Coral reefs damages caused by hurricanes in the Caribbean and their correlation with the characteristics of hurricanes and reefs.

Abstract

This technical report estimate hurricane damage to reefs in the Caribbean and identify the correlation between hurricane and reef characteristics and the damage caused. The results contribute to understand the severity of hurricane damage and to define the parameters that can trigger a parametric insurance whose funds are used to repair the damage caused to reefs.

Full text

CORAL REEFS DAMAGES
CAUSED BY HURRICANES
IN THE CARIBBEAN
Esmeralda Pérez Cervantes, Fernando Pardo Urrutia, Lorenzo Álvarez Filip,
Fernando Secaira Fajardo, Claudia Ruiz Alvarado, Mariana Alvarez Rocha
2020
Mesoamerian Reef Rescue Initiative (RRI)
Coastal Risk and Reslience Initiative, The Nature Conservancy
AND THEIR CORRELATION
WITH THE CHARACTERISTICS
OF HURRICANES AND REEFS

CORAL REEFS DAMAGES
CAUSED BY HURRICANES
IN THE CARIBBEAN
Esmeralda Pérez Cervantes, Fernando Pardo Urrutia, Lorenzo Álvarez Filip,
Fernando Secaira Fajardo, Claudia Ruiz Alvarado, Mariana Alvarez Rocha
2020
Mesoamerian Reef Rescue Initiative (RRI)
Coastal Risk and Reslience Initiative, The Nature Conservancy
AND THEIR CORRELATION WITH
THE CHARACTERISTICS OF
HURRICANES AND REEFS

AUTHORS
Esmeralda Pérez Cervantes - Consultant
Fernando Pardo Urrutia -Consultant
Lorenzo Álvarez Filip -UNAM- Laboratorio de Biodiversidad
Arrecifal y Conservación
Fernando Secaira Fajardo -The Nature Conservancy - CRR
Claudia Ruiz Alvarado -MAR Fund - RRI
Mariana Alvarez Rocha -Universidad Autónoma de Yucatán
DESIGN
Karla Paola Vazquez Mendoza
SUGGESTED CITATION
Pérez-Cervantes, Esmeralda; Pardo-Urrutia, Fernando; Alvarez-
Filip, Lorenzo; Secaira Fajardo, Fernando; Ruiz Alvarado, Claudia
y Álvarez Rocha, Mariana. 2020. Coral reefs damages caused
by hurricanes in the Caribbean and their correlation with
the characteristics of hurricanes and reefs. Mesoamerican
Reef Rescue Initiative - MAR Fund, UNAM and The Nature
Conservancy.
With the financial support of The Central American Commission
for Environment and Development (CCAD), the German
Government, through KfW and the International Coral Reef
Initiative (ICRI) through the United Nations Environment
Programme (UN/UNEP).
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs

CONTENT
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
1. IMPACT OF HURRICANES ON CORAL REEFS vi
2. METHODOLOGY 4
2.1 ANALYSIS OF THE IMPACT OF A HURRICANE ON REEFS 5
2.2 REEF CHARACTERISTICS. 5
2.2.1 Live Coral Cover 6
2.2.2 Rugosity 6
2.2.3 Reef Type 7
2.2.4 Reef Zone 8
2.2.5 Reef Depth 8
2.2.6 Wind and Wave Exposure 8
2.2.7 Fetch 9
2.2.8 Reef Data 10
2.3 HURRICANE CHARACTERISTICS 12
2.3.1 Maximum Sustained Wind 12
2.3.2 Central Pressure 12
2.3.3 Duration of the Exposition to Hurricane Winds 12
2.3.4 Minimum Distance Between the Site and the Hurricane 13
2.3.5 Wind Speed at Impact 13
2.3.6 Storm Surge 14
2.4 INDICATORS AND RATES TO MEASURE THE DAMAGE CAUSED BY HURRICANES 14
2.5 INFERENTIAL ANALYSIS 15
3. RESULTS 16
3.1 DAMAGE TO REEFS CAUSED BY HURRICANES IN THE CARIBBEAN 17
3.1.1 Change in Live Coral Cover by Initial Condition 17
3.1.2 Change in Live Coral Cover by Wind Speed at Impact Segregated by Initial Cover 20

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
3.2 EXPLORATORY ANALYSIS OF HURRICANE AND REEF CHARACTERISTICS 23
3.2.1 Change in Live Coral Cover by Initial Condition of Wind-Segregated Site at Impact 23
3.2.2 Change in Live Coral Cover by Wind Speed at Impact 28
3.2.3 Change in Live Coral Cover Due to Site Exposure 30
3.2.4 Change in Live Coral Coverage by Duration of Hurricane Impact (number of snapshots) 32
3.2.5 Change in Live Coral Cover by Fetch 34
3.2.6 Change in Live Coral Cover by Maximum Sustained Wind 36
3.2.7 Change in Live Coral Cover by Minimum Distance from the Hurricane Site 38
3.2.8 Change in Live Coral Cover by Reef Type 40
3.2.9 Change in Live Coral Cover by Reef Zone 43
3.2.10 Change in Live Coral Cover by Mean Depth of Site 45
3.2.11 Change in Reef Rugosity by Initial Rugosity Segregated by Maximum Mean Sustained Wind 47
3.2.12 Change in Reef Rugosity by Reef Type 48
3.2.13 Change in Reef Rugosity by Reef Zone 48
3.2.14 Change in Reef Rugosity due to Mean Depth 49
3.2.15 Change in Reef Rugosity by Initial Rugosity Segregated by Exposure 49
3.2.16 Change in Reef Rugosity by Mean Fetch 50
3.2.17 Change in Reef Rugosity Due to Central Pressure 51
3.2.18 Change in Reef Rugosity by Number of Snapshots 51
3.2.19 Change in Reef Rugosity by Minimum Distance from Site 52
3.2.20 Change in Reef Rugosity Due to Storm Surge 52
3.3 VARIABLES WITH STATISTICAL SIGNIFICANCE 53
4. CONCLUSIONS 56
5. REFERENCES 58
APPENDIX 1 61

IMPACT OF
HURRICANES ON
CORAL REEFS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
1

IMPACT OF HURRICANES ON CORAL REEFS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 1
DEAD CORAL
& BRANCHING
RUBBLE
STANDING
DEAD CORAL
HEALTHY
CORAL STATE
DEAD CORAL
& NO
COMPLEXITY
Coral reefs possess high biodiversity and provide numerous environmental benefits and services to
humanity. Their global economic value is approximately thirty-six trillion USD per year (Spalding et. al.,
2017). Climate change, human activities and natural phenomena have aected reefs in the Caribbean,
decreasing coral cover by 80% from 1970 to 2003 (Gardner et al., 2003; Alvarez-Filip et al., 2011).
IMPACT OF HURRICANES ON
CORAL REEFS
Figure 1. Gradual decrease of reef complexity (extracted from Rogers, et. al. 2017).
Hurricanes are natural phenomena that destroy coral colonies and cause severe damage (Hughes,
1994); corals have not been able to recover from such damage due to current conditions that hinder reef
regeneration. The Caribbean is impacted by an average of 6.2 hurricanes per year, therefore many reefs
are frequently impacted (Taylor and Alfaro, 2005). High intensity hurricanes cause an abrupt decrease in
live coral cover, between 15% up to 60% (Gardner et al., 2005 in Figure 2); the aected sites do not show
cover recovery during the following 8 years (Gardner et al., 2005), which is the average return period of a
hurricane. Similarly, rugosity in the Caribbean has declined since 1969 (Alvarez-Filip et al., 2009). Reefs
have lost between 3% and 17% of their rugosity due to the impact of a single hurricane (Alvarez-Filip et al.,
2011 in Figure 3).

IMPACT OF HURRICANES ON CORAL REEFS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
2
24
-60
-50
-40
-30
-20
-10
-30
5
12
0
10
20
30
3
50 70 90 110 130 150 170
36
14
6
5
4
23
10
25
36
15
Year-on-year rate of change
in coral percent cover
Maximum hurricane intensity (knots)
Figure 2. Relationship between annual rate of change of live coral cover and maximum hurricane wind speed (extracted from
Gardner et al., 2005).
The objective of this work is to estimate hurricane damage to reefs in the Caribbean and to identify if
there is a correlation between hurricane and reef characteristics and the damage caused (see Figure 3). The
results will contribute to understand the severity of hurricane damage and to define the parameters that can
trigger a parametric insurance whose funds are used to repair the damage caused to reefs.
% of anual change
Hurricanes impact on structural complexity
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
2
Previous During Not impacted
Figura 3. . Change in rugosity (extracted from Álvarez-Filip, et. al. 2011): Sites were grouped by the time of hurricane impact:
"previous": impacted during the 10 years prior to monitoring and "during": impacts during the period between monitoring, or
"not impacted": sites that were not impacted.

IMPACT OF HURRICANES ON CORAL REEFS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 3
Damage to reefs
caused by
hurricanes
Reef
characteristics
Live coral cover
Rugosity
Reef type
Reef depth
Exposure
Fetch
Loss of live coral cover
Loss of rugosity
CORRELATION
Hurricane
characteristics
Maximum sustained
wind speed (kt)
Central pressure (mb)
Duration of the exposition
to hurricane winds (m)
Minimum distance between
the site and the hurricane (m)
Wind speed at impact (kt)
Storm surge (m)
Figura 4. Scheme of correlation between hurricane damage, reef characteristics and hurricane characteristics (own
elaboration).

