Content uploaded by Alex E. Mercado-Molina
Author content
All content in this area was uploaded by Alex E. Mercado-Molina on Nov 17, 2019
Content may be subject to copyright.
REPORT
Demography of the threatened coral Acropora cervicornis:
implications for its management and conservation
Alex E. Mercado-Molina
1,2
•Claudia P. Ruiz-Diaz
2,3
•Marı
´aE.Pe
´rez
4
•
Ruber Rodrı
´guez-Barreras
1
•Alberto M. Sabat
1
Received: 6 October 2014 / Accepted: 24 August 2015
ÓSpringer-Verlag Berlin Heidelberg 2015
Abstract Populations of Acropora cervicornis have col-
lapsed throughout the Caribbean. This situation has prompted
the initiation of many restoration efforts; yet, there are insuf-
ficient demographic data and analyses to effectively guide
these initiatives. In this study we assessed the spatiotemporal
variability of A. cervicornis vital rates. We also developed a
population matrix model to (1) evaluate the risk of population
extinction, (2) estimate population growth rates (k)consid-
ering different rates of colony fragmentation and fragment
survival, (3) determine the demographic transition(s) that
contribute the most to spatiotemporaldifferences in ks, and (4)
analyze the effectiveness of outplanting coral fragments of
different sizes. The model was parameterized by following the
fate of 300 colonies from 2011 to 2013 at two localities in
Puerto Rico. Demographic transitions varied spatiotempo-
rally, with a significant interaction between location and time
period on colony fate. Spatiotemporal variations in kwere
also observed. During the first year, populations exhibited ks
below equilibrium (0.918 and 0.948), followed by a dramatic
decline at both sites (0.535 and 0.709) during the second year.
The lower ks were caused by a decrease in the probability of
stasis of large-sized colonies coupled with lack of sexual
recruits and a meager contribution of asexual recruitment.
Spatial variations in ks were largely due to differences in the
probability of medium-sized colonies advancing to the largest
size class. The viability analysis forecasts that the populations
will reach quasi-extinction levels of 25 % of the initial pop-
ulation size in B16 yrs. Numerical simulations indicate that
outplanting fragments C250 cm in total linear length (TLL)
would result in a higher asymptotic population size than out-
planting smaller fragments. We argue, however, that trans-
planting colonies B100 cm TLL will be a better management
strategy because they can be produced faster and in higher
numbers at coral nurseries.
Keywords Acropora cervicornis Coral demography
Population viability analysis Reef restoration
Introduction
Coral reefs are among the most threatened ecosystems on
earth. The structural and ecological integrity of approxi-
mately 60 % of coral reefs has been severely impacted by
anthropogenic activities (e.g., water pollution), biological
factors (e.g., diseases), and physical disturbances (e.g.,
bleaching events associated with elevated seawater tem-
perature); by 2030, 90 % of reefs are expected to be com-
promised (Burke et al. 2011). Today, approximately half of
the acroporid species, which are among the major reef-
building corals worldwide, are listed as threatened by the
International Union for Conservation of Nature (Carpenter
Communicated by Biology Editor Dr. Mark Vermeij
Electronic supplementary material The online version of this
article (doi:10.1007/s00338-015-1341-8) contains supplementary
material, which is available to authorized users.
&Alex E. Mercado-Molina
amolinapr@gmail.com
1
Department of Biology, University of Puerto Rico,
PO Box 23360, Rı
´o Piedras, Puerto Rico
2
Sociedad Ambiente Marino,
PO Box 22158, San Juan 00931-2158, Puerto Rico
3
Department of Environmental Sciences, University of Puerto
Rico, PO Box 70377, Rı
´o Piedras, Puerto Rico
4
Department of Mathematics, University of Puerto Rico,
PO Box 70377, Rı
´o Piedras, Puerto Rico
123
Coral Reefs
DOI 10.1007/s00338-015-1341-8
et al. 2008). This list includes Acropora cervicornis, one of
the most depleted species in the wider Caribbean. Popula-
tions of this coral have been decreasing dramatically (up to
97 % decline in abundance) throughout its geographical
range since the late 1970s (Aronson and Precht 2001; Miller
et al. 2002), and currently, it is considered a threatened
species under the US Endangered Species Act (NMFS 2006).
This is of concern not only because A. cervicornis is a major
contributor to reef accretion but also because its branching
morphology serves as nursery ground for multiple vertebrate
and invertebrate species, thereby promoting coral reef bio-
diversity (Precht et al. 2002; Quinn and Kojis 2006).
Due to the decline of A. cervicornis and its ecological
importance, efforts have been undertaken to revert its pop-
ulation decrease. In the USA and its territories, critical
habitats have been protected, restoration programs have been
launched, and research and monitoring programs have been
initiated. Surprisingly, none of these conservation and
restoration efforts are based on solid quantitative demo-
graphic analyses (Williams et al. 2006). The few population
studies available have focused on estimating rates of colony
growth, survival, and recruitment (Tunnicliffe 1981;
Knowlton et al. 1990), but none have evaluated how varia-
tion in these demographic parameters affects local popula-
tion growth (k). This hinders the effectiveness of
conservation efforts because initiatives intended to improve
the persistence of a population may fail to target those vital
rates that have the greatest contribution to k. Conservation
management strategies also depend on quantitative predic-
tions that can only be achieved by employing demographic
analysis and modeling. Indeed, the lack of such analyses has
been identified as an area of particular concern by the NOAA
Acropora Biological Review Team (ABRT 2005) and the
National Marine Fisheries Service (NMFS 2015).
Among the major goals of conservation and manage-
ment efforts are first to arrest the decline of threatened
populations and then to increase the rate of population
growth until a ‘‘safe’’ population size is reached. The rate
at which a population grows or declines is inevitably linked
to its individuals’ survival, growth, and reproduction.
Therefore, effective conservation initiatives require
knowledge of how variations in vital rates relate to varia-
tions in population growth. In this respect, population
matrix models have become an important tool in conser-
vation (Caswell 2001). However, relatively few scientists
have applied population matrix models to answer specific
management questions focusing on coral conservation (but
see Linares et al. 2008; Vardi et al. 2012). Multiple factors
make corals a complex subject for demographic studies. In
many instances, corals are physically and practically hard
to measure. Branching corals, for example, can form
thickets that make it difficult to identify the physical
boundaries of individual colonies. In fact, the literature
focusing on the relationship between scleractinian coral
vital rates and their population growth rates is scarce.
