Content uploaded by Annabelle Constance
Author content
All content in this area was uploaded by Annabelle Constance on Jun 08, 2021
Content may be subject to copyright.
Contents lists available at ScienceDirect
Global Ecology and Conservation
journal homepage: www.elsevier.com/locate/gecco
Original Research Article
Extent change of protected mangrove forest and its relation to
wave power exposure on Aldabra Atoll
Annabelle Constance
a,b,⁎
, Paul J. Haverkamp
a
, Nancy Bunbury
b,c
,
Gabriela Schaepman-Strub
a
a
Department of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland
b
Seychelles Islands Foundation, P.O. Box 853, Mahé, Victoria, Seychelles
c
Centre for Ecology and Conservation, University of Exeter, Cornwall Campus, Penryn TR10 9FE, UK
article info
Article history:
Received 21 October 2020
Received in revised form 22 March 2021
Accepted 29 March 2021
Keywords:
Mangrove extent
Wave power
Conservation
Seychelles
Islands
Landsat
abstract
Mangrove forests, vital for the conservation of biodiversity, protection of coastlines, and
carbon capture, are decreasing globally at a rate higher than most other tropical forests. They
are threatened by sea level rise, drought and storm surge, especially on low-lying islands
where forests are directly exposed to the elements and have limited land area. We in-
vestigated changes in the spatial extent of mangroves on Aldabra Atoll, Seychelles – a pro-
tected area without direct human pressures, over 21 years using Landsat images. Over the 21-
year study period, mean mangrove extent was 1283 ha with an overall net increase of 60 ha
(0.23% year
-1
). The majority of extent changes were small (< 2 ha) and contiguous to the
existing mangrove extent. We then assessed the relation of mangrove cover change along the
lagoon coastline with wave power (rate of energy transfer by waves), using fetch measures
and local wind data. We found lower wave power values for stable mangrove areas than for
areas that had gained or lost mangroves from 1997 to 2018. We identified wave power
thresholds of 2.3 W m
-1
for stable mangrove and 7.1 W m
-1
for mangrove occurrence. These
thresholds might be valuable for assessing threats and sites with the greatest potential for
mangrove restoration across similar areas worldwide. Our results highlight the importance
of quantifying mangrove extent changes at a local scale to assist with planning for the
protection and restoration of this ecologically important habitat, given its vulnerability to the
pressures associated with climate change.
© 2021 The Author(s). Published by Elsevier B.V.
CC_BY_4.0
1. Introduction
Mangroves grow along tropical coastlines, between the land and sea. They form a vital habitat for many species, including 69
mangrove-endemic vertebrates, and have a critical role in regulating material and energy fluxes across coastal ecosystems
(Luther and Greenberg, 2009; Carr et al., 2017). Mangrove forests store more carbon than most other forest types, protect coasts
from erosion, storms, wave action and sea level rise, and provide refuge for diverse coral reef communities from climate change
https://doi.org/10.1016/j.gecco.2021.e01564
2351-9894/© 2021 The Author(s). Published by Elsevier B.V.
CC_BY_4.0
]]]]
]]]]]]
⁎
Corresponding author at: Department of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, 8057 Zurich,
Switzerland.
E-mail addresses: annabelle.constance@ieu.uzh.ch (A. Constance), paul.haverkamp@ieu.uzh.ch (P.J. Haverkamp), nancy@sif.sc (N. Bunbury),
gabriela.schaepman@ieu.uzh.ch (G. Schaepman-Strub).
Global Ecology and Conservation 27 (2021) e01564
impacts (Donato et al., 2011; Duarte et al., 2013; Yates et al., 2014; Woodroffe et al., 2016). Mangrove forests are substantial
contributors to ocean biodiversity (Zhao et al., 2020). Socioeconomically, mangrove forests provide fish, crustaceans and timber
to billions of people (Martínez et al., 2007).
Mangroves are globally threatened across their range, with a world areal loss of 0.13% year
-1
between 2000 and 2016 and a
total loss of 35% since the 1980s (Valiela et al., 2001; Goldberg et al., 2020). Agricultural conversion is the greatest driver of
mangrove loss, although climate change is expected to exacerbate declines through sea level rise (Woodroffe et al., 2016;
Goldberg et al., 2020). At particular risk are low-lying atolls – small islands formed from coral reefs – where sea level rise is
expected to physically destabilize coastlines, inundate and exceed the occurrence threshold of vegetation, and salinise fresh-
water sources (Kench et al., 2018; Storlazzi et al., 2018). Combined with an increase in sea level, waves will lead to a shift in the
zone of wave activity on the coast, likely to cause twice as much land inundation on these atoll islands as predicted by sea level
rise effects alone (Storlazzi et al., 2018).
Local studies commonly report a net loss in mangrove extent (Kirui et al., 2013; Mafi-Gholami et al., 2020), although gains
have been identified for small islands and areas with minimal direct human pressure (Hu et al., 2018). Mangrove loss has been
predicted to result in changes in biodiversity, increased vulnerability of coastlines to sea level rise and reduced carbon storage
capacity (Curnick et al., 2019; Macreadie et al., 2019; Zhao et al., 2020). In comparison, gains in mangrove extent are often
associated with ecological shifts at the expense of other vegetation communities (Kelleway et al., 2017). Hence, monitoring and
understanding the processes of changing mangrove extent are essential for measuring changes in biodiversity (Turak et al.,
2017), for effective biodiversity conservation, and to develop locally-focused climate change adaptation and mitigation stra-
tegies (Duarte et al., 2013).
Aldabra Atoll, part of the Seychelles archipelago, hosts the largest area of mangroves in the country and has been strictly
protected since 1976. Mangrove habitat covers more than 11% of Aldabra’s total land area – one of the highest proportions of
mangrove cover worldwide relative to island size (Giri et al., 2011; Walton et al., 2019). Aldabra’s mangroves support several
globally threatened and endemic populations of terrestrial and marine species, including the largest breeding population of
frigatebirds in the Indian Ocean (Šúr et al., 2013). Red-footed boobies and wading birds, such as egrets and herons, use the
mangroves as their main nesting sites, and the mangroves also provide essential feeding, breeding and nursery habitat for
threatened turtles, giant tortoises, sharks, and many other marine species (Macnae et al., 1971; Diamond, 1971; Taylor et al.,
1971; Šúr et al., 2013). In particular, the lagoon and surrounding mangroves are an important nursery for the second largest
population of endangered green turtles in the Western Indian Ocean (Mortimer, 2012).