IMPACT OF HURRICANES ON CORAL REEFS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 4
METHODOLOGY
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
2

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 5
METHODOLOGY
2.1 ANALYSIS OF THE IMPACT OF A HURRICANE ON
REEFS
The reef characteristics used to quantify hurricane damage are live coral cover (LCC) and reef rugosity.
Both characteristics are key indicators of reef health, they are aected by hurricanes (see Figure 2 and Figure
3) and the majority of sites have live coral cover data for extended periods of time, some for as long as
50 years. This study defines hurricane impact as the dierence between pre-hurricane and post-hurricane
conditions of live coral cover and reef rugosity.
Sites impacted by a single hurricane and with pre-impact and post-impact cover or rugosity data were
used. Reefs are considered to be impacted by a hurricane when they are 100 km or less from the track of a
hurricane (Gardner et. al. 2005; Alvarez-Filip et. al. 2011). To select the sites, a 100 km circumference radius
was drawn around each site and overlapped with the hurricane tracks (see Figure 5).
SITE
Snapshot
each 6 hrs
Hurricane
track
Buer
(100km)
SITE
HURRICANE HURRICANE
Figura 5. .Determinación de sitios impactados con un polígono de 100 km de diámetro alrededor de un arrecife (elaboración
propia). Sitio izquierdo se considera impactado y sitio derecho no impactado.
2.2 REEF CHARACTERISTICS.
The characteristics of the reefs are conditions of the site structure, its health, degree of development,
and its location in relation to its surroundings. In this analysis, the characteristics were chosen considering
that they could be correlated with hurricane damage and that sucient data exist. The characteristics are
as follows:

METHODOLOGY
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
6
VERY GOOD
GOOD
FAIR 10% - 20%
POOR
CRITICAL
2.2.1 Live Coral Cover
Live coral cover is the proportion of live hard
coral area within a reef. It is obtained by dividing
the live coral area by the total area analyzed and is
expressed as a percentage (i.e. 20% live coral cover).
This percentage is independent of sample size so it
applies equally to small sites or sites with large reef
patches. Live coral cover is one of the best indicators
of reef health (Huang et al., 2018) although it does not
include the diversity of species or coral types nor the
degree of reef development. It has been commonly
used in the study, conservation and management of
coral reefs for more than 50 years (Eddy et al., 2018).
Methodologies for assessing live coral cover are
video transects, quadrats, and single transects.
Sites were grouped according to their initial
live coral cover using the classification proposed
by Healthy Reef Initiative (HRI) in their reef health
reports (2008, 2010, 2012, 2014, 2016, and 2018)
(see Table 1).
CATEGORY Cobertura de coral vivo
Tabla 1. Categories of live coral cover (Healthy Reefs
Initiative [HRI], 2012).
0 -
5%
Live coral cover
5% - 10%
20% - 40%
40% - 100%
2.2.2 Rugosity
The rugosity index is a measure of reef complexity
and is the ratio of the length of a reef profile outline
to the linear distance between the start and end
point of the profile. It is estimated using a chain to
measure the corresponding outline over the reef
and it is compared to the straight line distance
between the start and end point (see Figure 6).
Rugosity creates diversity of microhabitats on reefs
and provides shelter for many species in dierent
life cycles, which attracts a high diversity of species.
The friction caused by reef rugosity on waves
significantly decreases their energy (between 40
and 60%, according to Ruiz de Alegria-Arzaburu et
al., 2013), which contributes to coastal protection
(Secaira & Acevedo,2017). Therefore, rugosity is a
key characteristic of reef functionality and suers
severe damage upon hurricane impact. Values close
to 1 indicate low rugosity or complexity and higher
values are more complex, for example 3.
1m 2m
CHAIN
Figura 6. Methodology for the determination of the
rugosity index (taken and modified from Rogers, C. S.
1994).

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 7
METHODOLOGY
2.2.3 Reef Type
Reefs have dierent geological and morphological origins and Lang et. al. (2010) classifies them into:
atoll, barrier, shore, platform, bank, patch and shallow, which are detailed below:
ᛜATOLL
An irregular ovoid reef system surrounded by deep water. The reef crest is along
the windward edge and they have a central lagoon with patch reefs.
ᛜBARRIER:
Characterized by a linear, medium or border reef system with a nearly continuous
intertidal crest. The component reefs are oriented almost parallel to the coast and
are separated by a lagoon.
ᛜSHORE:
They are adjacent to the coast and close to it. They finish seaward, forming a bay or
lagoon.
ᛜPLATFORM:
They are characterized by being large, shallow and irregular. Some have an
intertidal crest, sand cays or islands, and shallow lagoons.
ᛜBANK:
Refers to a linear system of shore reefs with an intertidal crest that is more or less
parallel to the shore.
ᛜPATC H
An isolated and/or merged reef structure that rises above the surrounding seafloor
and independent of a larger reef system.
ᛜSHALLOW:
A shallow submerged reef structure with corals in the bay or lagoon and with no
other corals bordering it.

METHODOLOGY
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
8
2.2.4 Reef Zone
It refers to the location between the open sea and the coast, they are classified into: fore reef, reef crest,
reef lagoon and back reef (Lang, et. al. 2010) (see Figure 7).
CORAL REEF
FOREREEF
CREST
BACKREEF LAGOON BENTHIC
BOTTOM
DEPTH
SEAGRASS
Figura 7. Reef zones: fore reef, crest, back reef and reef lagoon (taken and modified from Navarro, E. E., 2020).
ᛜFORE REEF:
This is the part of the reef exposed to the open
sea.
ᛜREEF CREST:
This is the shallowest zone and sometimes
exposed to the air.
ᛜBACK REEF:
The part of the reef that faces the coast.
ᛜREEF LAGOON:
The shallow area between the open sea and the
coast, delimited by the reef crest.
2.2.5 Reef Depth
The vertical distance between mean sea level and the benthic reef bottom (see Figure 7).
2.2.6 Wind and Wave Exposure
It refers to the exposure of the site in relation to the dominant wind direction, that defines the wave
energy impacting the reef and the ability of the reef to adapt to that energy (Lang et. al., 2010). The dominant
wind in the area of study is the east-west direction, therefore, sites located on the east or windward coasts
are considered exposed to the wind and those located on the west leeward are considered not exposed.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 9
METHODOLOGY
Sites located on the north or
south coasts are considered
"intermediate" (see Figure 8).
2.2.7 Fetch
Fetch is defined as the distance
over the water where the wind can
blow without interruption from a
land feature such as an island or
peninsula and form waves. The
wind can generate waves with
greater energy and height when
there is a greater unobstructed
distance (more fetch) (see Figure
9). To calculate it, we used the
"R" "Waver" package (InVEST
Coastal Vulnerability Model and
Rohweder J. 2008), on the layer
of the American continent (Porto
Tapiquén C. E., 2015) and the
WINDWARD
WINDWARD
WIND & WAVES
DIRECTION
LEEWARD
LEEWARD
WINDWARD
Figura 8. Example of wind and wave exposure on Cozumel Island.
database of the coordinates of the sites analyzed. For the present study, the fetch was estimated for wind
directions coming from north, northeast, east, southeast, south, southwest and west.
Figura 9. Example of fetch calculation. The dots represent the sites and the arrows represent the wind directions: north,
northeast, east, southeast, south, southwest and west.
EXPOSITION+-