In this study, we examined the demography of A. cer-
vicornis at two localities in northeastern Puerto Rico from
2011 to 2013. The aims were to measure the spatiotem-
poral variability in vital rates and to use the demographic
data to perform a population viability analysis. For the
viability analysis, a size-based population matrix model
was constructed to (1) determine whether the studied
populations were stable (k=1), increasing (k[1), or
decreasing (k\1), (2) estimate quasi-extinction proba-
bilities, (3) identify the demographic transition(s) that
accounted for the observed spatiotemporal differences in k
by performing life table response analyses, (4) estimate k
under different rates of colony fragmentation and fragment
survival, and (5) analyze the efficacy of different out-
planting scenarios through numerical simulations. The
numerical simulations were performed to assess (not pre-
dict) possible population trajectories of A. cervicornis for
given conditions (Caswell 2001). This study is the first to
use demographic modeling to better understand the popu-
lation dynamics of the threatened coral A. cervicornis
while evaluating the effect of different transplanting sce-
narios on population growth.
Materials and methods
Demographic data
We surveyed suitable habitats within an area covering
approximately 150 km
2
along the northeast coast of Puerto
Rico to identify localities with a relatively high abundance
of A. cervicornis colonies. The surveyed area encompassed
coral reefs within La Cordillera Nature Reserve (LCNR),
the Canal Luis Pen
˜a Nature Reserve (CLPNR), and the reef
areas around the west side of the island of Culebra, zones
where this coral has historically been prevalent (Wirt et al.
2015). We found very few extant populations and only two
localities—Canal Luis Pen
˜a (CLP) and Palomino (PAL)—
with an adequate number of discernible colonies (no
thickets) to study the demography of A. cervicornis. CLP
(18°1801400N, 65°2001500 W) is located within CLPNR on
the island municipality of Culebra, approximately 30 km
east of the Puerto Rico Island. PAL (18°210300 N,
65°3305800W) is located within LCNR, 6 km from the
coastal municipality of Fajardo on the Puerto Rico main-
land. The reef structure at both sites has a northeast ori-
entation and is exposed to the easterly trade winds and
long-period swells during the winter. Consequently, both
sites are frequently exposed to moderate to strong wave
action. Water quality is good (i.e., low turbidity, no ter-
rigenous sediments) as none of the sites are directly
Coral Reefs
123
influenced by river-derived sediment, nutrient discharges,
or coastal development. At the two sites, A. cervicornis
colonizes a zone that was previously dominated by A.
palmata but is currently dominated by octocorals and
characterized by a low topographic relief. As mentioned
above, the studied populations are not thickets as defined
by the NMFS (2015), but rather have moderate densities of
easily identified individual colonies distributed throughout
the area. In this sense, the study can be understood as
describing the demography of two populations in the pro-
cess of forming (or not) a thicket. As A. cervicornis
thickets are currently rare, these two populations are likely
to be representative of the current state of populations of
this coral within its range.
One hundred and fifty colonies were tagged at each site,
at a depth of 3–5 m using numbered tags fixed with
masonry nails to non-living substrate adjacent to the col-
ony. Their survival, growth, and branch fragmentation
were measured annually from 2011 to 2013. Colonies were
considered dead if no live tissue was distinguishable or if
the colony disappeared during the study period. Growth
rates were estimated by comparing changes in total linear
length (TLL) between surveys. Initial and final colony sizes
were calculated as the sum of the lengths of all live por-
tions of branches (Knowlton et al. 1990). To do this, a set
of photographs was taken in situ from different angles with
a scale by side (Mercado-Molina et al. 2014), which
allowed us to measure all branches fully extended. This
approach is an accurate estimate of the actual colony size
of A. cervicornis as demonstrated by Mercado-Molina et al.
(2014). The free software CPCe version 4.1 (Kohler and
Gill 2006) was used to make all measurements.
Recruitment and natural fragmentation
At the two study sites, we randomly established thirty 1 m
2
permanent quadrats along a 60 m
2
belt transect to assess
sexual and asexual recruitment rates. All quadrats were
marked with nails to facilitate relocation in subsequent
surveys. Following Tunnicliffe (1981) and Knowlton et al.
(1990), we defined sexual recruitment as a small crust
showing a round or ellipsoidal morphology and measuring
less than 10 cm in height with a vertical orientation.
Asexual recruits were differentiated from sexual recruits by
their orientation (horizontal vs. vertical), signs of obvious
fragmentation, and size (TLL C10 cm) (Tunnicliffe 1981;
Knowlton et al. 1990; but see Williams and Miller 2006).
Rates of natural fragmentation were also estimated in situ
by counting the number of broken branches and/or scars
within a given colony and by comparing images between
surveys to determine the number of branches that were lost
(Tunnicliffe 1981; Mercado-Molina et al. 2015).
Visual inspections of the studied areas were performed
every 3–4 months to evaluate the overall health of the
marked colonies (looking for signs of disease or outbreaks
of coral-feeding snails).
Demographic analysis
A size-based matrix population model was constructed to
analyze and simulate the dynamics of A. cervicornis at the
study sites (Eq. 1).
s
m
l
0
@1
Atþ1
¼
Sss þðPfÞdðÞRsm þðPfÞdðÞ Rsl þðPfÞdðÞ
Gms Smm Rml þðPfÞdðÞ
Gls Glm Sll
0
@1
A
s
m
l
0
@1
At
ð1Þ
The number of small-sized colonies [s](B100 cm TLL),
medium-sized colonies [m] (101–250 cm TLL), and large-
sized colonies [l]([250 cm TLL) at time t?1 (one year
into the future) equals the current number of colonies in
each of the three size classes multiplied by a 3 93 matrix
of the transition probabilities among the size classes. The
diagonal elements in the matrix are the probabilities of
colonies surviving and remaining in their current size class.
