOVA,
F16, 147¼ 2.537, P , 0.002). Although there was a slight
but non-significant increase in growth for theþN trees in
the transition zone, values were still less than half their
pre-hurricane levels.
Leaf area index.—LAI values for individual trees in
the fertilization experiment at MI 23, which were made
11 times beginning with pre-hurricane conditions in
April 2003 through November 2011, provide a measure
of the defoliation caused by each of the storms and the
subsequent recovery of the canopy by nutrient treatment
and zone (Fig. 4). Over this time period, LAI changed
significantly (Table 2). That change in LAI over time
FIG. 3. Changes in growth rates based on shoot elongation (cm/month) of fertilized trees in scrub, transition, and fringe zonesat MI 23 before and after the September 2004 hurricanes.
FIG. 4. Damage and recovery of the mangrove canopy structure following the September 2004 hurricanes were determined withleaf area index (LAI), based analysis of hemispherical photographs of the canopy using Hemiview Canopy Analysis Software. Alldata are from 1 January.
ILKA C. FELLER ET AL.2964 Ecology, Vol. 96, No. 11
was significantly affected by both nutrient treatment and
zone, with a significant zone 3 nutrient treatment
interaction. Although Hurricane Jeanne was the stron-
ger of the two storms, most of the reduction in LAI was
caused during Hurricane Frances, the first of the two
storms (Fig. 4). Pre-hurricane LAI levels from April
2003 showed a significant relationship with LAI
measured immediately after Hurricane Frances (r2 ¼0.45, F1,78 ¼ 64.50, P , 0.0001), but not with LAI
measured immediately after Hurricane Jeanne (P .
0.05). But, the magnitude of the effect was not uniform.
Before the storms in 2003, the effect of nutrient
treatment differed significantly by zone with a statisti-
cally significant interaction between nutrient treatment
and zone (ANOVA, F4,71 ¼ 10.54, P , 0.001). Pairwise
contrasts indicated that the LAI values for control trees
were significantly lower in the scrub zone than in the
fringe (Tukey’s HSD test, P , 0.001) or transition
(Tukey’s HSD test, P , 0.001) zones. Also within the
scrub zone,þN trees had significantly greater LAI levels
than control (Tukey’s HSD test, P , 0.001) or þP(Tukey’s HSD test, P , 0.001) trees. However,
following the September 2004 hurricanes, LAI was
reduced to similar levels in all zones.
In the scrub zone, LAI levels were reduced by 46.5%6 4.1% (mean 6 SE) inþN trees compared to 32.1% 6
10.5% in controls and 26.9% 6 16.6% inþP trees. There
was a significant relationship between pre- and post-
Frances LAI levels for scrub trees (r2 ¼ 0.59, F1,26 ¼36.58, P , 0.0001), but there were no differences
between pre- and post-Jeanne LAI or between post-
Frances and post-Jeanne LAI. After Hurricane Jeanne,
LAI levels were reduced by 71.2% 6 2.4% in the þNtrees compared to 28.9% 6 12.2% in control trees and
16.2% 6 15.8% in theþP trees. Based on measurements
made over a seven-year period, LAI levels recovered to
pre-hurricane levels by January 2007 for Controls and
þP trees but not until August 2008 forþN trees (Fig. 4).
Subsequent to Aug. 2008, LAI continued to increase for
all nutrient treatments in the scrub zone to exceed pre-
hurricane levels.
In the transition and fringe zones, all nutrient
treatments sustained similar amounts of defoliation that
caused a reduction in LAI by 78.6% 6 1.0%, 80.0% 6
2.6%, and 83.3% 6 2.6% in control, þN, and þP trees,
respectively. The LAI for our experimental trees in the
both of these zones returned to pre-hurricane levels by
January 2007, with no significant differences based on
nutrient treatment (Fig. 4). Additionally, the fringe and
transition trees subsequently experienced large reduc-
tions in LAI in 2008 and in 2010 when there were no
storms, which may have been a result of delayed
mortality in the canopy or persistent drought conditions
in the region during that time period (SFWMD 2009,
Abtew et al. 2010). During these droughts, porewater
salinities increased significantly in all three zones
(repeated-measures ANOVA; F10, 120 ¼ 7.068, P ¼0.000). The scrub zone was consistently hypersaline,
but values fluctuated significantly over time (repeated-
measures ANOVA; F4,44 ¼ 13.2, P , 0.001), with
salinity lowest (47.2 6 4.3 ppt) in April 2005 and highest
(65.4 6 3.9 ppt) in August 2008. Salinity also fluctuated
significantly in the fringe (repeated-measures ANOVA;
F4,44 ¼ 34.8, P , 0.001), with values lowest (28.2 6 3.3
ppt) in October 2004 and the highest (45.2 6 1.1 ppt) in
January 2010. In the transition zone, salinity levels were
intermediate between scrub and fringe level, but also
fluctuated significantly with time (repeated-measures
ANOVA; F4,44 ¼ 21.8, P , 0.001), with values lowest
(39.4 6 1.8 ppt) in October 2004 and the highest (49.6 6
1.1 ppt) in January 2010.
Woody debris.—In April 2005, approximately seven
months after the hurricanes, there was relatively little
downed or standing CWD and FWD in the scrub,
transition, or fringe zones at MI 23 (Table 3). The
woody debris present at that time was primarily in the
sound-wood decay class, which indicated that it had
died recently as a result of the hurricanes (Table 3).
Thus, approximately 60% of the standing dead wood
biomass in the fringe and 80% in the scrub and
transition zones were attributable to the hurricanes.