METHODOLOGY
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
10
2.2.8 Reef Data
Live coral cover information was found for 414 sites: 48 sites in scientific articles, 356 sites in databases,
5 sites in reports and 5 sites in degree thesis. Rugosity data were found for 44 sites: 29 sites in scientific
articles, 27 sites in databases and 3 sites in reports. The databases used are: Sistema de Información
Arrecifal del Caribe (SIAC) (base of the Laboratorio de Biodiversidad y Conservación Arrecifal of UNAM)
and the database: "Time-series coral-cover data from Hawaii, Florida, Mo'orea, and the Virgin Islands": U.S.
Geological Survey (Guest, J.R, et al. 2018).
For the final analysis, 245 sites were chosen, with time series and coral cover data, and 37 sites with
reef rugosity data. For the selection of these sites, only those that presented impacts by a single hurricane
between two consecutive samplings were chosen in order to establish a cause-eect relationship. Data were
not used for sites that 1) were aected by more than one hurricane between samplings, 2) were not aected
by hurricanes, or 3) the samplings did not indicate when they were conducted. The sites are distributed
across 24 countries in the Caribbean Basin (see Figure 10 and Table 2).
The inter-sampling period of the 245 sites with coral cover data ranged from 40 days to 10 years; 40.3%
of the sites had periods of less than 1 year, 26.5% had periods of 1 to 2 years, 24% had periods of 2 to 3
years, and 9.2% had periods of more than 3 years. The influence of data with a period of more than 3 years
was analyzed because it is likely that in such a long period other factors may cause changes in coral cover,
so the reported change is not attributable to the hurricane. Damage to coral cover was analyzed breaking
it down by period between samplings and excluding data with a period longer than three years and it was
observed that the trend of the impact does not change (see Annex 1). However, there are changes in the
maximum damage values, showing that some reefs continued to deteriorate significantly. This analysis was
only elaborated for coral cover data.
25°N20°N15°N
80°W 70°W 60°W
25°N20°N15°N
800
km
0
Figura 10. Location of monitoring sites obtained from the databases consulted.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 11
METHODOLOGY
Tabla 2. List of the number of sites obtained per country and per year. Countries with ISO 3166-1 alpha-3 code.
COUNTRY
1973
19 74
1978
1979
1980
1981
1982
1983
1985
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
BLZ
4 4 2 2 84 18 68 44 4 18 48
GTM
6 4 5
MEX
5 15 5 15 2 2 1 1 1 1 6 111 63 24 24 48 18 32 16 14 14 2
HND
78 18 16 52 34 12 70
BHS
33 23 34 35 1
BRB
1 2
BMU
2 2 2 1 2
BES
4 4 3 1 2 2 2 6 1 3 4 4 4
COL
422222
CUB
4 2 2 2 8 8
CUW
4 4 3 1 1 4 4 4 4 4
USA
32 31 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32
NI-CS
1 1 1 1
CYM
2 2
VIR
3 6 3 9 16 11 12 4 4 4 4 4 4 4 4 6 7 7 7 8 8 9 9 9 9 9 9 7 7 7 7 7
JAM
7 5 2 1 1 1 2 2
MTQ
2
PRI
2 3 3 2 2
DOM
2 1
NL-BQ2
1 2 2 2 2
NL-BQ3
1 1 1 1 1 1
SLV
2 4 2 2 2
TTO
22221
VEN
4 6 2 2
TOTAL
4 4 7 3 11 3 2 85 9 31 16 27 10 616 32 40 64 67 51 40 39 50 47 40 46 153 272 109 115 209 75 202 106 97 171 40 1 2

METHODOLOGY
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
12
2.3.1 Maximum Sustained Wind
It is the average of the maximum wind speed at
10 meters of altitude sustained for 1 minute. It is
measured in knots (kt).
The maximum speed assigned to the site is
the maximum speed reported at the center of the
hurricane that passed within the 100 km polygon.
2.3.2 Central Pressure
Refers to the barometric pressure at the center
of the hurricane. It is measured in millibars (mb).
This variable is highly correlated with the maximum
sustained wind speed (r= -0.91, for all HURDAT 2
data); the lower the central pressure, the higher the
sustained wind speed.
The central pressure assigned to the site is the
pressure reported at the center of the hurricane that
passed within the 100-km polygon.
2.3.3 Duration of the Exposition to Hurricane
Winds
This characteristic seeks to establish the length of
time the hurricane was impacting a site. To indicate
the duration of a hurricane at the site, a proxy
indicator was used, which is the number of snapshots
occurring within the 100 km polygon. As indicated,
the HURDAT 2 database reports a snapshot every 6
hours, therefore, the number of snapshots reported
within the polygon allows inferring the time the
hurricane was impacting the site.
This proxy (number of snapshots) cannot be
converted into hours of impact because a snapshot
within the polygon can mean that the hurricane was
from few minutes to less than 12 hours (see Figure
11).
Figura 11. Proxy
indicator for
estimating the
duration of the site's
exposure of the site
to the hurricane
(own elaboration).
The elapsed time is
the number of
snapshot of the
hurricane that occur
within a radius of
100 km around the
site (color blue).
2.3 HURRICANE CHARACTERISTICS
Snapshot
not considered
Buer (100km)
HURRICANE
Snapshot
considered
SITE
The National Hurricane Center (NHC) of the National Oceanic and Atmospheric Administration (NOAA)
publishes hurricane characteristics. For this work we used Atlantic hurricane data from 1851 to 2017, the
data are found in the HURDAT 2 database (https://www.nhc.noaa.gov/data/), (Landsea and Franklin, 2013).
Since 2004, data are recorded every 6 hours referring to a particular point in the track of the hurricane
eye; each data series is called “Snapshot". By overlapping the hurricane trajectories on the 100 km radius
polygons around the sites under study, 44 hurricanes were identified (see Figure 5).
The hurricane variables used for this study are as follows:

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 13
METHODOLOGY
2.3.4 Minimum Distance Between the Site and the Hurricane
Refers to the distance between the site and the eye of the hurricane. The minimum distance between the
site and the hurricane is the weighted average of the inverse of the distance from the site to the eye of the
hurricane of each snapshot recorded within the polygon (see Figure 12).
Figura 12. Method for estimating the minimum distance between the hurricane and the site (own elaboration).
Snapshot
SITE
HURRICANE
Distance between
snapshot and
the Site
Buer (100km)
2.3.5 Wind Speed at Impact
Hurricane winds decrease as they move away
from the center of the hurricane. This characteristic
makes it necessary to establish whether the damage
to the reef is correlated to the wind over the site or
to the maximum wind at the center of the hurricane.
To obtain the wind speed over the site, the wind
speed dispersion around each selected hurricane
was calculated. NHC provides the distance from
the center of the hurricane to points located to the
northeast (NE), northwest (NW), southeast (SE)
and southwest (SW) where wind speeds greater
than 64, 50-64 , 34-50 and less than 32 knots occur.
With this information, polygons were generated
around the center of the hurricane for each wind
speed range (see Figure 13). Therefore, the wind
speed at impact is the wind speed range occurring
over the study site.
Figura 13. Speed range of each polygon (own
elaboration).
a b c d
Greater
than 64 50-
64 34-
50 Less
than 34
Greater
than 118 93-
118 63-
93 Less
than 63

METHODOLOGY
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
14
2.3.6 Storm Surge
Storm surge is the rise above sea level due to the action of wind forcing on the water surface (Meza-
Padilla et. al. 2015). Storm surge was obtained from the SURGEDAT database (The World's Storm Surge
Data Center at Louisiana State University). SURGEDAT has data from around the world since 1880 for more
than 700 hurricanes and storms (Needham & Keim, 2012).
2.4 INDICATORS AND RATES TO MEASURE THE
DAMAGE CAUSED BY HURRICANES
There are several ways to measure change before and after an event, as they must consider several factors
to be able to compare data between dierent sites:
1. The change compared to the initial value.
2. The direction of the change, i.e., whether the value increased or decreased.
3. The time elapsed between the pre- and post-event measurements.
We analyzed the advantages and disadvantages of dierent rates for measuring change (see Table 3) and
confirmed the use of the eective and logarithmic rates as the most appropriate for indicating changes in
coral cover and rugosity. Both the logarithmic rate and eective rate are widely used for reef change analysis
(Côté, et. al. 2005, Côté, et. al. 2006, Álvarez-Filip et. al. 2011, Paddack et. al. 2009, Suchley et. al. 2016,
Gardner et. al. 2003, Gardner et. al. 2005 and Grahan & Nash 2013).
NAME FORMULA INTERPRETATION
DOES IT
REFER TO
THE INITIAL
VALUE?
Gross change or ∆V∆V=Vf - V0Total change. No
Gross change rate or o ii= (Vf - V0)/t Average total change per unit of
time. No
Rate of change or pp= (Vf - V0)/V0 Percentage change from initial
sample. Yes
Simple change rate or i´ i´=(Vf - V0)/〖tV0 Percentage change per unit of
time. Yes
Logarithmic rate or δ δ= Total change in logarithms per unit
of time. Yes
Effective rate or i´´ i´´=eδ-1
The way to interpret δ as a
cumulative percentage change
similar to compound interest.
Yes
log(Vf )-log(V0)
t
Tabla 3. Dierent rates for measuring change. Where v0 is its initial value and vf is the final value (coral coverage or rugosity)
and t= is the time elapsed between the measurement of the values V_f y V_0.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 15
METHODOLOGY
2.5 INFERENTIAL ANALYSIS
The logarithmic rate was selected as an indicator of changes in coral cover and rugosity. Exploratory
analysis were then conducted to identify hurricane or reef characteristics (referred to as variables in this
section) that might be correlated or disassociated. This analysis consisted of generating:
a. Box diagrams, which indicate the median and quartiles of data dispersion.
b. Scatter plots, with local regressions (LOESS), which show the confidence interval of the trend found.
The final step was to perform inferential analysis of the variables or combinations of variables identified
in the exploratory analysis. Inferential analysis are used to identify which variables are statistically significant
with rates of change in coral cover and rugosity. The association of the variables with damage to live coral
cover was performed using a linear mixed model. The loss of coral cover is a multicausal phenomenon and
the linear mixed model allows including fixed and random factors, as well as involving dierent spaces and
times (Dicovskiy-Riobóo & Pedroza-Pacheco, 2017). Only those variables with the least amount of missing
data were considered in order to avoid reducing the sample size in the analysis.
The association of the variables with rugosity damage was analyzed using the multiple linear regression
model. Outliers were removed so that the data would reasonably fit the model, also a selection of variables
was made with the help of the AIC criterion (Akaike's information criteria, a measure of the relative quality
of a statistical model), since it is the best criteria for an inferential model when the sample is small and there
are random factors (Graft Acquah, 2010).
Inferential models calculate the P-value of each characteristic, which indicates the robustness of the
correlation. When the P-value of the characteristic is less than 0.05 it indicates that the characteristic has a
statistically significant correlation.
The characteristics or variables with statistical significance are those that we can use with certainty
independent of the data sample analyzed. In other words, we can select a site not included in the sample and
estimate the damage caused by a hurricane based on the variables with statistical significance.