In the case of small-sized colonies, this value is given by
the sum of two terms: (1) the probability of surviving and
not growing to the next size class (stasis, S
ss
) and (2) the
product (Pf)(d), which represents the probability of a
small-sized colonies producing another small-sized colo-
nies by means of branch fragmentation (Pf), multiplied by
the probability of the fragment surviving one year (d). For
each size class, the probability of colony fragmentation (Pf)
was estimated as the number of colonies that fragmented
divided by the total number of colonies. Annual fragment
survival rates of 24 % at CLP and 32 % at PAL were
estimated by following the fate of 100 loose fragments at
each site between February 2012 and August 2013 (Mer-
cado-Molina et al. 2014). The supradiagonal elements
represent the contribution of the large-sized colonies clas-
ses to smaller ones by regressing in size due to partial
mortality (R
sm
). medium-sized colonies and large-sized
colonies also contribute to the small-sized colonies size
class by branch fragmentation as described above. Lastly,
the subdiagonal elements represent the probabilities of a
colony growing to the next size class (G
ms
). Sexual
recruitment was not included in the transition matrix
because no larval-derived recruits were observed during
this study. With respect to asexual reproduction, we plan-
ned for two independent estimates of asexual recruitment
(branch fragmentation and the appearance of colony
Coral Reefs
123
fragments in the recruitment plots) because we did not
know a priori which would provide better data. Given that
no asexual recruits were observed in these plots, we only
use rate of branch fragmentation in the model. The tran-
sition probabilities were estimated by constructing transi-
tion frequency tables (Caswell 2001) for each locality and
census period (2011–2012 and 2012–2013; Electronic
Supplementary Materials, ESM, Table 1). The three size
classes were chosen to incorporate size-specific patterns of
vital transition rates while maintaining a sample size
greater than 25 colonies for each size category.
Four transition matrices were generated, one for each
site and census period. To determine whether the transition
probabilities in these four matrices were independent of
time and location, log-linear models were applied to a four-
way contingency table. Following Caswell (2001), fate
(F) was set as the response variable and time (T) and
location (L) were set as the explanatory factors, conditional
upon initial colony size (S). The model that best fit the data
was determined by means of the scaled Akaike information
criteria (AIC), calculated for log-linear models as
AIC =G
2
-2(df) (Caswell 2001), where G
2
is the
goodness-of-fit log-likelihood ratio statistic obtained from
comparing the model with the saturated model and df is the
degrees of freedom of the test. Log-linear models were also
applied to a three-way contingency table developed for
each of the size classes to examine the effect of time and
location on the fate of colonies within each size class. As
suggested by Fingleton (1984), 0.5 was added to each cell
value within the contingency tables to avoid estimation
problems for values equal to 0.
For each matrix, the dominant eigenvalue was calcu-
lated to obtain the asymptotic finite rate of increase in the
population (k) (Caswell 2001). Transition matrices were
subject to a bootstrapping re-sampling procedure (10,000
simulations) to obtain the 95 % confidence intervals for k.
To determine the demographic transition(s) that accounted
for most of the observed spatio-temporal differences in k,a
life table response experiment (LTRE) analysis was per-
formed rather than an elasticity analysis. LTRE is a ret-
rospective analysis that provides information on how much
variation in a particular life cycle transition actually con-
tributed to the observed differences in kbetween treat-
ments (years and localities in this study). Elasticity, on the
other hand, is a prospective (predictive) analysis that looks
at how kcould respond when a particular life cycle tran-
sition is perturbed (Caswell 2001).
Environmental stochasticity (Morris and Doak 2002)
was modeled by randomly selecting one of the two tran-
sition matrices constructed for the time periods 2011–2012
and 2012–2013. A total of 50,000 iterations (Stubben and
Milligan 2007) were performed in which the selected
matrix was post-multiplied by the initial population vector
to obtain the stochastic lambda (k
s
). During the iterations,
transition matrices had equal probability of being selected.
These stochastic trajectories were used to (1) estimate the
probability of populations reaching quasi-extinction
thresholds of 10, 15, 20, and 25 % of the initial population
size, (2) estimate ks for probabilities of colony fragmen-
tation and fragment survival varying between 0.10 and 1.0,
and (3) analyze the effectiveness of fragment transplanta-
tion as a management tool to aid in the recovery of A.
cervicornis populations. In the latter case, we modified
Eq. 1by adding a recruitment vector as expressed in Eq. 2,
where the elements R
s
,R
m
, and R
l
represent the numbers of
small-sized colonies, medium-sized colonies and large-
sized colonies to be transplanted, respectively. Equation 2
is an open population model, in which coral transplants
come from an external source (e.g., coral nurseries). To
determine the most effective number and size of fragments
to maintain a population equal or larger than the initial
population size, Eq. 2was projected for 20 yrs considering
the outplanting of 25, 50, 75, or 100 small-sized colonies,
medium-sized colonies, or large fragments per year.
s
m
l
0
@1
Atþ1
¼
Sss þPfðÞdðÞRsm þPfðÞdðÞ Rsl þPfðÞdðÞ
Gms Smm Rml þPfðÞdðÞ
Gls Glm Sll
0
@1
A
s
m
l
0
@1
At
þ
Rs
Rm
Rl
0
@1
At
ð2Þ
R.3.0.1 package popbio (Stubben and Milligan 2007;R
Development Core Team 2014) was used to perform all
demographics analyses. For log-linear analyses, the pack-
age MASS was used (Venables and Ripley 2002).
Results
Transition rates
During the first census period (2011–2012), the most fre-
quent transition was stasis (colonies remaining within their
original size classes), with large-sized colonies showing the
highest transition value at both sites (Table 1). At CLP, the
percentage of colonies (medium-sized colonies ?large-
sized colonies) that retrogressed to a lower size class
(28 %) was similar to the overall number of colonies
(small-sized colonies ?medium-sized colonies) that grew
to a larger size class (29 %). In contrast, at PAL more
colonies grew (38 %) than shrank (13 %). At both sites,
shrinkage occurred more frequently on medium-sized
colonies, whereas rates of size progression were similar
between small-sized colonies and medium-sized colonies
classes. Colony survival did not differ statistically between
Coral Reefs
123
size classes, varying between 82 and 96 % at CLP
(v
2
=3.77, df =2, p[0.05) and between 84 and 96 % at
PAL (v
2
=4.74, df =2, p[0.05). Overall survival rates
(size classes pooled) did not differ significantly between
sites (CLP: 88 %; PAL: 89 %; v
2
=0.0955, df =1,
p[0.05).
During the second year (2012–2013), shrinkage
became the most frequent transition, as both medium-
sized colonies and large-sized colonies experienced rela-
tively high rates of tissue loss. In this time period, 46 and
37 % of the colonies (medium-sized colonies?large-
sized colonies) transitioned to a smaller size class at CLP
and PAL, respectively. On the other hand, the percentages
of colonies (small-sized colonies ?medium-sized colo-
nies) growing into the next size class were lower than the
previous year, 9 % at CLP and 21 % at PAL. Differences
between sites in the proportion of colonies that progressed
in size were associated with the low number of medium-
sized colonies that grew to the largest class at CLP, as
rates of progression of small-sized colonies were similar
between sites (CLP: 13 %, PAL: 14 %). Significant dif-
ferences in survival rates between size classes were evi-
dent during the second year (CLP: v
2
=13.27, df =2,
p\0.05; PAL: v
2
=15.19, df =2, p\0.05). This was
mostly due to the low number of small-sized colonies that
survived from 2012 to 2013 (ESM Table 1). When
comparing between time periods within study sites, sur-
vival rates differed significantly (CLP: v
2
=24.93,
df =1, p\0.05; PAL: v
2
=9.15, df =1, p\0.05).