For the downed wood, approximately 50% of the dead
wood was a direct result of the hurricanes. There was
considerable variability among the plots in the transition
and fringe zones with no significant differences in the
amounts of dead wood in those parts of the mangrove
TABLE 2. Summary of results from the GLM procedure repeated-measures analysis of varianceunivariate tests of hypotheses for within-subject effects performed on leaf area index (LAI)measured over a seven-year period at fertilized Avicennia germinans (scrub and transition zones)and Rhizophora mangle (fringe zone) trees at MI 23.
Source df Type III SS MS F P
Adjusted P
G-G H-F-L
LAI 12 184.05 15.34 281.85 ,0.0001 ,0.0001 ,.0001LAI 3 treatment 24 3.80 0.16 2.91 ,0.0001 0.0006 0.0004LAI 3 zone 24 44.88 1.87 34.36 ,0.0001 ,0.0001 ,.0001LAI 3 treatment 3 zone 48 5.17 0.11 1.98 0.0001 0.0039 0.0026Error (LAI) 852 46.36 0.05G-G 0.5102H-F-L 0.5636
Note: G-G, Greenhouse-Geisser epsilon; H-F-L, Huynh-Feldt-Lecoutre epsilon.
November 2015 2965N ENRICHMENT INCREASES HURRICANE IMPACT
forest. When we resampled the plots two years later in
June 2007, approximately three years after the 2004
hurricanes, there was a large increase in the amount of
recently dead and standing CWD in the fringe zone as a
result of delayed mortality of R. mangle (Fig. 5). In
August 2010, we found a pulse of recently dead R.
mangle twigs and smaller parts of the canopy that had
subsequently fallen to the ground to become part of the
downed FWD in the fringe zone. In addition, delayed
mortality in L. racemosa trees also became evident in
August 2010 and caused a significant increase in the
amount of standing FWD in the transition zone. Unlike
the fringe and transitions zones, the scrub trees suffered
minor losses to the woody portions of their canopies
over the course of our study with no evidence of delayed
mortality.
DISCUSSION
Our results show that, while increasing nutrient
availability enhanced the productivity of scrub man-
groves in the IRL, it also increased their susceptibility to
hurricane damage and decreased their ability to recover
following such disturbances. These findings suggest that
nutrient over-enrichment of the coastal zone may
amplify the impact of hurricanes on mangroves and
lower their resilience to storms and other disturbances
particularly in scrub mangrove forests. In this forest
type, often referred to as dwarf forests, individual trees
are ,1.5 m tall but are relatively old (Lugo and
Snedaker 1974, Lugo 1997). Fertilization experiments
in several such old-growth scrub mangrove stands have
demonstrated that they are nutrient limited and that
addition of the limiting nutrient dramatically increased
their productivity and converted their stunted physiog-
nomy to resemble more vigorously growing trees in
fringing mangrove forests (Feller 1995, Feller et al.
2003a, b, Lovelock et al. 2006, 2009, Martin et al. 2010).
Although the aboveground biomass in scrub mangroves
is low compared to other types of mangrove forests,
these low-stature forests are widespread and have been
documented to dominate much of the mangrove
landscape not only at our study site along the IRL
(Feller et al. 2003b), but also in south Florida (Ross et
al. 2001, 2006, Simard et al. 2006), Belize (Murray et al.
2003, Rodriguez and Feller 2004), Panama (Lovelock et
al. 2005), and throughout the Caribbean (Rivera-
Monroy et al. 2004).
Several studies have observed that their low stature
allows scrub mangroves to escape hurricanes with
negligible defoliation (Craighead 1971, Roth 1992, Lugo
1997, Ross et al. 2006). Experimental studies have
shown that scrub mangroves are extremely sensitive to
increased nutrient availability and respond to nutrient
enrichment with dramatic increases in stature as a result
of increased shoot elongation, branching, and leaf
production (Feller 1995, Feller et al. 2003a, b, 2013,
Lovelock et al. 2004, 2006, Reef et al. 2010), which could
make them more vulnerable to the wind damage
associated with intense tropical storms. Similar to
observations in Australia by Lovelock et al. (2009), the
benefits of increased mangrove biomass in response to
nutrient over-enrichment of the coastal zone will be
offset by lower resilience of mangroves to increased
disturbance.
Despite the potential importance of interactions
between anthropogenic and natural disturbances for
many ecosystems (e.g., Mallin et al. 2002), few
experiments have evaluated the impacts of nutrient
loading on the level of damaged sustained during
disturbance events (Herbert et al. 1999, Lovelock et al.
2011). It is not known how increased nutrients from
anthropogenic sources might affect damage and recov-
ery of mangroves from hurricanes. In a fertilization
experiment in Hawaii, Herbert et al. (1999) found that
experimental nutrient additions had significant effects
on hurricane damage and recovery and that increased
availability of P, the limiting nutrient in a montane
Metrosideros polymorpha forest, resulted in higher
damage and/or decreased resistance coupled with faster
recovery, which indicated an increase in resilience.
TABLE 3. Summary of the biomass of woody debris in the mangrove forest at MI 23 collected inNovember 2005 in 4 3 4 m plots in fringe, transition, and scrub zones (N ¼ 3).
Zone andcanopy position
Sound Intermediate and rotten
FWD (g/m2) CWD (g/m2) FWD (g/m2) CWD (g/m2)
Fringe
Standing 23.0 6 3.1 36.3 6 26.2 1.6 6 0.8 16.6 6 10.8Downed 40.0 6 5.4 42.7 6 14.3 0.1 6 0.0 43.4 6 19.5
Transition
Standing 52.5 6 23.5 167.5 6 115.2 4.2 6 4.1 28.4 6 14.3Downed 39.5 6 5.2 59.4 6 9.1 0.7 6 0.6 136.4 6 98.0
Scrub
Standing 45.2 6 5.6 24.6 6 4.3 1.2 6 1.1 12.9 6 10.7Downed 20.0 6 5.2 3.2 6 1.4 2.3 6 2.0 12.7 6 8.8
Notes: Dead wood on the ground (downed) and in the canopy (standing) was collected andsorted into fine woody debris (FWD) and coarse woody debris (CWD). Samples were categorizedbased on decay class: sound, intermediate, rotten.