METHODOLOGY
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
16
RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
3

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 17
RESULTS
3.1 DAMAGE TO REEFS CAUSED BY HURRICANES IN
THE CARIBBEAN
Data from 245 sites with coral cover information during the period 1973 to 2017 show that hurricanes
caused severe damage to reefs, which lost up to 20% of the actual live coral cover after a hurricane hit. The
most obvious result of the study is that sites with higher live coral cover suered three times more actual
losses than sites with lower cover (see section 3.1.1). Sites with cover in "good" and "very good" condition
lost an average of 20% and 13% cover respectively. Sites in "critical" and "poor" condition lost an average of
3% and 5% live coral cover. This dierence is due to the fact that sites with lower cover have less coral that
can be destroyed. Another factor determining the severity of coral cover loss is the wind speed at impact, as
the higher the wind speed, the greater the loss in all reef condition categories (see section 3.1.2).
Data at 37 sites during the period 1978 to 2007 indicate that rugosity can decrease up to 20% after a
hurricane when rugosity is greater than 3. Sites with this value are considered highly complex (see section
3.2.11). The analysis also indicated that sites on the back reef lose more rugosity than those on the front reef
(see section 3.2.13).
Section 3.2 presents exploratory results relating hurricane characteristics to live coral cover and
rugosity. Section 3.3. presents a mixed linear explanatory model that integrates the joint eect of hurricane
characteristics with changes in coral loss and rugosity, in order to identify the characteristics that are
statistically significant (most likely to be determinant).
3.1.1 Change in Live Coral Cover by Initial Condition
The initial coral cover is the amount of coral found on the reef before the passage of a hurricane, this
variable in the multiple linear explanatory model was statistically significant in relation to coral loss after the
passage of a hurricane. The change in live coral cover by initial condition was analyzed with the following
rates: gross change, relative change, annual logarithmic rate and eective rate of change.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
18
PERCENT OF CHANGE
Initial condition
-100
0
20
-40
40
80
120
-60
-80
-20
100
60
Percentage of change in LCC by initial condition
% of change in LCC
Critical Poor Fair Good Very good
Figura 15. Percentage change in live coral cover by initial condition.
Figura 14. Gross change in live coral cover by initial condition.
Gross change in LCC by initial condition
Initial condition
-40
-20
-10
-30
0
10
20
30
Critical Poor Fair Good Very good
Gross change in LCC
GROSS CHANGE

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 19
RESULTS
Logarithmic rate of change in LCC by initial condition
Initial condition
-1.5
0.5
1.5
-0.5
-1
1
0
Logarithmic annual rate of LCC
Critical Poor Fair Good Very good
Figura 16. Annual logarithmic rate of change in live coral cover by initial condition.
ANNUAL LOGARITHMIC RATE
Figura 17. Eective annual rate of change in live coral cover by initial condition.
EFFECTIVE ANNUAL RATE
Eective rate of change in LCC by initial condition
Initial condition
-150
-100
-50
0
100
200
Critical Poor Fair Good Very good
150
50
Eective rate of change in LCC (%)

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
20
3.1.2 Change in Live Coral Cover by Wind Speed at Impact Segregated by Initial Cover
Loss of cover increases with increasing wind speed at impact for all reef condition categories. As expected,
the greatest gross loss occurs on reefs with higher cover.
Gross change in LCC
Maximum wind speed at impact (kt)
-40
-30
-20
-10
10
30
20
0
Critical [0%, 5%]
Less than 34 kt Between 34 kt & 50 kt Between 50 kt y 64 kt Greater than 64 kt
Figura 18. Gross change in live coral cover by wind at impact on reefs in critical condition.
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
20

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 21
RESULTS
Gross change in LCC
Maximum wind speed at impact (kt)
-40
-30
-20
-10
10
30
20
0
Poor [5%, 10%]
Less than 34 kt Between 34 kt & 50 kt Between 50 kt & 64 kt Greater than 64 kt
Figura 19. Gross change in live coral cover by wind at impact on reefs in poor condition.
Gross change in LCC
Maximum wind speed at impact (kt)
-40
-30
-20
-10
10
30
Less than 34 kt Between 34 kt & 50 kt Between 50 kt & 64 kt Greater than 64 kt
20
0
Fair [10%, 20%]
Figura 20. Gross change in live coral cover by wind at impact on reefs in regular state.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
22
Gross change in LCC
Maximum wind speed at impact (kt)
-40
-30
-20
-10
10
30
20
0
Good [20%, 40%]
Less than 34 kt Between 34 kt & 50 kt Between 50 kt & 64 kt Greater than 64 kt
Figura 21. Gross change in live coral cover by wind at impact on reefs in good condition.
Gross change of LCC
Maximum wind speed at impact (kt)
-100
-50
50
150
100
0
Very good [40%, 100%]
Less than 34 kt Between 34 kt & 50 kt Between 50 kt & 64 kt Greater than 64 kt
Figura 22. Gross change in live coral cover by wind at impact on reefs in very good condition.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 23
RESULTS
3.2 EXPLORATORY ANALYSIS OF HURRICANE AND
REEF CHARACTERISTICS
After determining the severity of hurricane damage to reefs, we proceeded to explore the relationship
between that damage and the characteristics of the hurricane and the characteristics of the reef. This
exploration uses linear models whose plots show the change in live coral cover or rugosity on the Y-axis, and
reef or hurricane characteristics, or a combination of both, on the X-axis. The exploratory analysis uses data
from all sites since the inclusion of data older than 3 years does not influence trends and extends the sample
size (see Annex 1).
ANALYZED REEF CHARACTERISTICS.
1. Pre-impact live coral cover or initial condition of
the site.
2. Site exposure.
3. Reef type.
4. Reef zone.
5. Average depth of the site.
6. Clear distance in front of the site or mean fetch.
CHARACTERISTICS OF THE HURRICANE IN
RELATION TO THE SITE UNDER ANALYSIS.
7. Maximum sustained wind.
8. Wind over the site or wind at impact.
9. Hurricane duration or number of snapshots.
10. Minimum distance between the eye of the
hurricane and the site.
As explained in the methodology, coral cover indicators and dierent rates were used to show the change
in coral cover: gross change, proportion of change compared to the initial condition, logarithmic rate and
eective rate (see Table 3), so a graph is presented for each indicator. In the case of rugosity, only the
eective rate was used to show the change in rugosity, so a graph is presented for each indicator compared
to the eective rate.
3.2.1 Change in Live Coral Cover by Initial Condition of Wind-Segregated Site at Impact
The analysis shows that corals in good condition sites experience more severe damage than those in
degraded sites. Degraded reefs will always lose less than reefs in good condition. This dierence is most
evident at wind speeds below 34 knots, under which degraded reefs experience very little damage (-3%)
while in good conditions they lose up to 34%.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
24
Initial condition
-40
-30
-20
-10
0
10
20
30
Gross change in LCC
Less than 34 kt
Critical Poor Fair Good Very good
-40
-30
-20
-10
0
10
20
30
Gross change in LCC
Critical Poor Fair Good Very good
-40
-30
-20
-10
0
10
20
30
Gross change in LCC
Critical Poor Fair Good Very good
-40
-30
-20
-10
0
10
20
30
Gross change in LCC
Critical Poor Fair Good Very good
Between 34-50 kt
Between 50-64 kt Greater than 64 kt
Gross change in live coral cover by wind speed at impact
and segregated by initial condition
Initial condition
Initial condition Initial condition
Figura 23. Gross change in live coral cover by initial condition and wind at impact.
GROSS CHANGE

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 25
RESULTS
-100
-80
-60
-40
-20
0
60
100
% change of LCC
Less than 34 kt
Critical Poor Fair Good Very good
% change of LCC
Critical Poor Fair Good Very good
% change of LCC
Critical Poor Fair Good Very good
% change of LCC
Critical Poor Fair Good Very good
Between 34-50 kt
Between 50-64 kt Greater than 64 kt
Percentage of change in LCC by maximum wind speed at impact & initial condition
20
40
80
-100
-80
-60
-40
-20
0
60
100
20
40
80
-100
-80
-60
-40
-20
0
60
100
20
40
80
-100
-80
-60
-40
-20
0
60
100
20
40
80
Initial condition Initial condition
Initial condition Initial condition
Figura 24. Percent of change in live coral cover by initial condition and wind at impact.
PERCENT OF CHANGE
The analysis shows that degraded sites have experienced very severe damage (up to 74%) with wind
speed at impact greater than 64 knots, losing almost all the little cover they had prior to impact. But on
average sites in good condition lose significantly more than others, losing twice as much (37%) as sites in
critical condition (20%), ten times more than in poor condition (2%) and more than in fair condition (28%).
No data were obtained for sites in very good condition aected by impact wind greater than 64 knots.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
26
Logarithmic rate of change in LCC
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
1.0
Logarithmic rate of change in LCC
Logarithmic rate of change in LCC
Logarithmic rate of change in LCC by maximum wind speed at impact & initial condition
Logarithmic rate of change in LCC
-0.5
0
0.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
1.0
-0.5
0
0.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
1.0
-0.5
0
0.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
1.0
-0.5
0.5
0
Less than 34 kt
Critical Poor Fair Good Very good Critical Poor Fair Good Very good
Critical Poor Fair Good Very good Critical Poor Fair Good Very good
Between 34-50 kt
Between 50-64 kt Greater than 64 kt
Initial condition Initial condition
Initial condition Initial condition
Figura 25. Annual logarithmic rate of change in live coral cover by initial condition and wind at impact.
ANNUAL LOGARITHMIC RATE