Overall survival rates (size classes pooled) differed sig-
nificantly between sites during the period 2012–2013,
being significantly lower at CLP (CLP: 60 %; PAL:
75 %; v
2
=13.13, df =1, p\0.05).
When considering the whole study period (2011–2013),
only 53 % of the tagged colonies survived at CLP, com-
pared with 67 % at PAL (v
2
=5.90, df =1, p\0.05).
Survival was size independent at CLP (v
2
=5.44, df =2,
p[0.05) but varied with respect to size at PAL
(v
2
=17.99, df =2, p\0.05).
Log-linear analysis
Results from the four-way contingency table indicate that
both time (T) and location (L) had a significant effect on the
demographic rates of A. cervicornis colonies (ESM Table 2).
When comparing each model to the saturated model (TLSF),
the model that considered the effects of time and location
simultaneously (TLS, FST, FSL) was the one with the best
goodness of fit (ESM Table 2). That is, transitions among
size classes are better explained by the interaction between
location and time than by the independent effect of each
factor. Nevertheless, the model that included the effect of
location while excluding the effect of time (TLS, FSL) had
comparable AIC values to that of the model TLS, FST, FSL
(ESM Table 3); therefore, TLS, FSL could also be consid-
ered a good approximation of the observed data. Indeed, the
DAIC value for TLS, FSL was below 2 indicating that the
model has better support than the models TLS, FS and TLS,
FST (ESM Table 3). This result suggests that location may
be a more parsimonious (‘‘better’’) predictor of colony fate
than time. This is supported by the fact that, in contrast to
location, time did not have a significant effect on the fate of
small-sized colonies and large-sized colonies (ESM
Table 2), either when tested against the null model (TL, F) or
when tested against the model in which the factor location
was included (ESM Table 2).
Table 1 Size class transition probabilities of small-sized colonies
sized (S: B100 cm in TTL), medium-sized colonies (M: 101–250 cm
in TTL) and large-sized (L: [250 cm TTL) colonies of Acropora
cervicornis at two localities (Canal Luis Pen
˜a reef, CLP, and
Palomino reef, PAL) in eastern Puerto Rico during two annual time
periods
Size class 2011–2012 2012–2013
SMLS ML
CLP
S 0.5010 0.3309 0.0581 0.3029 0.4262 0.3041
M 0.3272 0.3225 0.1851 0.1276 0.2325 0.2500
L 0 0.2580 0.7407 0 0.0465 0.2500
k0.918 (95 % CI 0.852–96) 0.535 (95 % CI 0.455–0.599)
PAL
S 0.4710 0.1456 0.0387 0.4365 0.3584 0.2507
M 0.3134 0.5000 0.1515 0.1428 0.2156 0.1951
L 0.0895 0.3400 0.7575 0 0.2549 0.4390
k0.948 (95 % CI 0.892–981) 0.709 (95 % CI 0.664–0.778)
k=estimated population growth rate; numbers within ( ) represent the lower and upper 95 % confidence intervals
Coral Reefs
123
Recruitment and colony fragmentation
We did not observe the establishment of any sexual recruits
or colony fragments within permanent quadrats. Branch
fragmentation was rare. At CLP, only 7 % of the colonies
fragmented during the first year, whereas 8 % of the
colonies fragmented during 2012–2013. At PAL, 7 % of
colonies showed signs of fragmentation during 2011–2012,
while 10 % of the colonies fragmented during 2012–2013.
The total number of fragments produced during the whole
study period (2011–2013) was 26 and 25 at CLP and PAL,
respectively. All resulting fragments were small-sized
colonies (\100 cm TLL).
Population growth
During the first year of the study (2011–2012), populations
at both localities exhibited growth rates that were close to,
but significantly below 1.0, with no differences between
locations (based on 95 % CI), at 0.91 for CLP and 0.94 for
PAL (Table 1). During the second year, ks decreased
considerably at both sites (Table 1), with the decline being
more dramatic at CLP than at PAL. Growth rates differed
significantly between sites during 2012–2013 as well as
when comparing between time periods (2011–2012 vs.
2012–2013) within locations (Table 1).
LTRE analyses indicate that at both sites, temporal
differences in ks (2011–2012 vs. 2012–2013) were largely
related to a reduction in the stasis of large-sized colonies
(P
ll
; Fig. 1a, b). The spatial difference in ks observed
during 2012–2013 can be explained mostly by differences
in growth between medium-sized colonies and large-sized
colonies (G
lm
), but also with an important contribution
from differences in the stasis of small-sized colonies and
large-sized colonies (P
ss
,P
ll
; Fig. 1c).
The estimated stochastic population growth rates were
0.717 (CI 0.715–0.718) and 0.844 (CI 0.843–0.845) at
CLP and PAL, respectively. At these rates of decline, the
population at CLP would reach any of the proposed quasi-
extinction levels in less than 14 yrs, whereas at PAL it
would take between 16 and 20 yrs (Fig. 2). Numerical
simulations indicate that the observed rates of colony
fragmentation are not sufficient to maintain populations at
equilibrium, even if all produced fragments show an
annual survival rate of 100 % (Fig. 3a). On the other
hand, given the observed fragment survival rates of 24 %
at CLP and 32 % at PAL, the population at CLP would
fail to reach equilibrium even for a probability of frag-
mentation of 1.0, whereas at PAL kC1 can be obtained
if the probability of fragmentation is greater than 80 %
(Fig. 3b).