Despite being strictly protected, Aldabra’s mangroves are vulnerable to climate change impacts. Changes in rainfall fre-
quency can influence mangrove extent by altering local conditions including salinity, sediment influx and even nutrient
availability (Asbridge et al., 2015). Rainfall patterns on Aldabra show an increase in drought frequency over the past two decades
(Haverkamp et al., 2017) and could already be influencing mangrove extent changes. Sea level rise impacts mangroves through
eroding coastlines, inundation stress, and increased salinity within the intertidal zone (Asbridge et al., 2015; Kench et al., 2018;
Storlazzi et al., 2018). Mean global sea level rise has increased from 2.4 ± 0.2 mm year
-1
in 1993–3.3 ± 0.3 mm year
-1
in 2014
(Chen et al., 2017). Estimates of sea level rise in the Indian Ocean are 5–10 mm year
-1
for the period 2000–2015 for the south-
west open ocean (Thompson et al., 2016; Jyoti et al., 2019), and 2–3 mm year
-1
for coastal areas within the Mozambique
Channel, including Aldabra (Testut et al., 2016). Increase in wave power can also be linked to climate change and represents a
threat to mangrove survival (Reguero et al., 2019, Cannon et al., 2020). Within the Indian Ocean, the mean wave power has
increased by 0.3% year
-1
since 1948 to 2.9 kW m
-1
in 2008 (Reguero et al., 2019). While increases in global wave power
correspond to the open oceans, further increases are likely to have important consequences for the coastal mangroves of
Aldabra, especially since mangrove persistence along sheltered coasts has been linked to wave height (wave power depends on
the square of wave height) thresholds of only 80 mm (Cannon et al., 2020). The interaction of a changing climate and physical
oceanic processes at the coastline, including wave action, are therefore expected to lead to changes in mangrove extent on the
atoll.
Here, we examine the spatio-temporal extent of mangroves on Aldabra and the relationship between wave power and
changes in mangrove extent by addressing three questions: (1) How has the mangrove extent changed on Aldabra over the past
two decades? (2) Where are extent changes occurring? And (3) Are observed mangrove extent changes occurring in areas with
the highest wave exposure? We anticipate that our results can be used to further our understanding of the vulnerability of this
crucial habitat and the associated fauna, to support conservation efforts on Aldabra, and to improve the detection of biodiversity
changes of mangrove habitats worldwide.
2. Methods
2.1. Research site
Aldabra Atoll (9°24'S, 46°20'E) is a UNESCO World Heritage Site in the Western Indian Ocean with a land area of 155 km
2
and
an average height above sea level of 8 m. Aldabra consists of four main islands – Picard, Polymnie, Malabar and Grande Terre –
with other smaller islands scattered within the lagoon (205 km
2
). Aldabra experiences a range of tides 2–3 m high with nearly
three-quarters of the lagoon draining at low tide through the channels separating the main islands (Hamylton et al., 2018).
Relative to sea level, lagoon depth ranges from 0.2 to 30 m. Most of the lagoon floor is flat and shallow (< 3 m) and is dominated
by sparse macroalgae and seagrass on sand. The lagoon coastline is rocky and exposed to wave activity. The wave action within
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
2
the enclosed lagoon is less intense than on the seaward shores of the atoll and is limited by the fetch (length of open water over
which the wind can blow).
Aldabra’s mangroves are restricted to the coastal areas inside the atoll’s lagoon, with an estimated total extent of 1720 ha
(Walton et al., 2019). Seven species of mangroves were recorded on Aldabra, with Avicennia marina, Bruguiera gymnorhiza,
Ceriops tagal and Rhizophora mucronata being most common (Macnae et al., 1971; Constance, 2016). Pemphis acidula, considered
a mangrove associate (not a true mangrove species) in this study, forms a homogeneous habitat on Aldabra and dominates the
landward side of mangroves, especially on Malabar. Otherwise, Aldabra’s landscape is largely covered with scrub vegetation of
varying height, either continuous or in a mosaic with open, rocky ground (Walton et al., 2019).
Aldabra’s climate shows pronounced seasonal variation. At least 75% of the yearly rainfall (mean 975 mm year
-1
) falls during
the northwest monsoon from November to April. May to October is the driest period, with predominantly southeast trade
winds. The seasonality in rainfall drives net primary productivity on the atoll for most vegetation except mangroves
(Haverkamp et al., 2017).
2.2. Mangrove extent over time
To quantify mangrove extent on Aldabra over the past two decades, we performed a supervised classification of Landsat
satellite imagery collected from 1997 to 2018. We followed a four-step methodology: scene selection, preprocessing, analysis,
and validation (Kennedy et al., 2009), using ENVI software (version 5.5, Exelis Visual Information Solutions, Boulder, Colorado)
for geoprocessing.
2.2.1. Landsat data selection
Landsat surface reflectance products (30 m spatial resolution) were downloaded from the USGS EROS Science Processing
Architecture website (http://espa.cr.usgs.gov). We filtered data over Aldabra from January 1980 to March 2020, with less than
10% land cloud cover, resulting in 50 images from Landsat 5 (L5), Landsat 7 (L7) and Landsat 8 (L8). Images were manually
assessed to determine remaining cloud cover over mangroves and the quality of atmospheric corrections (highest quality for L7
and L8, but no quality selection criteria used for L5 because scenes of high quality were unavailable). We selected final images
for the analyzes (see Table A1) between four and seven years apart, allowing us to detect changes in mangrove extent, whereas
shorter intervals are known not to identify changes (Coppin et al., 2010; Kirui et al., 2013). To address the lack of cloud-free data
from 2000 to 2009, we separately preprocessed two partly cloud-covered images (February 2003 and August 2004) and mo-
saicked the cloud-free parts of the images. The mosaicked image is referred to as scene 2004, given the 2004 image makes up
75% of the mosaic.
2.2.2. Preprocessing
We compared the geographic registration of all images to ensure that each pixel refers to the same geographic location over
time. The L5 images showed georegistration differences (> 5 pixels) to the rest of the images and were thus co-registered to the
2018 L8 image. We chose at least 10 clearly discernible Ground Control Points (GCPs) from the base L8 image and corresponding
location points in the L5 images. A polynomial transformation with nearest neighbour spectral matching was used with RMSE
below 0.5 pixel (15 m) (Jensen et al., 2009). We used a standardised projected coordinate system (WGS 1984 UTM Zone 38 S)
and cropped all scenes to only include the area immediately around Aldabra.