ILKA C. FELLER ET AL.2966 Ecology, Vol. 96, No. 11
Contrary to predictions that resilience and resistance are
inversely related (Hollings 1973, Attiwill 1994), LAI and
growth measurements in our fertilization experiment in
a mangrove forest along the Indian River Lagoon
following hurricanes Frances and Jeanne in September
2004 showed that increased availability of N, the
limiting nutrient at our site, resulted in more structural
damage to the canopy and slower recovery to pre-
hurricane levels. It took four years for the þN trees to
return to pre-hurricane LAI levels compared to two
years for the control trees. These results, which were
only evident in the scrub zone, partially supported our
predictions that the þN trees would sustain more
damage than control trees as a result of decreased
resistance to hurricane winds, but would recover faster
as a result of increased resilience.
For the mangrove forest at MI 23, earlier studies
experimentally identified N as the growth-limiting
nutrient (Feller et al. 2003b, 2013, Lovelock and Feller
2003) and determined that the addition of N fertilizer
(þN) caused a significant increase in plant growth while
the additionþP had no effect on growth. The September
2004 hurricanes resulted in a significant decrease in
growth, measured as shoot elongation, in all zones and
nutrient treatments (Fig. 3). Although the experimental
addition of fertilizer continued, that pattern of decreased
growth persisted over a three-year period following the
hurricanes. This contrasts with other studies that have
found a subsidy effect for mangrove productivity as a
result of increased delivery of nutrients by a storm
(Davis et al. 2004, Lovelock et al. 2011). In contrast to a
similar fertilization experiment in a mangrove forest in
Western Australia, Lovelock et al. (2011) found an
increase in growth in the year following the 2008
Cyclone Pancho, attributed to nutrient and freshwater
subsidies delivered by the storm. However, there was no
canopy damage associated with Cyclone Pancho, but
rather mangroves were exposed to enriched flood
waters.
FIG. 5. Coarse (CWD) and fine (FWD) woody debris from the ground (downed) and from the canopy (standing) in 4 3 4 mplots (N ¼ 3) in the fringe, transition, and zones at MI 23 collected three times over a six-year period.
November 2015 2967N ENRICHMENT INCREASES HURRICANE IMPACT
Although nutrient subsidies enhance primary produc-
tivity of coastal wetlands, nutrient-enrichment experi-
ments have shown that increased nutrients are
detrimental to the long-term maintenance of mangroves
and salt marshes (Verhoeven et al. 2006, Reef et al. 2010,
Deegan et al. 2012). For example, fringing mangroves in
Florida and Belize experienced increased belowground
decomposition, subsidence and loss of habitat stability
in response to nutrient enrichment (Feller et al. 2003b,
McKee et al. 2007). In mangrove fertilization experi-
ments in Australia, nutrient enrichment increased the
sensitivity to hypersalinity and drought in Avicennia
marina trees, which resulted in increased mortality rates
(Lovelock et al. 2009). Several studies found that
competitive interactions, plant-herbivore dynamics,
habitat suitability, and community structure were
altered in response to fertilization (Feller 1995, Lovelock
and Feller 2003, Feller and Chamberlain 2007, Minch-
inton and McKenzie 2008, Feller et al. 2013). Feller et
al. (1999) found that nutrient enrichment dramatically
reduced nutrient use efficiency and the ability of
mangroves to act as a sink for nutrients. In other
coastal wetlands, nutrient enrichment is also widely
recognized as a major driver of habitat loss, eroding
shorelines and reducing carbon storage (e.g., Morris and
Bradley 1999, Cloern 2001, Bertness et al. 2002, 2008,
Turner et al. 2009, Deegan et al. 2012, Turner 2012,
Morris et al. 2013).
Hurricane effects on mangrove canopy.—LAI values
for the þN trees in the scrub zone, which exhibited
vigorous growth in response to the experimental
treatment (Feller et al. 2003b), were reduced ;50% by
Hurricane Frances and another ;25% by Hurricane
Jeanne. Although nutrient availability had no detectable
effect on damage levels in the fringe and transition
zones, the þN trees in the scrub zone sustained higher
PLATE 1. A stand of black and white mangrove trees in Mosquito Impoundment 23 in Avalon State Recreation Area, St. LucieCounty, Florida (USA) on 16 October 2004; three weeks after Hurricane Jeanne and six weeks after Hurricane Frances. The stormsstripped most of the leaves and branches from the mangrove canopy. Photo credit: I. C. Feller.
ILKA C. FELLER ET AL.2968 Ecology, Vol. 96, No. 11
levels of damage than P-fertilized or control trees. Based
on our repeated measurements of LAI and growth, the
effects of these hurricanes continued to be evident in the
mangroves for at least three to four years after the
passage of these storms.
Results from our study at MI 23 showed that the
damage levels to the mangrove canopy caused by the
September 2004 hurricanes (hurricanes Frances and
Jeanne) along the IRL varied by structural and species
differences among zones and nutrient availability (Fig.
4). Contrary to other studies that found hurricane
damage in coastal forests was similar among mangrove
species and among size classes within species (Odum et
al. 1982, Gresham et al. 1991, Smith et al. 1994), we
found that mangroves of different size classes were
affected differently by hurricanes. Individual plants
(measured repeatedly) changed in LAI over time
differently depending on their zone and nutrient
treatment. In the A. germinans-dominated scrub zone
(Fig. 4), a comparison of before and after LAI values
suggested that these low-stature trees (,2 m tall) in the
interior portions of the mangrove forest sustained very
little defoliation or stem breakage. Similar observations
have been made after other hurricanes (Craighead 1971,
Smith et al. 1994, Lugo 1996, and Ross et al. 2006).