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 27
RESULTS
Eective rate of change in LCCEective rate of change in LCC
Eective rate of change in LCC
Eective rate of change in LCC by maximum wind speed at impact & initial
Eective rate of change in LCC
-100
-50
0
50
150
100
-100
-50
0
50
150
100
-100
-50
0
50
150
100
-100
-50
0
50
150
100
Less than 34 kt
Critical Poor Fair Good Very good Critical Poor Fair Good Very good
Critical Poor Fair Good Very good Critical Poor Fair Good Very good
Between 34-50 kt
Between 50-64 kt Greater than 64 kt
Initial condition Initial condition
Initial condition Initial condition
Figura 26. Eective annual rate of change in live coral cover by initial condition and wind at impact.
EFFECTIVE ANNUAL RATE

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
28
PERCENTAGE OF CHANGE
3.2.2 Change in Live Coral Cover by Wind Speed at Impact
When grouping all sites without dierentiating the initial condition, the wind speed at impact does not
significantly determine the magnitude of the caused damage. The average loss is 0% and the maximum
value is -13% for winds less than 34 knots; with winds greater than 64 knots the average increases to
-4% and the maximum value to -14%, data that do not show a significant trend.
GROSS CHANGE
Figura 27. Gross change in live coral cover by wind speed at impact.
Gross change in LCC by wind speed at impact
Maximum wind speed at impact (kt)
-40
-20
-10
-30
0
10
20
30
Less than 34 kt Between 34-50 kt Between 50-64 kt Greater than 64 kt
Gross change in LCC
Figura 28. Percent of change in live coral cover by wind speed at impact.
Percentage of change in live coral cover by wind speed at impact
Maximum wind speed (kt)
-100
-60
-40
-80
40
60
100
Less than 34 kt Between 34-50 kt Between 50-64 kt Greater than 64 kt
% of change in LCC
0
20
-20
80

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 29
RESULTS
The average percentage change does increase significantly with wind speed, going from -2% with wind
less than 34 knots to -27% with wind greater than 64 knots. However, the maximum proportional loss is
similar, with -81% for lower winds and -74% for higher winds.
ANNUAL LOGARITHMIC RATE
Figura 29. Annual logarithmic rate of change in live coral cover by wind speed at impact.
Logarithmic rate of change in live coral cover by wind speed at impact
Maximum wind speed at impact (kt)
-4.0
-3.0
-2.5
-3.5
-2.0
-1.5
-1.0
1.5
Less than 34 kt Between 34-50 kt Between 50-64 kt Greater than 64 kt
Logarithmic rate of change in LCC
-0.5
0.5
1.0
0
ANNUAL EFFECTIVE RATE
Figura 30. Eective annual rate of change in live coral cover by wind speed at impact.
Eective rate of change in live coral cover by wind speed at impact
Maximum wind speed at impact (kt)
-150
-100
-50
0
150
Less than 34 kt Between 34-50 kt Between 50-64 kt Greater than 64 kt
Eective rate of change in LCC (%)
50
100
Logarithmic and eective rates express more significant damage with winds greater than 64 knots.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
30
3.2.3 Change in Live Coral Cover Due to Site Exposure
The sites were classified according to their position compared to the dominant wind. Those sites located
at points oriented in both directions were classified as intermediate. The average change does not vary
significantly (0% in windward and intermediate, compared to -2% in leeward). However, the maximum
impact does increase in leeward (-16%) compared to windward (-5%) and intermediate (-7%).
GROSS CHANGE
Figura 31. Gross change in live coral cover by site exposure.
Gross change in live coral cover by site’s exposure
Exposure
-40
-20
-10
-30
0
10
20
30
Windward Intermediate Leeward
Gross change in LCC
-30
-50
Figura 32. Percent of change in live coral cover by site exposure.
Percent of change in LCC by site´s exposure
Exposure
-100
-40
-20
-80
0
20
40
60
Windward Intermediate Leeward
% change in LCC
-60
80
100
120
PERCENTAGE OF CHANGE

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 31
RESULTS
ANNUAL LOGARITHMIC RATE
Figura 33. Annual logarithmic rate of change in live coral cover by site exposure.
Logarithmic rate of change in LCC by exposure from the site
Exposure
-4.0
-2.5
-2.0
-3.5
-1.5
-1.0
-0.5
0
Windward Middle Leeward
Logarithmic rate of change in LCC
-3.0
0.5
1.0
1.5
EFFECTIVE ANNUAL RATE
Figura 34. Eective annual rate of change in live coral cover by site exposure.
Eective rate of change in LCC by exposure from the site
Exposure
-150
0
50
-100
100
150
200
Windward Middle Leeward
Eective rate of change in LCC (%)
-50

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
32
3.2.4 Change in Live Coral Coverage by Duration of Hurricane Impact (number of snapshots)
The analysis shows that there are no significant changes in coral cover related to the duration of the hurricane
at the site. It could be interpreted that the damage is caused in the first hours of the hurricane impact, since the
damage does not increase in the sites impacted for more than 54 hours (9 snapshots).
GROSS CHANGE
PERCENTAGE OF CHANGE
Gross change in live coral cover by number of snapshots
Number of snapshots
-40
-20
-10
-30
0
10
20
30
Gross change in LCC
-30
-50
' % $ + ,
1 2 3 4 5 6 7 8 9 10
Figura 35. Gross change in live coral cover by number of snapshots. NOAA reports one snapshot every 6 hours.
Percent of change in LCC by number of snapshots
Number of snapshots
-100
0
50
100
150
% of change in LCC
-50
1 2 3 4 5 6 7 8 9 10
Figura 36. Percent of change in live coral cover by number of snapshots. NOAA reports one snapshot every 6 hours.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 33
RESULTS
ANNUAL LOGARITHMIC RATE
EFFECTIVE ANNUAL RATE
Logarithmic rate of change in LCC by number of snapshots
Number of snapshots
-4
-2
-1
0
1
Logarithmic rate of change in LCC
-3
2
1 2 3 4 5 6 7 8 9 10
Figura 37. Logarithmic rate in live coral cover by number of snapshots. NOAA reports one snapshot every 6 hours.
Eective rate of change in LCC by number of snapshots
Number of snapshots
-150
-50
0
100
150
Eective rate of change in LCC (%)
-100
200
50
1 2 3 4 5 6 7 8 9 10
Figura 38. Eective annual rate in live coral cover by number of snapshots. NOAA reports one snapshot every 6 hours.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
34
3.2.5 Change in Live Coral Cover by Fetch
The fetch or the degree of wave protection is the distance that the wind can blow into the water without
interruption and form waves, for this analysis the wind direction used was: north, northeast, east, southeast,
south, southwest and west. The analysis does not show a relationship between the fetch distance of the site and
the damage caused by hurricanes, which means that the distance to form waves does not aect the damage
caused by the hurricane.
GROSS CHANGE
PERCENTAGE OF CHANGE
Gross change in LCC by mean fetch
Mean fecth (m)
-50
-30
-20
-10
0
Gross change in LCC
-40
10
20
30
500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000
Figura 39. Gross change in live coral cover per medium fetch.
Percent of change in LCC by mean fetch
Mean fetch (m)
-100
0
50
% of change in LCC
-50
100
150
500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000
Figura 40. Percent of change in live coral cover by mean fetch.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 35
RESULTS
ANNUAL LOGARITHMIC RATE
EFFECTIVE ANNUAL RATE
Logarithmic rate of change in LCC by mean fetch
Mean fetch (m)
-4
-2
-1
0
Logarithmic rate of change in LCC
-3
1
2
500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000
Figura 41. Logarithmic annual rate in live coral cover by mean fetch.
Eective rate of change in LCC by mean fetch
Mean fetch (m)
-150
-50
0
50
Eective rate of change in LCC
-100
150
200
100
500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000
Figura 42. Eective annual rate in live coral cover per mean fetch.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
36
3.2.6 Change in Live Coral Cover by Maximum Sustained Wind
The analysis shows no relationship between maximum sustained wind speed and reef damage.
GROSS CHANGE
Gross change in LCC by maximum sustained wind
Maximum sustained wind (kt)
-50
-30
-20
-10
0
Gross change in LCC
-40
10
20
30
20 40 60 80 100 120 140
Figura 43. Gross change in live coral cover by mean maximum sustained wind.
PERCENTAGE OF CHANGE
Percent of change in LCC by maximum sustained wind
-100
0
50
% of change in LCC
-50
100
150
Maximum sustained wind (kt)
20 40 60 80 100 120 140
Figura 44. Percent of change in live coral cover by mean maximum sustained wind.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 37
RESULTS
ANNUAL LOGARITHMIC RATE
Logarithmic rate of change in LCC by maximum sustained wind
-4
-2
-1
Logarithmic rate of change in LCC
-3
0
1
Maximum sustained wind (kt)
2
20 40 60 80 100 120 140
Figura 45. Logarithmic annual rate in live coral cover by mean maximum sustained wind.
EFFECTIVE ANNUAL RATE
Eective rate of change in LCC by maximum sustained wind
-150
-50
0
Eective rate of change in LCC (%)
-100
50
100
Maximum sustained wind (kt)
200
150
20 40 60 80 100 120 140
Figura 46. Eective annual rate in live coral cover by mean maximum sustained wind.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
38
3.2.7 Change in Live Coral Cover by Minimum Distance from the Hurricane Site
The analysis does not show a relationship between the distance between the site and the eye of the hurricane,
nor the damage caused to the reef.
GROSS CHANGE
Gross change in live coral cover by minimum distance
Gross change in LCC
Minimum distance (m)
-50
-30
-20
-10
0
-40
10
20
30
20000 40000 60000 80000 100000 1200000
Figura 47. Gross change in live coral cover by minimum distance from site.
PERCENTAGE OF CHANGE
Percent of change in live coral cover by minimum distance
-50
0
% of change in LCC
-100
50
100
Minimum distance (m)
150
20000 40000 60000 80000 100000 1200000
Figura 48. Percent change in live coral cover by minimum distance from site.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 39
RESULTS
ANNUAL LOGARITHMIC RATE
Logarithmic rate of change in live coral cover by minimum distance
-4
-3
-2
Logarithmic rate of change in LCC
-1
0
Minimum distance (m)
1
2
20000 40000 60000 80000 100000 1200000
Figura 49. Logarithmic annual rate in live coral cover by minimum distance from site.
EFFECTIVE ANNUAL RATE
Eective rate of change in live coral cover by minimum distance
-150
-100
-50
Eective rate of change in LCC (%)
0
50
Minimum distance (m)
150
100
200
20000 40000 60000 80000 100000 1200000
Figura 50. Eective annual rate in live coral cover by minimum distance from site.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
40
3.2.8 Change in Live Coral Cover by Reef Type
Analysis shows that atolls experience greater damage than other reef types.
GROSS CHANGE
Reef type
-50
-20
-10
-30
10
20
30
Atoll Bank Barrier Fringing Patch
-40
Gross change in live coral cover by reef type
Gross change in LCC
0
Figura 51. Gross change in live coral cover by reef type.
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
40