2.2.3. Analysis
We used a Maximum Likelihood Classification (MLC) to assess the mangrove extent on Aldabra Atoll across years. We
separated six land cover classes: mangrove; scrub; pemphis; surface; sand and; water. A description of the classes is given in
Walton et al. (2019). We created a training point dataset for each land cover class, based on image-derived GCPs (Walton et al.,
2019) and 580 GPS points collected in 2009 during a habitat survey (Seychelles Islands Foundation, unpublished data). From
these points, we selected a subset, based on an even spatial distribution across the atoll and on homogeneity for the given land
cover type (Jensen et al., 2009). Because of differences in tide conditions across images, the training points had to be updated for
each image. Mostly, sand and water training points varied between images. We extracted the values of six Landsat surface
reflectance bands (covering similar wavelength ranges across different Landsat sensors, see Table A1) at each training point
location and tested for class spectral separability. Mangrove training points had good separability against other land cover
classes (Transformed Divergence values ≥ 1.9). We developed a cloud mask based on all cloud (and shadow) pixels throughout
all years and images by manual digitization. We applied this cloud mask (5.6% of the total land area) before the classification to
keep the analyzed area the same across all images and dates. Based on the class statistics from training points and the cloud-
masked images, we ran a MLC in ENVI. We summed the resulting areal extent of pixels classified as mangrove in hectares and
reported this as the mangrove extent for the respective year.
2.2.4. Validation of classification
We used the ‘confusion matrix’ analysis in ENVI to validate the classification results. The method compares the land cover
class of each ground truth point with the class assigned by the MLC to the same point. We applied a subset of the initial ground
truth points not used for training the classifier to validate our results. We reported the accuracy of the classification based on
the overall accuracy matrix (number of correctly classified points/total number of ground truth points).
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
3
2.3. Mangrove extent change over time
We compared classified images across years to assess if changes in mangrove extent occurred. We used the ‘post-classifi-
cation change detection’ algorithm in ENVI to identify where land cover classes changed between dates, or stayed the same.
2.3.1. Analysis
We performed post-classification change detection for five different time intervals (1997–2004, 2004–2009, 2009–2014,
2014–2018, and 1997–2018) to assess longer-term spatial dynamics in mangrove forests on Aldabra. The output was five change
maps, one for each time interval showing changes in land cover (e.g., mangrove to pemphis, pemphis to scrub) and their areal
extent. To identify how the change affected mangroves, we grouped the classification from each change map into five land cover
change classes: (i) stable mangrove (mangrove to mangrove); (ii) mangrove loss (mangrove that changed to any other class);
(iii) mangrove gain (any class that changed to mangrove); (iv) other land covers (other stable or change classes not including
mangroves); and (v) water (stable water class). To account for the relative contribution of patch size to total changes in extent,
contiguous land cover change class pixels were grouped to patches applying the eight-neighbour rule, and the area of the patch
was calculated.
2.3.2. Evaluation of mangrove change classes
The evaluation of mangrove change classes was carried out by comparing the Landsat-based change map for the 2009–2014
period with high resolution drone imagery. We used a single time point (2015 drone imagery) so that our measure of accuracy is
relevant to the land cover class identified at the end of the change detection. In December 2015, RGB and NIR images were
captured by a senseFly eBee classic drone over ca. 10% of the land area of the atoll. The images were orthomosaicked using the
Photoscan version 1.2.6 software (AgiSoft LLC, St. Petersburg, Russia) resulting in a final mosaic with a spatial resolution of
1.5 m. As the drone imagery was acquired with a non-calibrated camera, we calculated the Normalized Difference Vegetation
Index (NDVI) to standardise the data across flight lines. Also, previous research showed that NDVI is a useful measure to
separate mangroves from other land covers on Aldabra (Haverkamp et al., 2017). We selected pixels at locations identified in the
field as mangroves with ± 1 m accuracy (Constance, 2016) to determine an NDVI reference range for mangrove. We tested the
difference between the mean NDVI for mangrove habitat (of different tree coverage) and other land covers (including surface,
sand, and non-mangrove vegetation) with a t-test. The mean NDVI for mangroves was higher (p < 0.001) at the 0.05 confidence
level (n = 362, mean = 0.4, s = 0.16) than for other land covers (n = 1228, mean = 0.2, s = 0.10). The 95% confidence interval of the
mean NDVI for mangroves (0.22–0.58) was used as a threshold range for separating mangrove presence and absence in the
drone imagery.
We overlaid a random selection of pixels from the Landsat-based change map on the drone imagery to test if the final land
cover class was assigned correctly. The number of Landsat pixels selected from each change class (mangrove loss, n = 13; stable
mangrove, n = 17; other land covers, n = 11) was proportional to the total number of pixels from each land cover change class
that overlapped with the drone imagery. Mangrove loss had a slightly higher proportion of its total to increase representation in
the rarer classes (Olofsson et al., 2014). The evaluation of pixels from the mangrove gain class could not be performed because
there was no spatial overlap with available drone imagery. Mangrove presence was assigned if the area of a Landsat pixel
contained ≥ 45% of drone pixels within the mangrove NDVI range (each Landsat pixel covered 20 ×20 drone pixels). Finally, we
calculated an overall accuracy measure for the Landsat change detection by dividing the number of correctly classified man-
grove absence or presence pixels by the total number of validated pixels. As an example, for Landsat land cover pixels of class
mangrove loss and other land covers, a drone evaluation of mangrove absence counted as a correct classification (Fig. 1). Further,
visual inspection of the mangrove loss change class pixels was used to qualitatively verify the land cover class at the beginning
of the change detection.
2.4. Relation of wave power with mangrove extent changes
We calculated wave power along the lagoon coastline to determine whether the presence of stable mangrove and mangrove
change could be related to exposure to wave power (Ekebom et al., 2003).
2.4.1. Wave power calculation
To define the interaction boundary of waves with the coast, we created a vector of the Aldabra lagoon coastline from the
Landsat 2014 classification layer of water (Section 2.2.3). From the coastline vector, we used ArcGis 10.2.2 software (Environ-
mental Systems Research Institute, Redlands, California, USA) to return 1069 points spaced at 240 m intervals along the lagoon
coastline for islands with an area greater than 3500 m
2
. The interval between points along the coastline was chosen based on
the coastal profile of the lagoon and for comparison with the Landsat-scale change detection. Each point was used as the origin
of 48 evenly distributed radiating lines (7.5 degrees apart). These radiating lines (fetch lines) were extended to the point where
they intersected with the nearest coastline in the direction of the fetch. The length of this line was the fetch length, and all
points had a maximum of 48 fetch lines in 48 directions (excluding fetch lines of zero length).