Following each hurricane, layers of wrack lodged in the
tops of scrub trees indicated that they were immersed by
high water during the storms, which likely protected
them from wind damage. In contrast, the taller trees (2–
3 m) in the A. germinans–L. racemosa transition zone,
just seaward of the scrub zone, sustained extensive
defoliation as well as breakage of most of the small
branches out of their canopies. This pattern of damage
further contrasts with the narrow R. mangle-dominated
fringe zone immediately along the water (4–5 m tall),
which experienced less defoliation but more breakage in
the canopy of major branches and boles. Over the course
of our study, recovery to pre-hurricane LAI in the fringe
and transitions zones seemed to be further exacerbated
by a persistent drought in the region.
Although the second of the two storms, Hurricane
Jeanne, was stronger than the first storm, Hurricane
Frances, before and after hurricane comparisons of LAI
at our study site showed that the first storm caused
significantly more damage to the mangrove canopy (Fig.
4). Frances was a category 2 hurricane with maximum
sustained winds of 167 km/h when it made landfall on
South Hutchinson Island on 5 September 2004 (NHC
2004). In contrast, Jeanne was a stronger category 3
hurricane with maximum sustained winds of 195 km/h
when it made landfall three weeks later along the same
portion South Hutchinson Island. It is likely that lower
damage by Hurricane Jeanne was due to less resistance
to the wind as a result of the prior defoliation that had
occurred during Hurricane Frances.
Based on large inputs of woody debris, measured at
three and six years after the September hurricanes, trees
in the fringe and transition zones also experienced
extensive additional damage as a result of delayed
mortality (Fig. 5). When we resampled our plots in 2007,
we found large dead branches in R. mangle trees in the
fringe zone that had been alive in November 2005. By
2010, the evidence of the delayed mortality was detected
in the fringe zone as a pulse of downed FWD, composed
of R. mangle twigs and small stems, and in the transition
zone as a pulse of recently dead standing FWD,
composed mainly small branches in the canopies of L.
racemosa that had died since our 2007 survey.
Krauss et al. (2005) found that the ratio of fine to
coarse woody debris increased with time after the
passage of Hurricane Andrew in south Florida. For
our site along the IRL, we found large increases in the
input of FWD and CWD for more than five years after
the Hurricanes Frances and Jeanne as a result of delayed
mortality in the canopy in the transition and fringe
zones, but with no evidence of delayed mortality in the
scrub zone regardless of nutrient treatment. However,
our data did not reveal any trends in FWD :CWD with
time. Consistent with observations made following
Hurricane Donna (Craighead and Gilbert 1962) and
Hurricane Andrew (e.g., Smith et al. 1994), these results
provide evidence that indirect damage was more severe
than the immediate effects of the storms. Delayed
mortality as an indirect effect of hurricane damage has
also been documented in other ecosystems, including
coral reefs in Jamaica (Knowlton et al. 1981), upland
forests in Puerto Rico (Walker 1995), and forests of the
southeastern United States (Zeng et al. 2009).
Storm recovery
In the year following hurricanes Frances and Jeanne,
the growth rates of both A. germinans and R. mangle
decreased in all three zones. By the second year, growth
rates started to increase but were still lower than they
were before the hurricanes (Feller et al. 2003b). Other
studies have found a similar pattern and have suggested
that pulses of dead wood result in short-term immobi-
lization of nutrients in microbial pools and lead to
decreased forest productivity (Zimmerman et al. 1995,
Scatena et al. 1996, Lugo 2008). Scatena et al. (1996)
showed that it took more than three years for leaf
nutrient concentrations to return to pre-hurricane
concentrations.
Numerous studies have investigated the impact of
hurricanes on coastal ecosystem processes (Greening et
al. 2006). Other researchers have suggested that because
mangrove forests are characterized by low floristic
diversity compared to other tropical and subtropical
forests, the recovery trajectory will be relatively simple
and biological diversity, forest structure, ecosystem
processes, and ecological functions will completely
regain their former states following hurricane distur-
bance (Lugo et al. 1976, Ogden 1992, Smith et al. 1994).
In this study, we found that recovery was not
straightforward, and that nutrient enrichment affected
both the level of impact and the rate of recovery from
November 2015 2969N ENRICHMENT INCREASES HURRICANE IMPACT
hurricane damage in a mangrove forest, with N-enriched
scrub trees taking longer to return to pre-hurricane
conditions than Control trees. Although there is
considerable uncertainty in climate model projections,
the intensity of hurricanes and perhaps the frequency
may increase in the future (Bender et al. 2010, Knutson
et al. 2010, Emanuel 2013, Wang et al. 2014). At the
same time, nutrient loading remains a threat to the
coastal zone and is projected to increase over the next
few decades as fertilizer use increases to meet the
demands of expanding human population (Deegan et
al. 2012). Given that scrub mangrove forests predomi-
nate so much of the mangrove landscape in south
Florida and the Caribbean (e.g., Feller and Mathis 1997,
Lugo 1997, Murray et al. 2004, Rodriguez and Feller
2004, Lovelock et al. 2005, Simard et al. 2006), our
study suggests that the coupling of these anthropogenic
stressors could influence the mechanisms and trajectory
of recovery of mangroves to large-scale disturbances and
thereby diminish their ability to provide valuable
ecosystem services.
ACKNOWLEDGMENTS
We thank the Smithsonian Marine Station in Fort Pierce forlogistical and field support and Anne Chamberlain and R.Feller for countless hours of fieldwork. This research wasfunded by the NASA Climate and Biological Responseprogram (NNX11AO94G) and the NSF MacroSystems Biol-ogy program (EF1065821). This is SMSFP Contribution No.995. We also thank Florida DEP Park Service and St. LucieCounty Mosquito Control and Coastal Management Servicesfor permission to work in Avalon SP.