Figures 4and 5show the projected size of A. cervicornis
populations under different outplanting (management)
scenarios. Transplanting large-sized colonies resulted in a
higher asymptotic population size in all cases. Neverthe-
less, the analysis also indicates that transplanting fewer
than 25 colonies per year at CLP, regardless of size, would
result in abundances below the initial population size. In
contrast, at PAL, transplanting 25 colonies measuring more
than 100 cm TTL would be sufficient to attain an asymp-
totic population size that surpasses the initial number of
colonies. Transplanting small-sized colonies would be
effective in attaining population viability (kC1) as long
as the number of transplants is C50.
abc
Fig. 1 Results of the life table response experiment analysis showing
the contribution of each life cycle transition to the population growth
rate of Acropora cervicornis after accounting for the effect of time,
2011–2012 versus 2012–2013, at Canal Luis Pen
˜a(a) and Palomino
(b) and between locations during the 2012–2013 period (c). Positive
and negative values indicate transitions that contribute to and
suppress local population growth, respectively. Note the difference
in scale in the y-axis in (c). S =stasis, G =growth, R =retrogres-
sion; s =small, m =medium, l =large
Coral Reefs
123
Discussion
Spatiotemporal variability in vital rates
Studies directed at understanding the population dynamics
of A. cervicornis have been relatively scarce. Knowlton
et al. (1990) found that the rates of survival and growth of
this coral varied both temporally and spatially. Our results
support these findings. The spatiotemporal variability
observed in this study is, however, somewhat surprising.
Knowlton et al. (1990) attributed most of the mortality they
observed to high predation rates by the snail Coralliophila
spp. and the polychaete Hermodice carunculata. In con-
trast, the 2 yrs in which we measured the vital rates can be
categorized as ‘‘normal’’ in the sense that no major dis-
turbances such as hurricanes, bleaching, predator out-
breaks, or epizootic events directly impacted the two study
sites. Moreover, the conditions at the study localities
remained similar in terms of water quality, wave exposure,
and biotic composition. Thus, this study suggests that even
in the absence of a major disturbance, the vital rates of this
species can be susceptible to minor or moderate local
variations in environmental parameters. The only poten-
tially stressful event that occurred during the study period
was a heavy rainfall event (356 cm, an anomaly of 215 %
in relation to mean monthly rainfall) during June 2013
(Herna
´ndez-Delgado et al. 2014), 4 months before the last
survey. Extreme rainfall events have been associated with
high mortality of scleractinian corals, including A. cervi-
cornis, due in part to a drop in seawater salinity (Morton
2002; Herna
´ndez-Delgado et al. 2014). There is evidence
that a reduction in salinity can limit the capacity of a coral
to survive short-term increases in seawater temperature
(Coles and Jokiel 1978). Low salinity can also impair coral
physiological efficiency (Kerswell and Jones 2003) with
possible costs to colony development. However, we cannot
Fig. 2 Population viability
analysis for Acoprora
cervicornis populations studied
at Canal Luis Pen
˜a (CLP) and
Palomino (PAL). Each
trajectory represents the
cumulative probability of quasi-
extinction over time. Various
trajectories are explored based
on the threshold at which the
population would be considered
functionally (or quasi-)
extinct—from 10 % (a 90 %
reduction) to 25 % of the
population size in 2011
ab
Fig. 3 Estimated population
growth rate for the two
Acropora cervicornis
populations studied as a
function of afragment survival
assuming the observed rates of
fragmentation at Canal Luis
Pen
˜a (CLP; solid line) and
Palomino (PAL; dashed line)
and bprobability of
fragmentation based on
fragment survival of 24 and
32 % at CLP (solid line) and
PAL (dashed line), respectively
Coral Reefs
123
determine the specific causal factors for the significant
decline in survival and growth rates during the period
2012–2013.
Life history trade-offs may explain the apparently high
susceptibility of this species to low/moderate levels of
environmental variation. It is understood that A. cervicor-
nis allocates most of its energetic resources toward rapid
growth at the expense of other biological functions (Palmer
et al. 2010). The resource allocation trade-off hypothesis
postulates that an individual’s energetic budget is finite and
that resources invested in one function cannot be used for
another (Bazzaz et al. 1987). Hence, it is likely that A.
cervicornis’ poor ability to cope with environmental
changes is a result of a resource allocation strategy favor-
ing rapid colony growth over other life history traits
directed to assure colony survival, such as maintenance,
tissue regeneration and repair, and a stronger immune
response (Palmer et al. 2010; Ruiz-Diaz et al. 2013). The
literature on life history trade-offs in corals is very limited;
nonetheless, the immune response of acroporids appears to
be weaker in relation to that of slower-growing massive
corals (Palmer et al. 2010).
Results of the log-linear analysis suggest that spatial
variation in the demography of A. cervicornis may be larger
than temporal variation within the same location. Such dif-
ferences were more evident in medium-sized colonies.
During 2011–2012, the probability of a medium-sized
colonies remaining within its size class or retrogressing into
the smaller size class was noticeably lower and higher,
respectively, at PAL. In contrast, during the second year,
Fig. 4 Simulated trajectories of the Acoprora cervicornis population size
over time at Canal Luis Pen
˜a using different numbers and sizes of
transplanted colonies based on the stochastic mean matrix. Solid line =
initial population size (N
0
=144); closed circles =population trajectory
with no outplants; open circles =small-sized outplants; closed trian-
gles =medium-sized outplants; open triangles =large-sized outplant
Coral Reefs
123
most of the spatial variation in the dynamics of medium-
sized colonies could be attributed to differences in their
probability of advancing to the large size class. It is unclear
why medium-sized colonies were more susceptible than
small-sized colonies and large-sized colonies to spatiotem-
poral variability. Nevertheless, we noticed that at the
beginning of the study, and particularly at CLP, medium-
sized colonies exhibited the highest proportions of dead
tissue. It has been suggested that differences in the ratio of
live to dead tissue can have important implications for the
demographic performance of corals (Meesters et al. 1997;
Oren et al. 1997; Ruiz-Diaz et al. unpublished data). For
instance, there is an inverse relationship between the
potential of a colony to regenerate lost tissue and the ratio
between the size of a lesion (e.g., dead tissue) and size of the
colony (Sousa 1984; Oren et al. 1997; Ruiz-Diaz et al.
unpublished data). Lower regeneration capability not only
limits the growth of a colony but also increases its probability
of dying (Meesters et al. 1997; Hall 2001).
The significant spatiotemporal variability of A. cervi-
cornis demographic rates observed in this study contrasts
with rates for A. palmata in Florida (Vardi et al. 2012).
That study estimated higher and more stable population
growth rates (ks=0.97–1.05) over 6 yrs (with the
exception of a hurricane year, k=0.71), and transition
rates that did not vary considerably among reefs. This
suggests, as previously noted by Williams et al. (2006), that
the demography of A. cervicornis is more dynamic and
variable than that of A. palmata, which is unexpected given
that the two species share similar life history traits. Con-
sidering the differences observed here, although only based
on two sites and 2 yrs, conservation recommendations and
initiatives should be species specific, not generic, when
possible.