For each fetch line, we used the fetch length and mean wind speed as input for the calculations of the wave height, peak
spectral period, wave energy, and finally wave power (P
i
) following Ekebom et al. (2003). We used wind data for 2018 from an
automatic weather station on Picard Island (SIF, unpublished data) to calculate mean wind speed and direction every 10 min for
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
4
wind sectors of 45 degrees. Finally, each wave power value was multiplied by 1/6 the wind occurrence from the direction of the
line in question (six lines for each wind sector). We calculated wave power for a given point by summing up the 48 wave
exposure values of each fetch from that point (Eq. 1).
=
=
P t P( )
pii i
1
48 1
645
(1)
Where P
p
is the wave power (W m
-1
of wave front length) for a point p; t
45i
is the probability that the wind blows from a 45°
sector including line i; and P
i
is the wave power in the direction of the fetch line i.
2.4.2. Wave power and mangrove extent change
We hypothesised that mangrove extent changes (gain or loss) occurred in areas with higher wave power than stable
mangrove cover because of shifting sediment dynamics in areas of higher wave power. To test this, we first extracted the land
cover from the change detection 1997–2018 for every point along the coastline. We used a t-test to calculate the difference
between the log-transformed wave power for: (1) stable mangrove and changed mangrove; and (2) changed mangrove and
other land covers. We did a power analysis for both t-tests on the basis of the mean, between-group comparison effect size
observed and alpha at 0.05, to assess whether our sample sizes (stable mangrove = 82, mangrove change = 74, other land covers
= 703) were large enough to obtain the statistical power at the recommended 0.8 level.
3. Results
3.1. Mangrove extent over time
Mangroves were distributed along the lagoon coastline on all four major islands of Aldabra. From 1997 to 2018, the overall
extent of mangroves increased by 60 ha on Aldabra, representing a 0.23% increase in mangrove area per year, and a mean
mangrove extent of 1283 ha over the 21-year period (Fig. 2). From 1161 ha in 1997, the mangrove area decreased slightly (40 ha)
to its lowest extent of 1121 ha in 2004. Over the next five years, a 38% increase (431 ha) in mangrove area occurred, with the
Fig. 1. Examples of evaluation of Landsat land cover change classes 2009–2014 against drone imagery captured in 2015. In each panel the border of one Landsat
pixel and the assigned mangrove presence or absence from drone imagery (dashed border = mangrove absence, solid border = mangrove presence) is shown. A.
Correctly classified as ‘other land covers’. B. Correctly classified as ‘mangrove loss’. C. Correctly classified as ‘stable mangrove’. D. Classified as ‘other land covers’
in Landsat but drone imagery indicates mangrove presence.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
5
largest mangrove area of the study, 1552 ha, in 2009. From 2009 to 2018 there was a total decrease in mangrove area of 331 ha
(21% decline). Overall mangrove extent classification accuracy ranged from 97% (1997 and 2004) to 98% (2009, 2014, and 2018).
3.2. Mangrove extent change over time
Larger contiguous extent change was unequally distributed around Aldabra’s coastline, with most changes occurring on the
landward margins of mangrove extent.
On Picard there was an increase of mangroves along the shorelines of the West Channels (Fig. 3). A marked increase also
occurred along Grand Poche from 1997 to 2018. Otherwise, the changes on Picard were highly variable over different time
periods, occurring on the seaward and landward margins.
On Malabar, the overall change trend was a notable increase along the landward mangrove extent (Fig. 3). The majority of
the landward gains occurred from 2004 to 2009, which was consistent with the large increase observed across the atoll for that
time period (Fig. 4). Over the next five years, while the landward extent continued to increase on Malabar, mangroves exposed
to the sea were lost. The loss in mangroves was almost parallel to the coastline while the landward gains occurred primarily to
the west or east of the stable mangrove areas at the mangrove-pemphis boundary.
On East Grande Terre (Fig. 4) changes were more varied, with some regions having persistent mangrove areas, while others
showed intermittent mangrove expansion and retraction. During 2009–2014, we identified the largest proportion (30%) of
mangrove loss to the scrub and sand land covers. The landward gains and losses around East Grande Terre were directional
along the Bras Cinq Cases waterway (Fig. 4). On the more exposed boundary between the lagoon and the mangroves, to the west
of the Bras Cinq Cases area, substantial loss of mangroves occurred between 2009 and 2014, which had not recovered by 2018
(Fig. 4). Our assessment of the change in the size of mangrove extent showed that 91% of the total gains and losses identified
between 1997 and 2018 occurred within areas of < 2 ha.
We found a high agreement between the 2009–2014 Landsat analysis and the drone imagery (overall accuracy of 87.8%)
when evaluating the mangrove change classes. The lowest accuracy was seen for pixels identified as ‘other land covers’ in
Landsat, which were determined to be mangroves in the drone imagery (three of 11 sample points incorrectly classified). We
visually identified dead or unhealthy mangrove trees on the drone imagery for all Landsat ‘mangrove loss’ pixels.
3.3. Relation of wave power with mangrove extent changes
The highest calculated wave power was 134 W m
-1
on the western coast of the lagoon. The winds prevailed from the E, SE
and NW sectors (see Fig. A2). The northern shore of the lagoon is exposed to waves from the dominant easterly and south-
easterly trade winds and had more points with higher wave power than the southern shore, which lies to the leeward side of
Grande Terre. The exception was for coastline points close to the West Channels on Grande Terre. Wave power intervals (at 95%
confidence level) were 0.8–2.3 W m
-1
for stable mangrove, 1.8–7.1 W m
-1
for changed mangrove, and 5.8–8.7 W m
-1
for other
land covers. Mean wave power was three times higher in areas with changed mangrove cover between 1997 and 2018 compared
to areas with stable mangrove cover (p < 0.001, power of t-test = 0.8; Fig. 5). Further, we found that areas along the coastline
that were not vegetated by mangroves during the study period had the highest (p < 0.05) mean wave power compared to areas
that gained or lost mangrove over that same period (power of t-test = 0.4).