LITERATURE CITED
Abtew, W., C. Pathak, R. S. Huebner, and V. Ciuca. 2010.Chapter 2: hydrology of the south Florida environment.Pages 1–73 in 2009 South Florida environmental report.Volume I. South Florida Water Management District, WestPalm Beach, Florida, USA.
Attiwill, P. M. 1994. The disturbance of forest ecosystems: theecological basis for conservative management. Forest Ecol-ogy and Management 63:247–300.
Baldwin, A. H., W. J. Platt, K. L. Gathen, J. M. Lessmann, andT. J. Rauch. 1995. Hurricane damage and regeneration infringe mangrove forests of southeast Florida, USA. Journalof Coastal Research 21:169–183.
Ball, M. C. 2002. Interactive effects of salinity and irradianceon growth: implications for mangrove forest structure alongsalinity gradients. Trees 16:126–139.
Bender, M. A., T. R. Knutson, R. E. Tuleya, J. J. Sirutis, G. A.Vecchi, S. T. Garner, and I. M. Held. 2010. Modeled impactsof anthropogenic warming on the frequency of intenseAtlantic hurricanes. Science 327:454–458.
Bertness, M. D., C. M. Crain, C. Holdredge, and N. Sala. 2008.Eutrophication and consumer control of New England saltmarsh primary productivity. Conservation Biology 22:131–139.
Bertness, M. D., P. Ewanchuk, and B. R. Silliman. 2002.Anthropogenic modification of New England salt marshlandscapes. Proceedings of the National Academy of SciencesUSA 99:1395–1398.
Bevan, J. L., II. 2014. Tropical Cyclone Report, HurricaneFrances, 25 August–8 September 2004. National HurricaneCenter, National Weather Service, NOAA, Miami, Florida,USA. http://www.nhc.noaa.gov/data/tcr/AL062004_Frances.pdf
Blake, E. C., E. N. Rappaport, J. D. Jarrell, and C. W. Sandsea.2005. The deadliest, costliest, and most intense United Statestropical cyclones from 1851 to 2004 (and other frequentlyrequested hurricane facts. NOAA Technical MemorandumNWS TPC-4. Tropical Prediction Center, National Hurri-cane Center, Miami, Florida, USA.
Boesch, D. F. 2002. Challenges and opportunities for science inreducing nutrient over-enrichment of coastal ecosystems.Estuaries 25:744–758.
Carlson, P. R., Jr., L. A. Yarbro, K. A. Kaufman, and R. A.Mattson. 2010. Vulnerability and resilience of seagrasses tohurricane and runoff impacts along Florida’s west coast.Hydrobiologia 649:39–53.
Carlson, P. R., L. A. Yarbro, C. F. Zimmermann, and J. R.Montgomery. 1983. Pore water chemistry of an overwashmangrove island. Florida Scientist 46:239–249.
Castaneda-Moya, E., R. R. Twilley, and V. H. Rivera-Monroy.2013. Allocation of biomass and net primary productivity ofmangrove forests along environmental gradients in theFlorida Coastal Everglades, USA. Forest Ecology andManagement 307:226–241.
Cloern, J. E. 2001. Our evolving conceptual model of thecoastal eutrophication problem. Marine Ecology ProgressSeries 210:223–253.
Craighead, F. C. 1971. The trees of south Florida. Volume I.Natural Press, Coral Gables, Florida, USA.
Craighead, F. C., and V. C. Gilbert. 1962. The effects ofHurricane Donna on the vegetation of southern Florida.Quarterly Journal of the Florida Academy of Sciences 25:1–28.
Davis, S. E., III, J. E. Cable, D. L. Childers, C. Coronado-Molina, J. W. Day, Jr., C. D. Hittle, C. J. Madden, E. Reyes,D. Rudnick, and F. Sklar. 2004. Importance of storm eventsin controlling ecosystem structure and function in a Floridagulf coast estuary. Journal of Coastal Research 20:1198–1208.
Deegan, L. A., D. S. Johnson, R. S. Warren, B. J. Peterson,J. W. Fleeger, S. Fagherazzi, and W. M. Wolheim. 2012.Coastal eutrophication as a driver of salt marsh loss. Nature490:388–394.
Dix, N. G., E. J. Phlips, and R. A. Gleeson. 2008. Water qualitychanges in the Guana Tolomato Matanzas National Estua-rine Research Reserve, Florida, associated with four tropicalstorms. Journal of Coastal Research 55:26–37.
Doyle, T. W., and G. F. Girod. 1997. The frequency andintensity of Atlantic hurricanes and their influence on thestructure of south Florida mangrove communities. Pages109–120 in H. F. Diaz and R. S. Pulwarty, editors.Hurricanes: climate and socioeconomic impacts. Springer-Verlag, Heidelberg, Germany.
Emanuel, K. A. 2013. Downscaling CMIP5 climate modelsshows increased tropical cyclone activity over the 21stcentury. Proceedings of the National Academy of SciencesUSA 110:12219–12224.
Feller, I. C. 1995. Effects of nutrient enrichment on growth andherbivory in dwarf red mangrove (Rhizophora mangle).Ecological Monographs 65:477–505.
Feller, I. C., and A. H. Chamberlain. 2007. Herbivore responsesto nutrient enrichment and landscape heterogeneity in amangrove ecosystem. Oecologia 153:607–616.
Feller, I. C., A. H. Chamberlain, C. Piou, S. K. Chapman, andC. E. Lovelock. 2013. Latitudinal patterns of herbivory inmangrove forests: consequences of nutrient over-enrichment.Ecosystems 16:1203–1215.
