ArticlePDF Available

Evidence of Závora Bay as a critical site for reef manta rays, Mobula alfredi, in southern Mozambique

Article

Evidence of Závora Bay as a critical site for reef manta rays, Mobula alfredi, in southern Mozambique

Abstract

The largest known reef manta ray (Mobula alfredi) population in Africa has been monitored for over 20 years at several locations on the coast of the Inhambane Province in southern Mozambique. However, prior to this study, little had been reported on the population dynamics of M. alfredi from Závora, a remote bay in the region. Photographic mark‐recapture was used to investigate the size and structure of M. alfredi that aggregate 'at ‘Red Sands’; a reef cleaning station in Závora Bay. An 11‐year photographic dataset was used to identify 583 M. alfredi individuals between 2010‐2021. More than half of M. alfredi individuals were resighted at least once, with most encounters (up to 18 for one individual) occurring during the peak sighting period in July‐November each year. An even sex ratio was observed, 44% females and 50% males, with no significant difference in resightings between the sexes. Pollock's Robust Design population models were used to estimate annual abundance, emigration, annual apparent survival, and capture probability at Red Sands from July‐November over a six year period (2016‐2021). Abundance estimates varied year to year, ranging from 35 (95% CI 30‐45), up to 233 (95% CI 224‐249) M. alfredi individuals. Given the seasonal affinity of M. alfredi observed at Red Sands, this study highlights the importance of understanding fine‐scale site use within the larger home range of this population to develop local management strategies. This article is protected by copyright. All rights reserved.
REGULAR PAPER
Evidence of Závora Bay as a critical site for reef manta rays,
Mobula alfredi, in southern Mozambique
Michelle Carpenter
1
| Nakia Cullain
2,3
| Stephanie Kathleen Venables
2
|
Yara Tibiriçá
4
| Charles Griffiths
1
| Andrea Denise Marshall
2
1
Department of Biological Sciences, University
of Cape Town, Cape Town, South Africa
2
Marine Megafauna Foundation, West Palm
Beach, Florida, USA
3
Department of Biology, Dalhousie University,
Halifax, Nova Scotia, Canada
4
Departamento de Biología, Facultad de
Ciencias del Mar y Ambientales, Universidad
de Cádiz, Cádiz, Spain
Correspondence
Michelle Carpenter, Department of Biological
Sciences, University of Cape Town, Private
Bag, Rondebosch, Cape Town 7700,
South Africa.
Email: crpmic001@myuct.ac.za
Funding information
University of Cape Town, Grant/Award
Number: 2020 Faculty Fellowship
Abstract
The largest known reef manta ray (Mobula alfredi) population in Africa has been moni-
tored for more than 20 years at several locations on the coast of the Inhambane
Province in southern Mozambique. Nonetheless, before this study, little had been
reported on the population dynamics of M. alfredi from Závora, a remote bay in the
region. Photographic mark-recapture was used to investigate the size and structure
of M. alfredi that aggregate at Red Sands,a reef cleaning station in Závora Bay. An
11 year photographic data set was used to identify 583 M. alfredi individuals
between 2010 and 2021. More than half of M. alfredi individuals were resighted at
least once, with most encounters (up to 18 for one individual) occurring during the
peak sighting period in JulyNovember each year. An even sex ratio was observed,
44% females and 50% males, with no significant difference in resightings between
the sexes. Pollock's robust design population models were used to estimate annual
abundance, emigration, annual apparent survival and capture probability at Red Sands
from July to November over a 6 year period (20162021). Abundance estimates var-
ied year to year, ranging from 35 (95% C.I. [30, 45]) up to 233 (95% C.I. [224, 249])
M. alfredi individuals. Given the seasonal affinity of M. alfredi observed at Red Sands,
this study highlights the importance of understanding fine-scale site use within the
larger home range of this population to develop local management strategies.
KEYWORDS
mark recapture, Mobulidae, photo-ID, population demography, robust design, southern Africa
SIGNIFICANCE STATEMENT
The purpose of this study is to better understand areas of critical habitat for reef manta rays in
southern Mozambique, specifically, the use of a cleaning station called Red Sands(RS) in Závora
Bay in the Inhambane Province. This is important because cleaning stations, areas on reef where
fish remove parasites or dead skin off of a clientanimal, are sites that manta rays have been
found to repeatedly return to. The authors found a large seasonal abundance of manta rays at RS,
displaying the importance of this site for the greater southern Mozambique population. As RS is
currently unprotected, the study demonstrates the need for immediate site-specific protection.
Received: 15 March 2022 Accepted: 7 June 2022
DOI: 10.1111/jfb.15132
FISH
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2022 The Authors. Journal of Fish Biology published by John Wiley & Sons Ltd on behalf of Fisheries Society of the British Isles.
J Fish Biol. 2022;112. wileyonlinelibrary.com/journal/jfb 1
1|INTRODUCTION
Understanding population trends of large, migratory, marine species can
be challenging because of the difficulty in assessing habitat use within
their estimated home range. Knowledge of fine-scale visitation patterns
at specific locations through focused site-specific studies can be benefi-
cial to understanding population dynamics, identifying priority areas for
protection and developing local management regimes. The use of pho-
tographic mark-recapture is rapidly co-evolving in parallel with both
technological advancements in camera equipment, open-source web-
sites, algorithm development and growth in citizen science initiatives (e.
g., public data submissions and increased internet access), making it an