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 41
RESULTS
PERCENTAGE OF CHANGE
Reef type
-100
-40
-20
-60
40
80
120
Atoll Bank Barrier Fringing Patch
-80
Percent of change in LCC by reef type
% of change in LCC
0
100
60
20
Figura 52. Percent of change in live coral cover by reef type.
ANNUAL LOGARITHMIC RATE
Reef Type
-4
-2.5
-2
-3
-0.5
0.50
1.5
Atoll Bank Barrier Patch Parche
-3.5
Logarithmic rate of change in LCC by reef type
Logarithmic rate of change in LCC
-1.5
1
0
-1
Figura 53. Logarithmic annual rate in live coral cover by reef type.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
42
EFFECTIVE ANNUAL RATE
Reef Type
-150
0
50
-50
200
Atoll Bank Barrier Fringing Patch
-100
Eective rate of change in LCC by reef type
Eective rate of change in LCC (%)
100
150
Figura 54. Eective annual rate in live coral cover by reef type.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 43
RESULTS
3.2.9 Change in Live Coral Cover by Reef Zone
Analysis does not show a relationship between reef zone and reef damage.
GROSS CHANGE
Reef zone
-50
-20
0
-30
30
Crest Force Lagoon Back
-40
Gross change in LCC by reef zone
Gross change in LCC
10
20
-10
Figura 55. Gross change in live coral cover by reef zone.
PERCENTAGE OF CHANGE
Reef zone
Crest Fore Lagoon Back
Percent of change in LCC by reef zone
-100
-40
-20
-60
40
80
120
-80
% of change in LCC
0
100
60
20
Figura 56. Percent of change in live coral cover by reef zone.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
44
ANNUAL LOGARITHMIC RATE
Reef zone
Crest Fore Lagoon Back
Logarithmic rate of change in LCC by reef zone
Logarithmic rate of change in LCC
-4
-2.5
-2
-3
-0.5
0.5
1.5
-3.5
-1.5
1
0
-1
Figura 57. Logarithmic annual rate in live coral cover by reef zone.
EFFECTIVE ANNUAL RATE
Reef zone
Crest Fore Lagoon Back
Eective rate of change in LCC by reef zone
Eective rate of change in LCC (%)
-150
0
50
-50
200
-100
100
150
Figura 58. Eective annual rate in live coral cover by reef zone.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 45
RESULTS
3.2.10 Change in Live Coral Cover by Mean Depth of Site
The analysis shows that the deeper the depth, the greater the damage to the reefs, although the dierence
is not significant. The average damage changes from 0% in reefs with depths less than 10 meters, to 5% loss
with depths greater than 31 meters.
GROSS CHANGE
Figura 59. Gross change in live coral cover by mean depth.
Reef depth (m)
-50
-20
-10
-40
0
10
20
30
Gross change in LCC
-30
0 5 10 15 20 25 30 35 40 45
Gross change in LCC by reef depth
PERCENTAGE OF CHANGE
Figura 60. Percentage of change in live coral cover by mean depth.
Reef depth (m)
% of change in LCC
-100
0
50
100
150
-50
0 5 10 15 20 25 30 35 40 45
Percent of change in LCC by reef depth

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
46
ANNUAL LOGARITHMIC RATE
Figura 61. Annual logarithmic rate in live coral cover by mean depth.
Reef depth (m)
Logarithmic rate of change in LCC
-4
-2
-1
1
1
-3
0
-1
-3
-2
-4
0 5 10 15 20 25 30 35 40 45
Logarithmic rate of change in LCC by reef depth
EFFECTIVE ANNUAL RATE
Figura 62. Eective annual rate in live coral cover by mean depth.
Reef depth (m)
Eective rate of change in LCC
-4
-2
-1
1
1
-3
0
-1
-3
-2
-4
0 5 10 15 20 25 30 35 40 45
Eective rate of change in LCC by reef depth

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 47
RESULTS
3.2.11 Change in Reef Rugosity by Initial Rugosity Segregated by Maximum Mean Sustained
Wind
The analysis of hurricane damage to reef rugosity used data from 37 sites due to limited data availability.
This contrasts with the coral cover damage analysis, that used data from 245 sites. Therefore, the information
presented here describes the damage, but does not allow us to make statistical projections.
The multiple linear explanatory model analysis indicates that hurricanes cause greater damage when the
pre-impact rugosity is greater, since it is the variable with the greatest statistically significant dierence.
The analysis also indicates that there is a wide variability in damage (see Figure 63), where the gray area
indicates the dispersion of data, and the blue line indicates the average damage.
EFFECTIVE ANNUAL RATE
Figura 63. Eective annual rate in rugosity change by initial rugosity and mean maximum sustained wind (kt).
3
Eective rate of change in rugosity (%) vs initial rugosity
Eective rate of change in rugosity (%)
Initial rugosity
-40
-20
0
1 2

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
48
Figura 64. Eective annual rate of change in rugosity by reef type.
Eective rate of change in rugosity (%)
Eective rate of change in rugosity (%) vs reef type
-40
-20
0
Barrier Fringing
Reef type
3.2.12 Change in Reef Rugosity by Reef Type
EFFECTIVE ANNUAL RATE
3.2.13 Change in Reef Rugosity by Reef Zone
EFFECTIVE ANNUAL RATE
Figura 65. Eective annual rate of change in rugosity by reef zone.
Eective rate of change in rugosity (%)
Eective rate of change in rugosity (%) vs reef zone
-40
-20
0
Crest Front Back
Reef zone

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 49
RESULTS
3.2.14 Change in Reef Rugosity due to Mean Depth
EFFECTIVE ANNUAL RATE
Figura 66. Eective annual rate of change in rugosity by mean depth.
10
Eective rate of change in rugosity (%)
vs reef depth (m)
Eective rate of change in rugosity (%)
Reef depth (m)
-30
0
30
25 57.5 12.5
3.2.15 Change in Reef Rugosity by Initial Rugosity Segregated by Exposure
The analysis indicates that windward exposed sites suered more damage than leeward sites. It also
indicates that windward and higher rugosity index sites experience more damage.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
50
2
Sotavento Barlovento
Eective rate of change in rugosity (%)
vs initial rugosity by exposition
Eective rate of change in rugosity (%)
Initial rugosity
25
0
25
1.5 2.0 3
2.5
-50
1
Figura 67. Eective annual rate of change in rugosity per initial rugosity segregated by site exposure.
3.2.16 Change in Reef Rugosity by Mean Fetch
EFFECTIVE ANNUAL RATE
1000
Mean fetch (km)
-40
1500
-20
0
500 2000
Eective annual rate change in rugosity (%)
vs east/southeast/south mean fetch (km)
Eective rate of change in rugosity (%)
Figura 68. Eective annual rate of change in rugosity per average fetch.
EFFECTIVE ANNUAL RATE