Feller, I. C., C. E. Lovelock, U. Berger, K. L. McKee, S. B.Joye, and M. C. Ball. 2010. The biocomplexity of mangroveecosystems. Annual Review of Marine Science 2:395–416.
Feller, I. C., and W. N. Mathis. 1997. Primary herbivory bywood-boring insects along a tree-height architectural gradi-ent of Rhizophora mangle L. in Belizean mangrove swamps.Biotropica 29:440–451.
ILKA C. FELLER ET AL.2970 Ecology, Vol. 96, No. 11
Feller, I. C., K. L. McKee, D. F. Whigham, and J. P. O’Neill.2003a. Nitrogen vs. phosphorus limitation across andecotonal gradient in a mangrove forest. Biogeochemistry62:145–175.
Feller, I. C., D. F. Whigham, K. L. McKee, and C. E.Lovelock. 2003b. Nitrogen limitation of growth and nutrientdynamics in a disturbed mangrove forest, Indian RiverLagoon, Florida. Oecologia 134:405–414.
Feller, I. C., D. F. Whigham, J. P. O’Neill, and K. M. McKee.1999. Effects of nutrient enrichment on within-stand nutrientcycling in mangrove ecosystems in Belize. Ecology 80:2193–2205.
Greening, H., P. Doering, and C. Corbett. 2006. Hurricaneimpacts on coastal ecosystems. Estuaries and Coasts 29:877–879.
Gresham, C. A., T. M. Williams, and D. J. Lipscomb. 1991.Hurricane Hugo wind damage to southeastern U.S. coastalforest tree species. Biotropica 23:420–426.
Herbert, D. A., J. H. Fownes, and P. M. Vitousek. 1999.Hurricane damage to a Hawaiian forest: nutrient supply rateaffects resistance and resilience. Ecology 80:908–920.
Hollings, C. S. 1973. Resilience and stability of ecologicalsystems. Annual Review of Ecology and Systematics 4:1–23.
Imbert, D., P. Labbe, and A. Rousteau. 1996. Hurricanedamage and forest structure in Guadeloupe, French WestIndies. Journal of Tropical Ecology 12:663–680.
Jiang, J., D. L. DeAngelis, G. H. Anderson, and T. J. Smith III.2014. Analysis and simulation of propagule dispersal andsalinity intrusion from storm surge on the movement of amarsh–mangrove ecotone in South Florida. Estuaries andCoasts 37:24–35.
Knowlton, N., J. C. Lang, M. C. Rooney, and P. Clifford. 1981.Evidence for delayed mortality in hurricane-damaged Jamai-can staghorn corals. Nature 294:251–252.
Knutson, T. R., J. L. McBride, J. Chan, K. Emanuel, G.Holland, C. Landsea, I. Held, J. P. Kossin, A. K. Srivastava,and M. Sugi. 2010. Tropical cyclones and climate change.Nature Geoscience 3:157–163.
Krauss, K. W., R. R. Twilley, T. J. Smith III, K. R. T. Whelan,and J. K. Sullivan. 2005. Woody debris in the mangroveforests of South Florida. Biotropica 37:9–15.
Lapointe, B. E., B. J. Bedford, and R. Baumberger. 2006.Hurricanes Frances and Jeanne remove blooms of theinvasive green alga Caulerpa brachypus forma parvifolia(Harvey) Cribb from coral reefs off northern Palm BeachCounty, Florida. Estuaries and Coasts 29:966–971.
Liu, K. 2007. Uncovering prehistoric hurricane activity.American Scientist 95:126–133.
Lovelock, C. E., M. C. Ball, B. Choat, B. M. J. Engelbrecht,N. M. Holbrook, and I. C. Feller. 2006. Linking physiolog-ical processes with mangrove forest structure: phosphorusdeficiency limits canopy development, hydraulic conductivityand photosynthetic carbon gain in dwarf Rhizophora mangle.Plant, Cell and Environment 29:793–802.
Lovelock, C. E., M. C. Ball, K. Martin, and I. C. Feller. 2009.Nutrient enrichment increases mortality of mangroves. PLoSONE 4:e5600.
Lovelock, C. E., and I. C. Feller. 2003. Photosyntheticperformance and resource utilization of two mangrovespecies coexisting in hypersaline scrub forest. Oecologia134:455–462.
Lovelock, C. E., I. C. Feller, M. F. A. Adame, R. R. Reef,H. M. Penrose, L. Wei, and M. C. Ball. 2011. Intense stormsand the delivery of materials that relieve nutrient limitationsin mangroves of an arid zone estuary. Functional PlantBiology 38:514–522.
Lovelock, C. E., I. C. Feller, K. L. McKee, B. M. J.Engelbrecht, and M. C. Ball. 2004. The effect of nutrientenrichment on growth, photosynthesis and hydraulic con-ductance of dwarf mangroves in Panama. FunctionalEcology 18:25–33.
Lovelock, C. E., I. C. Feller, K. L. McKee, and R. Thompson.2005. Variation in mangrove forest structure and sedimentcharacteristics in Bocas del Toro, Panama. CaribbeanJournal of Science 41:456–464.
Lugo, A. E. 1996. Caribbean island landscapes: indicators ofthe effects of economic growth on the region. Environmentand Development Economics 1:128–136.
Lugo, A. E. 1997. Old-growth mangrove forests in the UnitedStates. Conservation Biology 11:11–20.
Lugo, A. E. 2008. Visible and invisible effects of hurricanes onforest ecosystems: an international review. Austral Ecology33:368–398.
Lugo, A. E., M. Sell, and S. C. Snedaker. 1976. Mangroveecosystem analysis. Pages 113–145 in B. C. Patten, editor.Systems analysis and simulation in ecology 4. AcademicPress, New York, New York, USA.