Fig. 5 Simulated trajectories of Acoprora cervicornis population size
over time at Palomino using different numbers and sizes of transplanted
colonies based on the stochastic mean matrix. Solid line =initial
population size (N
0
=150); closed circles =population trajectory
with no outplants; open circles =small-sized outplants; closed trian-
gles =medium-sized outplants; open triangles =large-sized outplant
Coral Reefs
123
Population dynamics and viability analysis
During the first year of this study, colony survival varied
between 88 and 89 %. These annual rates are comparable
to those found by Knowlton et al. (1990) during the initial
years of their study. Yet the lack of sexual recruitment, as
well as relatively low rates of colony fragmentation,
resulted in a population decline at both sites (k\1). The
absence of sexually derived recruits was not a surprise, as
this coral is characterized by limited sexual reproduction
(Tunnicliffe 1981; Knowlton et al. 1990). On the other
hand, the rarity of natural fragmentation was unexpected as
it is accepted that colony fragmentation is the main mode
of reproduction in A. cervicornis, even in areas where the
degree of wave exposure is low (Tunnicliffe 1981;
Knowlton et al. 1990). A possible explanation for the low
rates of fragmentation is that current size structures are
below a threshold at which water-related drag forces or
mechanical instability can induce branch breakage (High-
smith 1982). Tunnicliffe (1982) identified 40 cm as the
minimum colony height at which waves can induce branch
breakage in A. cervicornis; at our study sites, only ten
colonies exceeded that height (eight at CLP and two at
PAL). Mean colony height varied between 19.42 cm
(±8.88; SD) at PAL and 23.00 cm (±9.26; SD) at CLP.
During the second year, reductions in population growth of
*24 and 38 % were documented at PAL and CLP, respec-
tively. These lower ks were due not only to insignificant rates of
recruitment but also to the combination of lower colony sur-
vival, little colony growth, and higher levels of colony
shrinkage. As indicated by the life table response analysis, the
lower probability of large-sized colonies surviving and
remaining within the largest size class was the demographic
transition that contributed most to the observed temporal
reduction in ks. Although large-sized colonies survived better
than small-sized colonies- and medium-sized colonies, they
exhibited higher transitions to smaller size classes, particularly
to the smallest size class. Higher proportions of colonies
\250 cm TTL may increase the risk of population extinction.
First, small-sized colonies survive poorly compared with large-
sized colonies. Second, it is recognized that there is a positive
linear relationship between sexual reproductive output and
colony size (Hughes 1984). Third, large-sized colonies tend to
produce a greater number of branches than small-sized colonies
(Mercado-Molina et al. 2015), which is essential for a species
that relies mostly on branch fragmentation for propagation. It
can be argued that populations are currently shifting toward
smaller populations dominated by small-sized colonies, which
are generally characterized by low demographic performance.
This exacerbates the poor prospect of population persistence.
Determining the minimum viable size at which a pop-
ulation is capable of recovering is critical for its conser-
vation. We know that the number of colonies required to
maintain a coral population needs to be large, particularly
for broadcast spawners (Herna
´ndez-Pacheco et al. 2011);
but determining an exact number is not currently possible.
Quasi-extinction thresholds are a way of estimating the
probability of extinction without knowing the critical level
at which a given population becomes functionally extinct.
Four different quasi-extinction thresholds were considered
for the studied populations, and even under the most con-
servative scenario (10 %), the studied populations can
become functionally extinct in B20 yrs. This result pro-
vides quantitative support to the generalized perception
that populations of A. cervicornis continue to decline
rapidly throughout their geographical range (NMFS 2015).
Therefore, populations of A. cervicornis should be of
conservation concern, especially when our analyses did not
consider the consequence of possible extreme events such
as hurricanes or disease outbreaks. The insignificant con-
tribution of asexual reproduction to local population
growth together with the fact that loose fragments of A.
cervicornis survive poorly (Mercado-Molina et al. 2014)
leads us to believe that the studied populations will not
recover or sustain themselves without intervention. This
underscores the need for human involvement to ensure the
viability of A. cervicornis in northeastern Puerto Rico.
Restoration
Outplanting nursery-grown coral fragments (colonies) to
selected sites using stabilization techniques such as epoxy/
cement and cable ties has been considered an effective tool
for restoring depleted populations of Caribbean acroporid
species (Bowden-Kerby 2001; Williams and Miller 2010;
Mercado-Molina et al. 2015). Simulations indicate that
transplanting large fragments would result in higher
asymptotic population sizes. Nevertheless, considering that
coral nurseries are usually initiated with coral fragments
B15 cm TLL (Quinn and Kojis 2006; Herlan and Lirman
2009; Lirman et al. 2010,2014), it would take at least 3 yrs
in a good growth environment for a colony to reach
C250 cm TLL (Lirman et al. 2010,2014; Herna
´ndez-
Delgado et al. 2014). This time lag may be critical for the
persistence of the targeted populations, if, as shown in this
study, populations experience a ‘bad’ year in which
approximately half of the colonies are lost. An alternative
approach is to focus the management plan on transplanting
at least 50 (*33 % of the initial population size in the case
of this study) small (B100 cm TLL) colonies per year.
Results from Mercado-Molina et al. (2015) suggest that
25 cm TLL is an effective initial size for colony trans-
plantation as the coral outplants survive, grow, and produce
branches relatively well. Nursery-reared fragments can
reach this size in less than 5 months (Herna
´ndez-Delgado
et al. 2014).
Coral Reefs
123
Coral conservation initiatives can be costly (Edwards
et al. 2010; Vardi et al. 2012). Yet reef restoration projects
could be both economically and logistically feasible as
long as members of the local community are part of the
effort (Herna
´ndez-Delgado and Suleima
´n-Ramos 2014;
Herna
´ndez-Delgado et al. 2014; Forrester et al. 2014). In
Puerto Rico, for instance, the community-based non-gov-
ernment organization Sociedad Ambiente Marino (SAM)
outplanted over 2000 colonies per year (employing nails
and cable ties) with an initial budget of US$90,000 yr
-1
,
compared with *US$600,000 yr
-1
estimated for a similar
(but not community-based) project in the Florida Keys (see
Vardi et al. 2012). Taking into consideration the total cost
of the SAM operation, the unit cost of a transplant is
estimated at US$45.