Fig. 2. Mangrove extent on Aldabra Atoll from Landsat images over a 21-year period.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
6
Fig. 3. Spatial extent of stable mangrove habitat, mangrove gain and loss areas on Aldabra from 1997 to 2018. Numbered boxes indicate regions with substantial
expansion and retraction over the study period: Box 1. Subset of Malabar, Box 2. Subset of East Grande Terre shown in greater detail in Fig. 4.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
7
4. Discussion
Our study investigates extent of mangroves over the past 21 years on Aldabra Atoll, a critical habitat for the protection of
global biodiversity and for climate change mitigation and adaptation efforts (Luther and Greenberg, 2009; Duarte et al., 2013;
Zhao et al., 2020). Sea level rise, and increasing storm activity, wave exposure and flooding can lead to mangrove loss at seaward
margins, expansion into landward areas, and shifts in species composition (Asbridge et al., 2015). This means that mangrove
extent is spatially variable at local and regional scales, with both positive and negative consequences. On Aldabra, we found an
overall net gain in mangrove area of 60 ha from 1997 to 2018. The change trends were unequally distributed around the
coastline and the majority of change occurred at numerous different but restricted locations.
Fig. 4. Stable mangrove, mangrove gain and loss areas for two regions on Aldabra Atoll; Malabar (left column) and East Grande Terre (right column).
Fig. 5. Wave power range (W m
-1
) of areas of stable mangrove, changing mangrove and other land cover on Aldabra Atoll.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
8
The range of mangrove forest extent determined in our study was 1121–1552 ha, higher than a global forest cover map
estimate of 1087 ha for the entire Seychelles (Giri et al., 2011), but lower than a recent estimate of 1720 ha using high spatial
resolution satellite imagery (Walton et al., 2019). The accuracy of the global map (Giri et al., 2011) using remote sensing data
was not provided with the final product, and the inconsistency of our data with Walton et al. (2019) was likely due to their
classification of mangrove and pools together, while our study classified water bodies (including pools) separately. It is clear
from all studies that Aldabra contains more mangrove habitat than the remaining Seychelles islands combined, underscoring
the ecological importance of the atoll on a national and regional level.
The evaluation of Landsat-derived mangrove extent changes using drone imagery suggests that the reported mangrove
extent from our study underestimated mangrove cover for non-continuous stretches of mangroves (Fig. 1D). Another possible
source of uncertainty was in evaluating mangrove loss pixels from 2009 to 2014 using data from a single time point. Despite not
knowing what land cover mangroves changed from, we saw dead or unhealthy mangrove trees on the drone imagery for all
mangrove loss pixels, confirming that losses were correctly identified at the Landsat scale (Fig. 1B). Our results show that data
from the Landsat surface reflectance archive combined with GCPs can be used to map mangrove extent on islands and small
areas, which might not be accurately represented in global extent studies (Giri et al., 2011). Furthermore, drone imagery is a
useful source for the evaluation of mangrove extent changes.
Substantial seaward mangrove loss on East Grande Terre was offset and surpassed by landward expansion on Malabar and
Picard. While these gains and losses are spatially dynamic and balance each other to a large extent, the changes are concerning.
The losses along the seaward margins affect primarily established mangrove forests, dominated by tall, mature trees (Macnae
et al., 1971), which are the preferred habitat of breeding seabirds on Aldabra (Diamond, 1971; Šúr et al., 2013). In contrast, the
gains in mangrove extent, which are mostly at the landward margins, are primarily successional stage vegetation on the
boundary. Here, young mangrove trees expand into other land covers, for example grassland, which is a globally-unique habitat
to Aldabra, essential for the threatened Aldabra giant tortoise (Walton et al., 2019). Thus, while the total extent of mangroves
increased slightly across two decades, the spatial dynamics of mangrove change may have key impacts on Aldabra’s biodi-
versity.
In general, droughts impact mangrove extent, however, the effect varies by location (Rogers et al., 2006; Asbridge et al.,
2015). Along the Oman Sea, mangroves contracted severely during droughts (Mafi-Gholami et al., 2020). Conversely, on Aldabra,
the greatest extent of mangroves in our study was in 2009, which was among the driest drought periods on the atoll since 2000
(Haverkamp et al., 2017). Our study showed that mangrove extent increased inland into the scrub and grassland on East Grande
Terre (Fig. 4) during the longer drought. The expansion could have been due to the dependence of the surrounding vegetation
on rainfall (Haverkamp et al., 2017). Inherently, mangrove trees are also remarkably adept at maintaining water uptake in saline
conditions (Reef and Lovelock, 2015), including through condensation and foliar water uptake (Hayes et al., 2020). During a
drought, established mangrove stands promote the growth of new trees by conserving soil moisture and preventing desiccation
of mangrove propagules (Asbridge et al., 2015). This could explain why mangrove gains along East Grande Terre occurred
primarily alongside existing mangrove areas, emphasizing the importance of the continued protection of mangroves. Such
expansion highlights community-specific vulnerabilities of the surrounding grassland vegetation that is likely to worsen as
drought frequency increases (Haverkamp et al., 2017).
Mangrove retraction along the coastline occurred primarily on Picard and East Grande Terre. In contrast to the landward
expansion, the seaward changes are more likely to result from coastal processes including interactions between sea level,
waves, storms and sediment dynamics. Over the past two decades, sea level in the Western Indian Ocean region has increased
by 2–3 mm year
-1
(Testut et al., 2016) and further increases will likely exacerbate the impact of waves and storms on atolls by
allowing larger, more powerful waves to reach the coastline (Storlazzi et al., 2018). These coastal processes determine flooding
and erosion, affecting the distribution of sediments that are essential for mangrove persistence (Reguero et al. 2019; Shope and
Storlazzi, 2019). Thus, these change drivers can ultimately affect mangrove extent and the associated protection of the coastline,
biodiversity, and carbon capture capabilities. Increased efforts should be placed on monitoring mangrove extent changes for
island coastlines that are particularly vulnerable to the effects of climate change. Aldabra’s designation as a UNESCO World
Heritage Site, a Ramsar Wetland Site of International Importance and a Blue Park are, in part, due to its healthy and extensive
mangrove forest. Therefore, monitoring, maintaining and understanding this habitat will be crucial for future management of
the atoll.