Lugo, A. E., and S. C. Snedaker. 1974. The ecology ofmangroves. Annual Review of Ecology and Systematics 5:39–64.
Mallin, M. A., M. H. Posey, M. R. McIver, D. C. Parson, S. H.Ensign, and T. D. Alphin. 2002. Impacts and recovery frommultiple hurricanes in a piedmont-coastal plain river system.BioScience 52:999–1010.
Martin, K. C., D. Bruhn, C. E. Lovelock, I. C. Feller, J. R.Evans, and M. C. Ball. 2010. Nitrogen fertilization enhanceswater use efficiency in a saline environment. Plant, Cell andEnvironment 33:344–357.
McCoy, E. D., H. R. Mushinsky, D. Johnson, and W. E.Meshaka, Jr. 1996. Mangrove damage caused by HurricaneAndrew on the southwestern coast of Florida. Bulletin ofMarine Science 59:1–8.
McKee, K. L., D. Cahoon, and I. C. Feller. 2007. Caribbeanmangroves adjust to rising sea-level through biotic controlson soil elevation change. Global Ecology and Biogeography16:546–556.
Milbrandt, E. C., J. M. Greenawalt-Boswell, P. D. Sokoloff,and S. A. Bortone. 2006. Impact and response of southwestFlorida mangroves to the 2004 hurricane season. Estuariesand Coasts 29:979–984.
Minchinton, T. E., and L. A. McKenzie. 2008. Nutrientenrichment affects recruitment of oysters and barnacles in amangrove forest. Marine Ecology Progress Series 354:181–189.
Morris, J. T., and P. M. Bradley. 1999. Effects of nutrientloading on the preservation of organic carbon in wetlandsediments. Limnology and Oceanography 44:699–702.
Morris, J. T., G. P. Shaffer, and J. A. Nyman. 2013. BrinsonReview: perspectives on the influence of nutrients on thesustainability of coastal wetlands. Wetlands 33:975–988.
Murray, M. R., S. A. Zisman, P. A. Furley, D. M. Munro, J.Gibson, J. Ratter, S. Bridgewater, C. D. Minty, and C. J.Place. 2003. The mangroves of Belize: Part 1. Distribution,composition and classification. Forest Ecology and Manage-ment 174:265–279.
NHC. 2004. 2004 tropical cyclone advisory archive. NationalHurricane Center, National Weather Service, NOAA,Miami, Florida, USA. http://www.nhc.noaa.gov/archive/2004/
NHC. 2012. The Saffir-Simpson hurricane wind scale. NationalHurricane Center, National Weather Service, NOAA,Miami, Florida, USA. http://www.nhc.noaa.gov/pdf/sshws.pdf
Odum, W. E., C. C. McIvor, and T. J. Smith III. 1982. Theecology of the mangroves of South Florida: a communityprofile. FWS/OBS—81/24. U.S. Fish and Wildlife Service,Office of Biological Services, Washington, D.C., USA.
Ogden, J. C. 1992. The impact of Hurricane Andrew on theecosystems of south Florida. Conservation Biology 6:488–490.
Paperno, R. D., M. Tremain, D. H. Adams, A. P. Sebastian,J. T. Sauer, and J. Dutka-Gianelli. 2006. The disruption andrecovery of fish communities in the Indian River Lagoon,
November 2015 2971N ENRICHMENT INCREASES HURRICANE IMPACT
Florida, following two hurricanes in 2004. Estuaries andCoasts 29:1004–1010.
Piou, C., I. C. Feller, U. Berger, and F. Chi. 2006. Zonationpatterns of Belizean offshore mangrove forests 41 years aftera catastrophic hurricane. Biotropica 38:365–374.
Proffitt, C. E., E. C. Milbrandt, and S. E. Travis. 2006. Redmangrove (Rhizophora mangle) reproduction and seedlingcolonization after Hurricane Charley: comparisons of Char-lotte Harbor and Tampa Bay. Estuaries and Coasts 29:972–978.
R Core Team. 2009. R: a language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna,Austria. http://www.R-project.org/
Reef, R., I. C. Feller, and C. E. Lovelock. 2010. Nutrition ofmangroves. Tree Physiology http://dx.doi.org/10.1093/treephys/tpq048
Rey, J., and T. Kain. 1991. A guide to the salt marshimpoundments of Florida. University of Florida, FloridaMedical Entomology Laboratory, Vero Beach, Florida,USA.
Rey, J. R., R. A. Crossman, T. R. Kain, and D. S. Taylor. 1986.An overview of impounded mangrove forests along asubtropical lagoon in east-central Florida, USA. Pages341–350 in L. J. Bhosale, editor. Proceedings of NationalSymposium on Biology, Utilization and Conservation ofMangroves, 18–20 November 1985. Shivaji University,Kolhapur, India.
Rey, J. R., J. Shaffer, R. Crossman, and D. Tremain. 1990.Above-ground primary production in impounded, ditched,and natural Batis-Salicornia marshes along the Indian RiverLagoon, Florida, USA. Wetlands 10:151–171.
Rey, J. R., J. Shaffer, T. Kain, R. Stahl, and R. Crossman.1992. Sulfide variation in the pore and surface waters ofartificial salt-marsh ditches and a natural tidal creek.Estuaries 15:257–269.
Ridler, M. S., R. C. Dent, and D. A. Arrington. 2006. Effects oftwo hurricanes on Syringodium filiforme, manatee grass,within the Loxahatchee River Estuary, Southeast Florida.Estuaries and Coasts 29:1019–1025.
Rivera-Monroy, V. H., et al. 2004. A biogeochemical concep-tual framework to develop long term ecological research andsustain coastal management in the wider Caribbean Region.BioScience 54:843–856.