In conclusion, this study is based on demographic data
collected for 2 yrs at two localities. It is likely that this data set
has not captured all of the temporal and spatial variability in
vital rates that this species can exhibit. Particularly, the
impacts of hurricanes, mass bleaching, and disease outbreaks
are absent from our analyses because none occurred during the
study period. In spite of these limitations, this study has
identified important aspects of the demography and dynamics
of A. cervicornis. (1) This study highlights significant spa-
tiotemporal variability in vital rates, even in the absence of a
major disturbance. This stresses the apparent susceptibility of
this species to low/moderate levels of environmental vari-
ability. (2) The overall effect of spatial variation on the
demographic fate of A. cervicornis was stronger than the
effect of temporal variation. Accordingly, restoration efforts
should be partitioned among several small projects rather than
allocating all the resources into one site. This action will
enhance the persistence of the species if localized extirpation
occurs due to spatial variability. Such results also imply that
continuous demographic monitoring is essential, as real-time
knowledge of the spatial and temporal variation may be nec-
essary to fine-tune management plans in accordance with the
observed population trends and/or conditions. (3) The
observed contribution of asexual recruitment is meager and
therefore not a significant factor in the population dynamics of
the species. This challenges the generalized but insufficiently
demonstrated belief that under current coral reef conditions,
colony fragmentation is capable of sustaining local popula-
tions (NMFS 2006). (4) Rather, it is variation in the stasis of
large-sized colonies, plus the growth and partial tissue mor-
tality of medium-sized colonies that contributes most to
variations in population growth rates.
The numerical projections performed in this study are
not aimed at predicting what will happen to the population,
but rather at evaluating the possible influences of the
observed spatiotemporal demographic attributes on the
population dynamics of A. cervicornis (Caswell 2001).
Such projections are informative in guiding management
and conservation strategies given that they incorporate the
effects of the prevailing environmental conditions on col-
ony vital rates (Caswell 2001). In this sense, even in the
absence of a major perturbation, population growth rates
are below equilibrium and the population viability analysis
indicates quasi-extinction in less than two decades.
Including the effect of major disturbances in the analysis
will only make more ominous an already dismal scenario
for A. cervicornis. Restoring populations by outplanting
fragments is a feasible strategy, but one that requires sus-
tained human intervention (Mercado-Molina et al. 2015).
Finally, this study raises the first quantitatively supported
red flag with regard to the long-term persistence of A.
cervicornis. Major restoration efforts along the lines pro-
posed in this study are necessary at a regional scale if we
want to reverse the demise of this important Caribbean
coral. We understand, however, that only through the
collective publication of demographic data and analyses
from other Caribbean localities and time periods by other
research groups will we start to gain a comprehensive
appreciation of the spatial and temporal variability in the
demography and dynamics of this species.
Acknowledgments This project was partially funded by UPR Sea
Grant (PD-294), institutional funds of the UPR-RP, UPR Sea Grant
(NOAA award NA10OAR41700062, Project R-92-1-10), the Center
for Applied Tropical Ecology and Conservation (NSF grant HRD
#0734826), Puerto Rico Center for Environmental Neuroscience
(NSF-HRD #1137725), and Idea Wild Foundation. Logistical support
was provided by the SAM. Thanks to all members of SAM.
References
ABRT (Acropora Biological Review Team) (2005) Atlantic Acropora
status review document. Report to National Marine Fisheries
Service, Southeast Regional Office, 152 pp
Aronson RB, Precht WF (2001) White-band disease and the changing
face of Caribbean coral reefs. Hydrobiologia 460:25–38
Bazzaz FA, Chiariello NR, Coley PD, Pitelka LF (1987) Allocating
resources to reproduction and defense. BioScience 1987:58–67
Bowden-Kerby A (2001) Low-tech coral reef restoration methods
modeled after natural fragmentation processes. Bull Mar Sci
69:1–18
Burke L, Reytar K, Spalding M, Perry A (2011) Reefs at risk
revisited. World Resources Institute, Washington, DC
Carpenter K, Abrar M, Aeby G, Aronson RB, Banks S, Bruckner A,
Chiriboga A, Corte
´s J, Delbeek JC, DeVantier L, Graham JE,
Edwards AJ, Fenner D, Guzma
´n HM, Hoeksema BW, Hodgson
G, Johan O, Licuanan WY, Livingstone SR, Lovell ER, Moore
JA, Obura DO, Ochavillo D, Polidoro BA, Precht WF, Quibilan
MC, Reboton C, Richards ZT, Rogers AD, Sanciangco J,
Sheppard A, Sheppard C, Smith J, Stuart S, Turak E, Veron JEN,
Wallace C, Weil E, Wood E (2008) One-third of reef-building
corals face elevated extinction risk from climate change and
local impacts. Science 321:560–563
Caswell H (2001) Matrix population models, 2nd edition. John Wiley
and Sons, Ltd
Coral Reefs
123
Coles SL, Jokiel PL (1978) Synergistic effects of temperature, salinity
and light on the hermatypic coral Montipora verrucosa. Mar Biol
49:187–195
Edwards A, Guest J, Rinkevich B, Omori M, Iwao K, Levy G, Shaish
L (2010) Reef rehabilitation manual. Coral Reef Targeted
Research & Capacity Building for Management Program, St
Lucia, Australia
Fingleton B (1984) Models of category counts. Cambridge University
Press, Cambridge, UK
Forrester GE, Ferguson MA, O’Connell-Rodwell CE, Jarecki LL
(2014) Long-term survival and colony growth of Acropora
palmata fragments transplanted by volunteers for restoration.
Aquat Conserv 24:81–91
Hall VR (2001) The response of Acropora hyacinthus and Montipora
tuberculosa to three different types of colony damage: scraping
injury, tissue mortality and breakage. J Exp Mar Biol Ecol
264:209–223
Herlan J, Lirman D (2009) Development of a coral nursery program
for the threatened coral Acropora cervicornis in Florida. Proc
11th Int Coral Reef Symp 1:1244–1249
Herna
´ndez-Delgado EA, Suleima
´n-Ramos SE (2014) E.S.A. coral
species listing: a roadblock to community-based engagement in
coral reef conservation and rehabilitation across the U.S.