On Picard, mangrove extent increased along Grand Poche inlet over the whole study period. Grand Poche hosts one of four
frigatebird colonies on Aldabra, the largest breeding population of frigatebirds in the Indian Ocean. Aldabra’s other frigatebird
colonies are distributed mostly along the lagoon margins of Malabar (Šúr et al., 2013). Although the mangrove trees at the edge
of these colonies are often leafless and stag-headed (Macnae et al., 1971; Constance, 2016) due to the seabirds, the mangrove
forest gain along Grand Poche indicates that the nesting frigatebirds do not damage their habitat, and they may promote it in
the form of nutrient-rich guano. Future research is recommended to assess the role of seabird colonies on the nutrient cycle of
Aldabra’s mangroves. Finally, given the recent and major 9-year decline in mangrove extent on Aldabra, and the loss of mature
seaward mangrove habitat, we recommend regular assessments of mangrove extent changes on the atoll. Aldabra’s seabird
colonies are likely to be affected by a long-term decline in mangrove extent, so continued and expanded monitoring of seabird
distribution and population sizes, and other associated fauna, in this habitat is advisable.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
9
We investigated the relation of wave power with areas of stable mangroves, mangrove change, and mangrove absence. Up to
2.3 W m
-1
, mangroves persisted over time. Our results support recent research (Cannon et al., 2020) showing that wave ex-
posure plays a major role in mangrove extent stability along sheltered coastlines around the world. A wave height threshold of
8.0 cm has been identified for the persistence of mangroves along the estuarine coast of Florida (Cannon et al., 2020), a value
which is comparable to the mean wave height used to calculate wave power in our study (6.2 ± 4.1 cm). Our findings identify a
wave power threshold for mangrove presence of 7.1 W m
-1
of the wave front length. Above this value, it is unlikely that Aldabra’s
coastline will sustain a mangrove population. The wave power thresholds we identified have wide-ranging practical implica-
tions for planning and threat assessment, for example, for determining suitable areas for mangrove reintroductions or estab-
lishment, or identifying mangrove areas at greatest risk. Future research should investigate the impact of wave power at the
scale of single mangrove species, which would provide more opportunities to address mangrove extent protection, resilience,
and restoration at broad management scales.
Within the Indian Ocean, the mean wave power of 2.9 kW m
-1
(Reguero et al., 2019) is 22 times higher than the maximum
wave power calculated on Aldabra’s coastline in our study. Although projections of wave changes for coastal areas have sub-
stantial uncertainty (Young et al., 2012; Seneviratne et al., 2012), the direct connection between wind and local wave power
suggests that predicted increases in extreme maximum wind speed of storms will have a major impact on maximum wave
power experienced at a local scale (Seneviratne et al., 2012). The coastline protection provided by mangroves is likely to be
weakened by extreme seasonal events where initial erosion patterns following disturbance will shift the sediment dynamics
necessary for mangrove development and hinder habitat recovery in affected areas (Balke et al., 2015; Sánchez-Núñez et al.,
2019). Further research should therefore consider the seasonal and predicted extreme wave power gradients to identify vul-
nerable mangrove areas at local scale.
In our study, we measured and tracked mangrove extent, a valuable and high priority measure for biodiversity change (Turak
et al., 2017). While mangrove habitats across the world are in decline (Valiela et al., 2001; Goldberg et al., 2020), we found that
the protected mangrove forest on Aldabra has expanded over the past two decades, whereby mangroves have persisted in
certain areas and rapidly colonized newly suitable areas over time. Our results highlight the importance of quantifying changes
in mangrove extent at local scale to assist with planning for the protection and restoration of this valued habitat, and to improve
detection of changes occurring in global biodiversity.
CRediT authorship contribution statement
Annabelle Constance: Conceptualization, Methodology, Formal analysis, Writing - original draft, Writing - review & editing.
Paul J. Haverkamp: Methodology, Writing - review & editing. Nancy Bunbury: Writing - review & editing. Gabriela
Schaepman-Strub: Conceptualization, Methodology, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have ap-
peared to influence the work reported in this paper.
Acknowledgement
The authors would like to express their gratitude to the Seychelles Islands Foundation (responsible for the management of
Aldabra) and Chief Executive Officer Frauke Fleischer-Dogley for field data and support of this manuscript. Specific thanks to
Rowana Walton for cleaning the habitat reference points from 2009; Justin Prosper, Lindsay Chong-Seng, Christina Quanz and
Christian Fleischer for the collection of the habitat reference points in 2009 and Environment Trust Fund for funding the flights
to Aldabra; April Burt, Jakawan Hoareau, Jude Brice, Marcus Dubel, Marvin Roseline, Shiira Padayachy, Ervin Estico for the
collection of the habitat reference points in 2016. We appreciate the USGS Earth Resources Observation and Science (EROS)
Centre for freely available access to Landsat data; Philip Jörg from the Department of Geography at the University of Zurich for
mosaicking the drone data; Michael Schaepman and Mathias Kneubühler for flying the drone; Dennis Hansen for widespread
support and guidance; Cornelia Krug for helping with the revisions; and the anonymous reviewers for their useful comments on
the manuscript. The contribution by Annabelle Constance and Gabriela Schaepman-Strub was supported by the University of
Zurich Research Priority Program on Global Change and Biodiversity.
Appendix A
Landsat bands used for spatial analysis (Table A1); spatial representation of wave power values on land cover change classes
(Fig. A1).
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
10
Table A1
Landsat surface reflectance scenes and bands used in the classification of mangrove extent on Aldabra Atoll from 1997 to 2018. Only bands listed here were used
for the classification.
Acquisition date Sensor Band name Wavelength (µm)
03 Feb 1997 Landsat 5 Thematic Mapper Band 1 0.45–0.52
01 Aug 2004 Band 2 0.52–0.60
05 Dec 2009 Band 3 0.63–0.69
Band 4 0.76–0.90
Band 5 1.55–1.75
Band 7 2.08–2.35
28 Feb 2003 Landsat 7 Enhanced Thematic Mapper Plus Band 1 0.45–0.52
Band 2 0.52–0.60
Band 3 0.63–0.69
Band 4 0.77–0.90
Band 5 1.55–1.75
Band 7 2.09–2.35
30 Sep 2014 Landsat 8 Operational Land Imager Band 2 0.45–0.51
Band 3 0.53–0.59
09 Sep 2018 Band 4 0.64–0.67
Band 5 0.85–0.88
Band 6 1.57–1.65
Band 7 2.11–2.29
Fig. A1. Wave power exposure values calculated for the coastline study points with 240-m intervals illustrated for north-west Aldabra. The wind rose chart for
Aldabra shows the frequency of the mean wind speed every 10 min in 2018 at the Aldabra research station.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
11
References
Asbridge, E., Lucas, R., Accad, A., Dowling, R., 2015. Mangrove response to environmental changes predicted under varying climates: case studies from Australia.
Curr. For. Rep. 1, 178–194.