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 51
RESULTS
3.2.17 Change in Reef Rugosity Due to Central Pressure
EFFECTIVE ANNUAL RATE
980
Eective rate of change in rugosity (%) (kt)
vs central pressure (mb)
Eective rate of change in rugosity (%)
Central pressure (mb)
-30
0
30
920 940 960 1000
Figura 69. Eective annual rate of change in rugosity by mean central pressure (mb).
3.2.18 Change in Reef Rugosity by Number of Snapshots
EFFECTIVE ANNUAL RATE
2.5
Eective rate of change in rugosity (%) (kt)
vs number of snapshots to hurricane impact
Eective rate of change in rugosity (%)
Number of snapshots to hurricane impact
-30
0
30
1.0 1.5 2.0 3.0
Figura 70. Eective annual rate of change in rugosity per number of snapshots.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
52
3.2.19 Change in Reef Rugosity by Minimum Distance from Site
EFFECTIVE ANNUAL RATE
Eective rate of change in rugosity (%) (kt)
vs minimum distance (km)
Eective rate of change in rugosity (%)
Minimum distance (km)
-30
0
30
025 50 75
Figura 71. Eective annual rate of change in rugosity by minimum distance to site (km).
3.2.20 Change in Reef Rugosity Due to Storm Surge
EFFECTIVE ANNUAL RATE
3.5
Eective rate of change in rugosity (%) (kt)
vs mean surge imputation (m)
Eective rate of change in rugosity (%)
Mean surge imputation (m)
-30
0
30
1.5 2.5 4.5 5.5
Figura 72. Eective annual rate of change in rugosity per mean surge per imputation (m).

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 53
RESULTS
2
Front
2.5
Back
(n=39)
Rate of change in rugosity vs initial rugosity and reef zone
Rate of change in rugosity
Initial rugosity
-0.4
-0.2
0
1.5
Reef zone
0.2
3.3 VARIABLES WITH STATISTICAL SIGNIFICANCE
After quantifying the damage caused by hurricanes to reefs and exploring the correlation between
damage and the characteristics of hurricanes and reefs, we proceeded to identify the existence of a statistical
correlation through inferential analysis. In addition, the eective rate of change was used for the model
analysis, since it is easily interpreted as a cumulative percentage change (%) and refers to the initial value
(see Table 4).
We selected eleven variables with the highest apparent correlation to facilitate the application of the
model:
1. Daily logarithmic rate of change of coral cover
(Y).
2. Maximum wind speed at impact kt
(categorical).
3. Percent of coral cover initial sample.
4. Interaction between maximum impact wind
speed kt, and percentage of coral cover in the
initial sample.
5. Number of snapshots at impact.
6. Mean central pressure mb.
Figura 73. Annual logarithmic rate of change vs. mean rugosity initial sample profiled by reef zone, where the back reef has
greater loss after hurricane passage.

RESULTS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
54
VARIABLE VALUE STD.
ERROR DF T-VALUE P-VALUE
Percentage of coral coverage initial sample -0.02 0.01 298.00 -3.90 0.0001
Maximum wind speed at impact kt cat.C: initial sample
of coral coverage percentage -0.02 0.01 298.00 -2.67 0.008
Windward Exposure 0.17 0.07 298.00 2.59 0.0101
Maximum wind speed at impact kt cat.Q: initial sample
of coral coverage percentage -0.02 0.01 298.00 -1.86 0.0642
Maximum wind speed at impact kt cat.L: initial sample
of coral coverage percentage -0.02 0.01 298.00 -1.60 0.1116
Intermediate Exposure 0.21 0.15 298.00 1.43 0.1528
Number of snapshots impact 0.03 0.02 298.00 1.23 0.2213
Intermediate Exposure: mean fetch km 0.00 0.00 298.00 -1.15 0.2526
Windward exposure: mean fetch km 0.00 0.00 298.00 -1.02 0.3078
Medium fetch km 0.00 0.00 298.00 0.98 0.3257
Maximum wind speed at impact kt cat.L -0.26 0.26 298.00 -0.98 0.3261
Maximum wind speed at impact kt cat.Q -0.11 0.18 298.00 -0.62 0.5372
Maximum wind speed at impact kt cat.C -0.05 0.10 298.00 -0.51 0.6071
Mean central pressure mb 0.00 0.01 298.00 -0.43 0.6643
(Itntercep) 3.73 8.94 298.00 0.42 0.677
Mean maximum sustained wind kt 0.00 0.01 298.00 -0.40 0.6865
Difference final sample hurricane output days 0.00 0.00 298.00 -0.08 0.9343
Tabla 4. Linear mixed explanatory model of coral cover. Itntercept=intercept, value=value of , std error=standard error,
df=degrees of freedom, t-value=value of t, p-value=value of p to accept or reject the null hypothesis. Statistical correlation
exists if the p-value is less than 0.05.
7. Mean maximum sustained wind kt.
8. Exposure.
9. Average fetch km.
10. Interaction between exposure and mean fetch
km.
11. Dierence between final sample and hurricane
outflow (in days).
The inferential analysis indicates the variables
most correlated with the loss of coral cover after a
hurricane (see Table 4). Variables with a p-value of
less than 0.05 are statistically significant and are:
1. Percentage of initial sample coral coverage.
2. Windward exposure.
3. Maximum wind velocity impact with
percentage of coral cover of initial sample.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 55
RESULTS
Tabla 5. Multiple linear explanatory model. Itntercep= intercept, estimate=estimate, std error=standard error, t-value=t-value,
Pr (>|T|) = value to accept or reject the null hypothesis. Statistical correlation exists if the value is less than 0.05.
VARIABLE ESTIMATE STD.
ERROR T VALUE PR(>|T|)
Initial sample average rugosity -0.094 0.029 -3.216 0.003
(Intercept) 0.144 0.053 2.715 0.011
Back reef zone -0.038 0.023 -1.648 0.109
However, after the selection of variables with sucient data, only the mean rugosity in the initial sample
and the reef zone remained. The variable with statistical significance is mean initial rugosity (see Table 5).
The variables included for the final rugosity
damage model were:
4. Mean maximum sustained wind kt.
5. Initial sample mean rugosity.
6. Dierence final sample hurricane outflow years.
7. Exposure.
8. Interaction between exposure and mean rugosity
in the initial sample.
9. Reef zone.

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 56
RESULTS
CONCLUSIONS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
4

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 57
CONCLUSIONS
Hurricanes are natural phenomena that,
depending on their characteristics, can
cause damage to coral reefs. In the case of
the Caribbean Basin, hurricanes have been
identified as contributing to the loss of coral
cover since 1980 (Gardner et al., 2003). Studies
such as Gardner's in 2005 indicate that reefs
that have been impacted by hurricanes have
faster rates of coral loss than those that have
not been impacted (Gardner et al., 2005).
The analysis presented indicates that wind
speed at impact was one of the variables most
closely related to the damage that a hurricane can
cause as it passes over a reef. Another variable
related to damage was the initial condition of
the site, since sites with greater coral cover or
greater rugosity had more damage than those
with little cover and little rugosity. Finally, site
exposure was also a significant variable, where
sites with less exposure had more damage.
Our study provides new information that can
help reef managers estimate the damage that a
hurricane can cause on a reef. Reefs, which act
as protectors of the coast to the action of storms
and waves, are currently under constant threat
and not only hurricanes cause impact on reefs,
threats such as climate change, acidification of
the seas, diseases, among others are added and
are the cause of its current decline, so knowing
the damage a hurricane can cause to reefs can
help to be prepared for a possible impact on
sites with high coral cover and high rugosity.
©JuanCarlosHuitrón

CONCLUSIONS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
58
REFERENCES
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
5