Rodriguez, W., and I. C. Feller. 2004. Mangrove landscapecharacterization and change in Twin Cays, Belize using aerialphotography and IKONOS satellite data. Atoll ResearchBulletin 513:1–22.
Ross, M. S., P. L. Ruiz, J. P. Sah, D. L. Reed, J. Walters, andJ. F. Meeder. 2006. Early post-hurricane stand developmentin fringe mangrove forests of contrasting productivity. PlantEcology 185:283–297.
Ross, M. S., P. L. Ruiz, G. J. Telesnicki, and J. F. Meeder.2001. Estimating above-ground biomass and production inmangrove communities of Biscayne National Park, Florida(U.S.A.). Wetlands Ecology and Management 9:27–37.
Roth, L. C. 1992. Hurricanes and mangrove regeneration:effects of Hurricane Joan, October 1988, on the vegetation ofIsla del Veneto, Bluefields, Nicaragua. Biotropica 24:375–384.
Sallenger, A. H., H. F. Stockdon, L. Fauver, M. Hansen, D.Thompson, C. W. Wright, and J. Lillycrop. 2006. Hurricanes2004: an overview of their characteristics and coastal change.Estuaries and Coasts 29:880–888.
Scatena, F. N., S. Moya, C. Estrada, and J. D. Chinea. 1996.The first five years of reorganization of aboveground biomassand nutrient use following Hurricane Hugo in the Bisleyexperimental watersheds, Luquillo Experimental Forest,Puerto Rico. Biotropica 28:424–440.
SFWMD. 2009. South Florida Water Management District,Media Advisory for March 10, 2009. http://sfwmd.gov/portal/page/portal/xrepository/sfwmd_repository_pdf/ma_2009_0310_drought_ebriefing.pdf
Sherman, R. E., T. J. Fahey, and P. Martinez. 2001. Hurricaneimpacts on a mangrove forest in the Dominican Republic:damage patterns and early recovery. Biotropica 33:393–408.
Silver, W. L., S. Brown, and A. E. Lugo. 1996. Effects ofchanges in biodiversity on ecosystem function in tropicalforests. Conservation Biology 10:17–24.
Simard, M., K. Zhang, V. H. Rivera-Monroy, M. S. Ross, P. L.Ruiz, E. Castaneda-Moya, R. R. Twilley, and E. Rodriguez.2006. Mapping height and biomass of mangrove forests inEverglades National Park with SRTM elevation data.Photogrammetric Engineering and Remote Sensing 72:299–311.
Smith, T. J., III, G. H. Anderson, K. Balentine, G. Tiling, G. A.Ward, and K. R. T. Whelan. 2009. Cumulative impacts ofhurricanes on Florida mangrove ecosystems: sedimentdeposition, storm surges and vegetation. Wetlands 29:24–34.
Smith, T. J., III, M. B. Robblee, H. R. Wanless, and T. W.Doyle. 1994. Mangroves, hurricanes, and lightning strikes.BioScience 44:256–262.
Steward, J. S., R. W. Virnstein, M. A. Lasi, L. J. Morris, J. D.Miller, L. M. Hall, and W. A. Tweedale. 2006. The impactsof the 2004 hurricanes on hydrology, water quality, andseagrass in the central Indian River Lagoon. Estuaries andCoasts 29:954–965.
Systat. 2004. Systat 11. Systat Software, San Jose, California,USA.
Tomasko, D. A., C. Anastasiou, and C. Kovach. 2006.Dissolved oxygen dynamics in Charlotte Harbor and itscontributing watershed, in response to Hurricanes Charley,Frances, and Jeanne—impacts and recovery. Estuaries andCoasts 29:932–938.
Turner, R. E. 2012. Beneath the salt marsh canopy: loss of soilstrength with increasing nutrient loads. Estuaries and Coasts34:1084–1093.
Turner, R. E., B. L. Howes, J. M. Teal, C. S. Milan, E. M.Swenson, and D. D. Goehringer-Toner. 2009. Salt marshesand eutrophication: an unsustainable outcome. Limnologyand Oceanography 54:1634–1642.
Verhoeven, J. T. A., B. Arheimer, C. Yin, and M. M. Hefting.2006. Regional and global concerns over wetlands and waterquality. Trends in Ecology and Evolution 21:96–103.
Vogt, J., A. Skora, I. C. Feller, C. Piou, G. Coldren, and U.Berger. 2011. Investigating the role of impoundment andforest structure on the resistance and resilience of mangroveforests to hurricanes. Aquatic Botany 97:24–29.
Walker, L. R. 1995. Timing of post-hurricane tree mortality inPuerto Rico. Journal of Tropical Ecology 11:315–320.
Wang, H., et al. 2014. How well do global climate modelssimulate the variability of Atlantic tropical cyclones associ-ated with ENSO? Journal of Climate 27:5673–5692.
Wang, P., and M. H. Horwitz. 2007. Erosional and depositionalcharacteristics of regional overwash deposits caused bymultiple hurricanes. Sedimentology 54:545–564.
Zeng, H., J. Q. Chambers, R. I. Negron-Juarez, G. C. Hurtt,D. B. Baker, and M. D. Powell. 2009. Impacts of tropicalcyclones on U.S. forest tree mortality and carbon flux from1851 to 2000. Proceedings of the National Academy ofSciences USA 106:7888–7892.
Zimmerman, J. K., W. M. Pulliam, D. J. Lodge, V. Quinones-Orfila, N. Fetcher, S. Guzman-Grajales, J. A. Parrotta, C. E.Asbury, L. R. Walker, and R. B. Waide. 1995. Nitrogenimmobilization by decomposing woody debris and therecovery of tropical wet forest from hurricane damage. Oikos72:314–322.
ILKA C. FELLER ET AL.2972 Ecology, Vol. 96, No. 11