Caribbean? Reef Encounter 29:11–15
Herna
´ndez-Delgado EA, Mercado-Molina AE, Alejandro-Camis PJ,
Candelas-Sa
´nchez F, Fonseca-Miranda JS, Gonza
´lez-Ramos
CM, Guzma
´n-Rodrı
´guez R, Olivo-Maldonado I, Me
`ge P,
Suleima
´n-Ramos SE (2014) Community-based coral reef reha-
bilitation in a changing climate: Lessons learned from hurri-
canes, extreme rainfall, and changing land use impacts. Open J
Ecol 4:918–944
Herna
´ndez-Pacheco R, Herna
´ndez-Delgado EA, Sabat AM (2011)
Demographics of bleaching in a major Caribbean reef-building
coral: Montastraea annularis. Ecosphere 2:art9
Highsmith RC (1982) Reproduction by fragmentation in corals. Mar
Ecol Prog Ser 7:207–226
Hughes TP (1984) Population dynamics based on individual size
rather than age: a general model with a reef coral example. Am
Nat 123:778–795
Kerswell AP, Jones RJ (2003) Effects of hypo-osmosis on the coral
Stylophora pistillata: nature and cause of ‘‘low-salinity bleach-
ing’’. Mar Ecol Prog Ser 253:145–154
Kohler KE, Gill SM (2006) Coral point count with Excel extensions
(CPCe): a visual basic program for the determination of coral
and substrate coverage using random point count methodology.
Comput Geosci 32:1259–1269
Knowlton N, Lang JC, Keller BD (1990) Case study of natural
population collapse: post-hurricane predation on Jamaican
staghorn corals. Smithson Contrib Mar Sci 31:1–25
Linares C, Coma R, Zabala M (2008) Restoration of threatened red
gorgonian populations: an experimental and modeling approach.
Biol Conserv 141:427–437
Lirman D, Thyberg T, Herlan J, Hill C, Young-Lahiff C, Schopmeyer
S, Huntington B, Santos R, Drury C (2010) Propagation of the
threatened staghorn coral Acropora cervicornis: methods to
minimize the impacts of fragment collection and maximize
production. Coral Reefs 29:729–735
Lirman D, Schopmeyer S, Galvan V, Drury C, Baker AC, Baums IB
(2014) Growth dynamics of the threatened Caribbean staghorn
coral Acropora cervicornis: influence of host genotype, symbiont
identity, colony size, and environmental setting. PLoS One
9:e107253
Mercado-Molina AE, Ruiz-Diaz CP, Sabat AM (2014) Survival,
growth, and branch production of unattached fragments of the
threatened hermatypic coral Acropora cervicornis. J Exp Mar
Biol Ecol 457:215–219
Mercado-Molina AE, Ruiz-Diaz CP, Sabat AM (2015) Demographics
and dynamics of two restored populations of the threatened reef
building coral Acropora cervicornis. J Nat Conser 24:17–23
Meesters EH, Wesseling I, Bak RP (1997) Coral colony tissue
damage in six species of reef-building corals: partial mortality in
relation with depth and surface area. J Sea Res 37:131–144
Miller M, Bourque A, Bohnsack J (2002) An analysis of the loss of
acroporid corals at Looe Key, Florida, USA: 1983–2000. Coral
Reefs 21:179–182
Morris WF, Doak DF (2002) Quantitative conservation biology: the
theory and practice of population viability analysis. Sinauer
Associates, Sunderland, MA. 480 pp
Morton B (2002) Effects of extreme rainfall, typhoons and declaration
of marine reserve status on corals beached at Cape d’Aguilar
(1998 and 1999). J Mar Biol Assoc UK 82:729–743
NMFS (National Marine Fisheries Service) (2006) Endangered and
threatened species: final listing determinations for elkhorn coral
and staghorn coral. Federal Register 71:26852–26872
NMFS (National Marine Fisheries Service) (2015) Recovery plan for
elkhorn (Acropora palmata) and staghorn (A. cervicornis) corals.
Prepared by the Acropora Recovery Team for the National
Marine Fisheries Service, Silver Spring, Maryland
Oren U, Benayahu Y, Loya Y (1997) Effect of lesion size and shape
on regeneration of the Red Sea coral Favia favus. Mar Ecol Prog
Ser 146:101–107
Palmer CV, Bythell JC, Willis BL (2010) Levels of immunity
parameters underpin bleaching and disease susceptibility of reef
corals. FASEB J 24:1935–1946
Precht W, Bruckner A, Aronson R, Bruckner R (2002) Endangered
acroporid corals of the Caribbean. Coral Reefs 21:41–42
Quinn NJ, Kojis BL (2006) Evaluating the potential of natural
reproduction and artificial techniques to increase the Acropora
cervicornis populations at Discovery Bay, Jamaica. Rev Biol
Trop 54:105–116
R Development Core Team (2014) R: a language and environment for
statistical computing. R Foundation for Statistical Computing,
Vienna, Austria
Ruiz-Diaz CP, Toledo-Herna
´ndez C, Sabat AM, Marcano M (2013)
Immune response to a pathogen in corals. J Theor Biol 332:141–148
Sousa WP (1984) The role of disturbance in natural communities.
Annu Rev Ecol Syst 15:353–391
Stubben C, Milligan BG (2007) Estimating and analyzing demo-
graphic models using the popbio package in R. J Stat Softw
22:1–23
Tunnicliffe V (1981) Breakage and propagation of the stony coral
Acropora cervicornis. Proc Natl Acad Sci USA 78:2427–2431
Tunnicliffe V (1982) The effects of wave-induced flow on a reef
coral. J Exp Mar Biol Ecol 64:1–10
Vardi T, Williams DE, Sandin SA (2012) Population dynamics of
threatened elkhorn coral in the northern Florida Keys, USA.
Endanger Species Res 19:157–169
Venables WN, Ripley BD (2002) Modern applied statistics with S,
4th edn. Springer, New York
Williams DE, Miller MW (2006) Morphology offers no clues to
asexual vs. sexual origin of small Acropora cervicornis (Scler-
actinia: Acroporidae) colonies. Rev Biol Trop 54:145–151
Williams DE, Miller MW (2010) Stabilization of fragments to
enhance asexual recruitment in Acropora palmata, a threatened
Caribbean coral. Restor Ecol 18:446–451
Williams DE, Miller MW, Kramer KL (2006) Demographic moni-
toring protocols for threatened Caribbean Acropora spp. corals.
NOAA Technical Memorandum NMFS-SEFSC-543. 91 pp
Wirt KE, Hallock P, Palandro D, Lunz KS (2015) Potential habitat of
Acropora spp. on reefs of Florida, Puerto Rico, and the US
Virgin Islands. Global Ecology and Conservation 3:242–255
Coral Reefs
123