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 59
REFERENCES
Álvarez-Filip, L., Dulvy, N. K., Gill, J. A., Côté, I. M., & Watkinson, A.
R. (2009). Flattening of Caribbean coral reefs: region-wide declines in
architectural complexity. Biological Sciences, 276 (1669), 3019-3025.
Álvarez-Filip, L., Gill, J. A., Dulvy, N. K., Perry, A. L., Watkinson, A.
R., & Côté, I. M. (2011). Drivers of region-wide declines in architectural
complexity on Caribbean reefs. Coral Reefs, 30(4), 1051- 1060.
Côté, I. M., Gill, J. A., Gardner, T. A., & Watkinson, A. R. (2005).
Measuring coral reef decline through meta-analyses. Philosophical
Transactions of the Royal Society of London B: Biological Sciences,
360(1454), 385-395.
Côté, I., Gardner, T., Gill, J., Hutchinson, D., & Watkinson, A.
(2006). New approaches to estimating recent ecological changes on
coral reefs. In Coral reef conservation. Cambridge University Press.
Dicovskiy-Riobóo, L. M. & Pedroza-Pacheco, M. E. 2017. Modelos
lineales generales y mixtos en la caracterización de la variable
calificación, ingeniería agroindustrial, Uni-norte. Nexo, 30 (2), 84- 95.
Eddy, T. D., W. W. L. Cheung & Bruno, J. F. 2018. Historical baselines
of coral cover on tropical reefs as estimated by expert opinion. PeerJ 6:
01-16. DOI 10.7717/peerj.4308
Gardner, T. A., Cote, I. M., Gill, J. A., Grant, A., & Watkinson, A.
R. (2005). Hurricanes and Caribbean coral reefs: impacts, recovery
patterns, and role in longterm decline. Ecology, 86(1), 174-184.
Gardner, T. A., Côté, I. M., Gill, J. A., Grant, A., & Watkinson, A. R.
(2003). Long-term region-wide declines in Caribbean corals. Science,
301(5635), 958-960.
Graham, N. A. J., & Nash, K. L. (2013). The importance of structural
complexity in coral reef ecosystems. Coral Reefs, 32(2), 315-326.
Graft Acquah, H. (2010). Comparison of Akaike information
criterion (AIC) and Bayesian information criterion (BIC) in selection
of an asymmetric price relationship. Journal of Development and
Agricultural Economics, 2(1), 1-6.
Huang, R., K. Yu, Y. Wang, W, Wang, L. Mu & J. Wang. 2018.
Method to design a live coral cover sensitive index for multispectral
satellite images. Optics express, 26 (10), 374-397.
Hughes, T. P. (1994). Catastrophes, phase shifts, and large-scale
degradation of a Caribbean coral reef. Science, 265(5178), 1547-1551.
Healthy Reefs Initiative (2012). Reporte de la Salud Ecológica del
Arrecife Mesoamericano. Una Evaluación de la Salud del Ecosistema.
https://www.healthyreefs.org/cms/wp-content/uploads/2012/12/
Reporte-2012.pdf
InVEST Coastal Vulnerability Model. http://data.
naturalcapitalproject.org/invest-releases/documentation/2 2 0/
coastal vulnerability.html
Landsea, C. W. & Franklin, J. L. (2013). Atlantic Hurricane Database
Uncertainty and Presentation of a New Database Format. Monthly
weather review, 141, 3576-3592.
Lang, J. C., Marks, K. W., Kramer, P. A., Kramer, P. R., & Ginsburg, R.
N. (2010). AGRRA protocols version 5.4. Atlantic and Gulf Rapid Reef
Assessment Program, Florida, USA.
Navarro, E. E. (2020). Efectos de los cambios de cobertura vegetal
y de uso de suelo sobre la cobertura de coral en los arrecifes del Caribe
Mexicano. Tesis maestría. UNAM
Needham, H.F. & Keim, B. D. (2012). A Storm Surge Database for
the U.S. Gulf Coast. International Journal of Climatology, 32 (14), 2108-
2123. DOI: 10.1002/joc.2425.
Meza-Padilla, R., Appendini, C.M., Pedrozo-Acuña, A. & González-
Villarreal, F. (2015). Evaluación de la marea de tormenta en sitios con
escasez de datos: río Pánuco, México. Ribagua, 2, 61-70, DOI: 10.1016/j.
riba.2015.09.001
Paddack, M. J., Reynolds, J. D., Aguilar, C., Appeldoorn, R. S., Beets,
J., Burkett, E. W., ... & Forrester, G. E. (2009). Recent region-wide
declines in Caribbean reef fish abundance. Current Biology, 19(7), 590-
595
Porto Tapiquén C. E. (2015). América. Orogénesis Soluciones
Geográficas. Porlamar, Basado en capas de Enviromental Systems
Research Institute (ESRI). Distribución Gratuita. Venezuela.
Descargado desde http://tapiquen-sig.jimdo.com.
R Development Core Team. 2009. R: a language and environment
for statistical computing. R Foundation for Statistical Computing,
Vienna, Austria. www.r-project.org
Rogers, A., J. L. Blanchard & P. J. Mumby. (2017). Fisheries
productivity under progressive coral reef degradation. Journal of
Applied Ecology, 55, 1041–1049.

REFERENCES
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
60
Ruiz de Alegria-Arzaburu, A., Mariño-Tapia, I., Enriquez, C., Silva, R.
& Gonzalez-Leija, M. (2013). The role of fringing coral reefs on beach
morphodynamics. Geomorphology, 198, 69-83.
Secaira, F. & César, A. (2017). Importancia de los arrecifes y dunas
en la protección de la costa. Serie técnica. El papel de los sistemas
naturales en la dinámica costera en el caribe mexicano y el impacto de las
actividades humanas en su condición actual. The Nature Conservancy,
México
Rohweder, J., Rogala, J. T., Johnson, B. L., Anderson, D., Clark, S.,
Chamberlin, F., Potter, D. & Runyon, K. (2008). Application of Wind Fetch
and Wave Models for Habitat Rehabilitation and Enhancement Projects.
USGS report. http://www.umesc.usgs.gov/management/dss/wind fetch
wave models.html
Spalding, M., Burke, L., Wood, S. A., Ashpole, J., Hutchison, J., & Zu
Ermgassen, P. (2017). Mapping the global value and distribution of coral
reef tourism. Marine Policy, 82, 104-113.
Suchley, A., McField, M. D., & Alvarez-Filip, L. (2016). Rapidly
increasing macroalgal cover not related to herbivorous fishes on
Mesoamerican reefs. PeerJ, 4, e2084.
Taylor, H. T., Ward, B., Willis, M., & Zaleski, W. (2012). The Sar-
Simpson Hurricane Wind Scale. HYPERLINK "https://www.nhc.noaa.
gov/pdf/sshws.pdf" https://www.nhc.noaa.gov/pdf/sshws.pdf
Taylor, M. A., & Alfaro, E. J. (2005). Climate of Central America and
the Caribbean. In: J. E. Oliver (editor) Encyclopedia of World Climatology.
Netherlands: Springer,p. 183-188.
WWF. 2006. Mejores prácticas de pesca en arrecifes coralinos.
Guía para la colecta de información que apoye el Manejo de Pesquerías
Basado en Ecosistemas. WWF México/Centroamérica. Pp. 81
APPENDIX

CONCLUSIONS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
61
1
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
APPENDIX

APPENDIX
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
62
IMPACT OF THE RESAMPLING PERIOD ON THE RESULTS.
When the resampling periods are very long (longer than 3 years, for example), coral cover can be
influenced by other phenomena or events dierent from the hurricane, aecting the data and conclusions
we can obtain. In this study, we have 72 resamples with resampling periods between 3 and up to 10 years
and 382 with shorter periods. To determine whether the inclusion of these 72 resamples influences the
results, we generated the results with and without these data. The results show the same trends; therefore,
this study used all 454 resamples.
GROSS CHANGE
Figura 1. Gross change by initial condition, excluding data with a period between samples of more than 3 years. Number of
resamples: 382.
Gross change in LCC by initial condition
Initial condition
-40
-20
-10
-30
0
10
20
30
Critical Poor Fair Good Very good
Gross change in LCC
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
62

Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs 63
APPENDIX
Gross change in LCC
Initial condition
-50
-20
-10
-30
0
10
20
30
Critical Poor Fair Good Very good
-40
Gross change in live coral cover by initial condition
Figura 2. Gross change by initial condition. Number of resamples: 455.
Initial condition
-10
50
100
0
150
200
250
Critical Poor Fair Good Very good
-50
Percentage of change in live coral cover by initial condition
% of change in LCC
Figura 3. Percent change from initial condition, segregated by initial condition category, excluding data from sites
with more than 3 years between samples. Number of resamples: 382.
PERCENTAGE OF CHANGE

APPENDIX
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
64
Initial condition
-100
0
20
-40
40
80
120
-60
-80
-20
100
60
Percentage of change in LCC by initial condition
% of change in LCC
Critical Poor Fair Good Very good
Figura 4. Percent change from initial condition, segregated by initial condition category, including data from sites
with more than 3 years between samplings. Number of resamples: 455.
CORAL REEFS DAMAGES CAUSED BY HURRICANES IN THE CARIBBEAN
AND THEIR CORRELATION WITH THE CHARACTERISTICS OF HURRICANES AND REEFS
Esmeralda Pérez Cervantes, Fernando Pardo Urrutia, Lorenzo Álvarez Filip,
Fernando Secaira Fajardo, Claudia Ruiz Alvarado, Mariana Alvarez Rocha
2020
Mesoamerian Reef Rescue Initiative (RRI)
Coastal Risk and Reslience Initiative, The Nature Conservancy

CONCLUSIONS
Coral reefs damages caused by hurricanes in the Caribbean and
their correlation with the characteristics of hurricanes and reefs
65
CORAL REEFS DAMAGES CAUSED BY HURRICANES IN THE CARIBBEAN
AND THEIR CORRELATION WITH THE CHARACTERISTICS OF HURRICANES AND REEFS
Esmeralda Pérez Cervantes, Fernando Pardo Urrutia, Lorenzo Álvarez Filip,
Fernando Secaira Fajardo, Claudia Ruiz Alvarado, Mariana Alvarez Rocha
2020
Mesoamerian Reef Rescue Initiative (RRI)
Coastal Risk and Reslience Initiative, The Nature Conservancy