Balke, T., Swales, A., Lovelock, C.E., Herman, P.M.J., Bouma, T.J., 2015. Limits to seaward expansion of mangroves: translating physical disturbance mechanisms
into seedling survival gradients. J. Exp. Mar. Biol. Ecol. 467, 16–25.
Cannon, D., Kibler, K., Donnelly, M., Mcclenachan, G., Walters, L., Roddenberry, A., Phagan, J., 2020. Hydrodynamic habitat thresholds for mangrove vegetation on
the shorelines of a microtidal estuarine lagoon. Ecol. Eng. 158.
Carr, M.H., Robinson, S.P., Wahle, C., Davis, G., Kroll, S., Murray, S., Schumacker, E.J., Williams, M., 2017. The central importance of ecological spatial connectivity
to effective coastal marine protected areas and to meeting the challenges of climate change in the marine environment. Aquat. Conserv.: Mar. Freshw.
Ecosyst. 27, 6–29.
Chen, X., Zhang, X., Church, J.A., Watson, C.S., King, M.A., Monselesan, D., Legresy, B., Harig, C., 2017. The increasing rate of global mean sea-level rise during
1993–2014. Nat. Clim. Change 7, 492–495.
Constance, A., 2016. Mangroves on Aldabra – Habitat Change Trends, Stand Structure and Species Composition. M.Sc. Dissertation, University of Zurich.
Coppin, P., Jonckheere, I., Nackaerts, K., Muys, B., Lambin, E., 2010. Review ArticleDigital change detection methods in ecosystem monitoring: a review. Int. J.
Remote Sens. 25, 1565–1596.
Curnick, D.J., Pettorelli, N., Amir, A.A ., Balke, T., Barbier, E.B., Crooks, S., Dahdouh-Guebas, F., Duncan, C., Endsor, C., Friess, D.A., Quarto, A., Zimmer, M., Lee, S.Y.,
2019. The value of small mangrove patches. Science 363 239-239.
Diamond, A.W., 1971. The ecology of the sea birds of Aldabra. Philos. Trans. R. Soc. Lond. Ser. B: Biol. Sci. 260, 561–571.
Donato, D.C., Kauffman, J.B., Murdiyarso, D., Kurnianto, S., Stidham, M., Kanninen, M., 2011. Mangroves among the most carbon-rich forests in the tropics. Nat.
Geosci. 4, 293–297.
Duarte, C.M., Losada, I.J., Hendriks, I.E., Mazarrasa, I., Marbà, N., 2013. The role of coastal plant communities for climate change mitigation and adaptation. Nat.
Clim. Change 3, 961–968.
Ekebom, J., Laihonen, P., Suominen, T., 2003. A GIS-based step-wise procedure for assessing physical exposure in fragmented archipelagos. Estuar. Coast. Shelf
Sci. 57, 887–898.
Giri, C., Ochieng, E., Tieszen, L.L., Zhu, Z., Singh, A., Loveland, T., Masek, J., Duke, N., 2011. Status and distribution of mangrove forests of the world using earth
observation satellite data. Glob. Ecol. Biogeogr. 20, 154–159.
Goldberg, L., Lagomasino, D., Thomas, N., Fatoyinbo, T., 2020. Global declines in human-driven mangrove loss. Glob. Chang Biol. 26, 5844–5855.
Hamylton, S., Hagan, A., Bunbury, N., Fleischer-Dogley, F., Spencer, T., 2018. Mapping the lagoon at Aldabra Atoll, Western Indian Ocean. Atoll Res. Bull. 45–59.
Haverkamp, P.J., Shekeine, J., De Jong, R., Schaepman, M., Turnbull, L.A., Baxter, R., Hansen, D., Bunbury, N., Fleischer-Dogley, F., Schaepman-Strub, G., 2017. Giant
tortoise habitats under increasing drought conditions on Aldabra Atoll—Ecological indicators to monitor rainfall anomalies and related vegetation activity.
Ecol. Indic. 80, 354–362.
Hayes, M.A., Chapman, S., Jesse, A., O’brien, E., Langley, J.A., Bardou, R., Devaney, J., Parker, J.D., Cavanaugh, K.C., 2020. Foliar water uptake by coastal wetland
plants: a novel water acquisition mechanism in arid and humid subtropical mangroves. J. Ecol. 1–13.
Hu, L., Li, W., Xu, B., 2018. The role of remote sensing on studying mangrove forest extent change. Int. J. Remote Sens. 39, 6440–6462.
Jensen, J., Im, J., Hardin, P., Jensen, R., 2009. Image classification. In: Warner, T., Nellis, D., Foody, G. (Eds.), The SAGE Handbook of Remote Sensing, first ed. SAGE
Publications Ltd, London.
Jyoti, J., Swapna, P., Krishnan, R., Naidu, C.V., 2019. Pacific modulation of accelerated South Indian Ocean sea level rise during the early 21st century. Clim. Dyn.
53, 4413–4432.
Kelleway, J.J., Cavanaugh, K., Rogers, K., Feller, I.C., Ens, E., Doughty, C., Saintilan, N., 2017. Review of the ecosystem service implications of mangrove en-
croachment into salt marshes. Glob. Change Biol. 23, 3967–3983.
Kench, P.S., Ford, M.R., Owen, S.D., 2018. Patterns of island change and persistence offer alternate adaptation pathways for atoll nations. Nat. Commun. 9, 605.
Kennedy, R.E., Townsend, P.A., Gross, J.E., Cohen, W.B., Bolstad, P., Wang, Y.Q., Adams, P., 2009. Remote sensing change detection tools for natural resource
managers: understanding concepts and tradeoffs in the design of landscape monitoring projects. Remote Sens. Environ. 113, 1382–1396.
Kirui, K.B., Kairo, J.G., Bosire, J., Viergever, K.M., Rudra, S., Huxham, M., Briers, R.A., 2013. Mapping of mangrove forest land cover change along the Kenya
coastline using landsat imagery. Ocean Coast. Manag. 83, 19–24.
Luther, D.A., Greenberg, R., 2009. Mangroves: a global perspective on the evolution and conservation of their terrestrial vertebrates. BioScience 59, 602–612.
Macnae, W., Westoll, T.S., Stoddart, D.R., 1971. Mangroves on Aldabra. Philos. Trans. R. Soc. Lond. B: Biol. Sci. 260, 237–247.
Macreadie, P.I., Anton, A., Raven, J.A., Beaumont, N., Connolly, R.M., Friess, D.A., Kelleway, J.J., Kennedy, H., Kuwae, T., Lavery, P.S., Lovelock, C.E., Smale, D.A.,
Apostolaki, E.T., Atwood, T.B., Baldock, J., Bianchi, T.S., Chmura, G.L., Eyre, B.D., Fourqurean, J.W., Hall-Spencer, J.M., Huxham, M., Hendriks, I.E., Krause-
Jensen, D., Laffoley, D., Luisetti, T., Marba, N., Masque, P., Mcglathery, K.J., Megonigal, J.P., Murdiyarso, D., Russell, B.D., Santos, R., Serrano, O., Silliman, B.R.,
Watanabe, K., Duarte, C.M., 2019. The future of Blue Carbon science. Nat. Commun. 10, 3998.
Mafi-Gholami, D., Zenner, E.K., Jaafari, A., 2020. Mangrove regional feedback to sea level rise and drought intensity at the end of the 21st century. Ecol. Indic. 110,
105972.
Martínez, M.L., Intralawan, A., Vázquez, G., Pérez-Maqueo, O., Sutton, P., Landgrave, R., 2007. The coasts of our world: ecological, economic and social im-
portance. Ecol. Econ. 63, 254–272.
Mortimer, J.A., 2012. Seasonality of Green Turtle (Chelonia mydas) reproduction at Aldabra Atoll, Seychelles (1980-2011) in the regional context of the Western
Indian Ocean. Chelonian Conserv. Biol. 11, 170–181.
Olofsson, P., Foody, G.M., Herold, M., Stehman, S.V., Woodcock, C.E., Wulder, M.A., 2014. Good practices for estimating area and assessing accuracy of land
change. Remote Sens. Environ. 148, 42–57.
Reef, R., Lovelock, C.E., 2015. Regulation of water balance in mangroves. Ann. Bot. 115, 385–395.
Reguero, B.G., Losada, I.J., Mendez, F.J., 2019. A recent increase in global wave power as a consequence of oceanic warming. Nat. Commun. 10, 205.
Rogers, K., Wilton, K.M., Saintilan, N., 20 06. Vegetation change and surface elevation dynamics in estuarine wetlands of southeast Australia. Estuar. Coast. Shelf
Sci. 66, 559–569.
Sánchez-Núñez, D.A., Bernal, G., Mancera Pineda, J.E., 2019. The relative role of mangroves on wave erosion mitigation and sediment properties. Estuaries Coasts
42, 2124–2138.
Seneviratne, S., Nicholls, N., Easterling, D., Goodess, C., Kanae, S., Kossin, J., Luo, Y., Marengo, J., Mcinnes, K., Rahimi, M., Reichstein, M., Sorteberg, A., Vera, C. &
Zhang, X., 2012. Changes in Climate Extremes and their Impacts on the Natural Physical Environment: An Overview of the IPCC SREX Report.
Shope, J.B., Storlazzi, C.D., 2019. Assessing morphologic controls on Atoll island alongshore sediment transport gradients due to future sea-level rise. Front. Mar.
Sci. 6, 245.
Storlazzi, C.D., Gingerich, S.B., Van Dongeren, A., Cheriton, O.M., Swarzenski, P.W., Quataert, E., Voss, C.I., Field, D.W., Annamalai, H., Piniak, G.A., Mccall, R., 2018.
Most atolls will be uninhabitable by the mid-21st century because of sea-level rise exacerbating wave-driven flooding. Sci. Adv. 4, eaap9741.
Šúr, M., Bunbury, N., Van De Crommenacker, J., 2013. Frigatebirds on Aldabra Atoll: population census, recommended monitoring protocol and sustainable
tourism guidelines. Bird Conserv. Int. 23, 214–220.
Taylor, J.D., Westoll, T.S., Stoddart, D.R., 1971. Intertidal zonation at Aldabra Atoll. Phiolos. Trans. R. Soc. Lond. B: Biol. Sci. 260, 173–213.
Testut, L., Duvat, V., Ballu, V., Fernandes, R.M.S., Pouget, F., Salmon, C., Dyment, J., 2016. Shoreline changes in a rising sea level context: the example of Grande
Glorieuse, Scattered islands, Western Indian Ocean. Acta Oecol.-Int. J. Ecol. 72, 110–119.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
12
Thompson, P.R., Piecuch, C.G., Merrifield, M.A., Mccreary, J.P., Firing, E., 2016. Forcing of recent decadal variability in the Equatorial and North Indian Ocean. J.
Geophys. Res.: Oceans 121, 6762–6778.
Turak, E., Brazill-Boast, J., Cooney, T., Drielsma, M., Delacruz, J., Dunkerley, G., Fernandez, M., Ferrier, S., Gill, M., Jones, H., Koen, T., Leys, J., Mcgeoch, M., Mihoub,
J.-B., Scanes, P., Schmeller, D., Williams, K., 2017. Using the essential biodiversity variables framework to measure biodiversity change at national scale. Biol.
Conserv. 213, 264–271.
Valiela, I., Bowen, J.L., York, J.K., 2001. Mangrove forests: one of the world’s threatened major tropical environments. BioScience 51, 807–815.
Walton, R., Baxter, R., Bunbury, N., Hansen, D., Fleischer-Dogley, F., Greenwood, S., Schaepman-Strub, G., 2019. In the land of giants: habitat use and selection of
the Aldabra giant tortoise on Aldabra Atoll. Biodivers. Conserv. 28, 3183–3198.
Woodroffe, C.D., Rogers, K., Mckee, K.L., Lovelock, C.E., Mendelssohn, I.A., Saintilan, N., 2016. Mangrove sedimentation and response to relative sea-level rise.
Ann. Rev. Mar. Sci. 8, 243–266.
Yates, K.K., Rogers, C.S., Herlan, J.J., Brooks, G.R., Smiley, N.A., Larson, R.A., 2014. Diverse coral communities in mangrove habitats suggest a novel refuge from
climate change. Biogeosciences 11, 4321–4337.
Young, I.R., Vinoth, J., Zieger, S., Babanin, A.V., 2012. Investigation of trends in extreme value wave height and wind speed. J. Geophys. Res.: Oceans 117, 1–13.
Zhao, Q., Stephenson, F., Lundquist, C., Kaschner, K., Jayathilake, D., Costello, M.J., 2020. Where marine protected areas would best represent 30% of ocean
biodiversity. Biol. Conserv. 244, 108536.
A. Constance, P.J. Haverkamp, N. Bunbury et al. Global Ecology and Conservation 27 (2021) e01564
13