increasingly powerful tool for long-term population monitoring of wide-
ranging species across many taxa and habitat types (Karanth, 1995;
Dala-Corte et al., 2016;Marshall&Pierce,2012;McConkey,1999;
Schofield et al., 2008; Towner et al., 2013; Wiirsig & Jefferson, 1990).
Predictable aggregations at certain sites allow snapshots of population
sizes, trends and movement patterns of these elusive species.
Manta rays (order Myliobatiformes) are wide-ranging, pelagic filter
feeders that aggregate at inshore reefs, islands or seamounts (Couturier
et al., 2012;Murieet al., 2020;Harriset al., 2020). Their aggregations
can be directly related to foraging, or at cleaning stations in close prox-
imity to feeding grounds, to solicit cleaning services by symbiotic fish
and sometimes engage in social or reproductive interactions (Couturier
et al., 2011; Limbaugh, 1961; Stevens, 2016). The unique and stable
ventral markings of individual rays have facilitated photo-identification
(photo-ID) studies at these aggregation sites, providing the foundation
for manta ray research in many locations across the globe (e.g., Coutu-
rier et al., 2014; Deakos et al., 2011;Germanovet al., 2019;Harris
et al., 2020; Homma et al., 1999; Kumli & Rubin, 2008;Marshallet al.,
2011; Stevens, 2016). This technique has been used to assess home
range (Deakos et al., 2011;Kashiwagiet al., 2011), longevity (Clark,
2010; Couturier et al., 2014; Kashiwagi, 2014; Rubin, 2002), migration
patterns (Armstrong et al., 2019; Germanov & Marshall, 2014), site
affinity (Couturier et al., 2011; Germanov et al., 2019;Marshallet al.,
2011), reproductive ecology (Deakos et al., 2011;Marshall&Bennett,
2010a;Stevens,2016) and estimating abundance (Beale et al., 2019;
Couturier et al., 2014;Venables,2020). Regional reef manta ray
[Mobula alfredi (Kreft, 1868)] photo-ID databases vary substantially in
the total number of individuals identified over time, ranging from popu-
lations in the low hundreds (Axworthy et al., 2019;Carpentieret al.,
2019; Deakos et al., 2011; Kashiwagi, 2014;Peel,2019) to those in the
thousands (Armstrong et al., 2019;Stevens,2016; Venables, 2020).
Mark-recapture population modelling provides a tool through
which photo-ID data can be analysed to provide abundance estimates.
Predictable patterns in the use of critical habitats, such as cleaning
stations and feeding locations, make M. alfredi a suitable candidate for
this technique (Couturier et al., 2011; Venables, 2020). With an initial
photo of the ventral spot patterning signifying an individual's mark
and subsequent photos representing their recaptures,these data
are further analysed through models to estimate population parame-
ters (Couturier et al., 2014; Grusd et al., 2019). Previous mark-
recapture studies of M. alfredi have used various model types [i.e.,
CormackJollySeber (CJS); Petersen's method; Pollock's robust
design (PRD)] (Couturier et al., 2014; Deakos et al., 2011; Kitchen-
Wheeler et al., 2012; Marshall et al., 2011). More recently, the robust
design has proven useful, with the ability to account for temporary
emigration and capture heterogeneity, which are inherent in mobile
marine species (Couturier et al., 2014; Venables, 2020). PRD models
are characterised by marginal dependence between abundance and
survival estimators, as well as estimation of temporary emigration, all
of which improve the precision of population estimates and interpre-
tations of the relationship between abundance and survival (Grusd
et al., 2019; Kendall et al., 1995; Pollock, 1981; Pollock et al., 1990).
A 16 year study in Mozambique documented the largest photo-
identified population of M. alfredi in Africa, with the number of identi-
fied individuals currently reported to be 1209 (Marshall et al., 2011;
Venables, 2020). With increased annual sampling effort, M. alfredi
continues to exhibit long-term affinity to monitored cleaning stations
in Mozambique; some individuals returning to the same sites for more
than 15 years (Venables, 2020). Estimations of annual abundances in
the Praia do Tofo region peaked at 836 individuals in 20042005
(Venables, 2020). However, sightings declined of up to 88% between
2003 and 2011 (Rohner et al., 2013), and estimates of only 100 indi-
viduals sighted in Tofo after 2013 (Venables, 2020) have raised imme-
diate concern about the health of this population. M. alfredi is listed
on the IUCN Red List as Vulnerable (Marshall et al., 2018) and is glob-
ally threatened from direct harvesting of the gill plates for the Asian
market, by-catch, destructive fisheries methods and coastal develop-
ment, which in turn leads to increased boat strikes, habitat loss and
pollution (Couturier et al., 2012; Croll et al., 2016; Fernando & Stewart
2021; Lawson et al., 2017; O'Malley et al., 2016). Monitoring of popu-
lations of this threatened species is thus crucial for future IUCN Red
List assessments and further development of local management
actions, such as the designation of new marine-protected areas and
regulations surrounding fisheries and tourism operations.
Although the Tofo region has been consistently monitored since
2003, an aggregation site for M. alfredi 90 km south in Závora has not
yet been assessed. The authors of this study aim to better understand
this specific aggregation and assess its importance for the larger
southern Mozambican population of M. alfredi. They use an 11 year
photo-ID database of individuals to describe population demo-
graphics, site affinity and resightings data. Using PRD mark-recapture
modelling, they estimate annual abundance and population parame-
ters including apparent survival, emigration and recapture probability
at RS between 2016 and 2021. The findings can be used to inform
the development of local conservation strategies and guide the design
and implementation of spatial management approaches, such as
marine-protected areas, in the Závora Bay region of the Inhambane
coastline.
2|MATERIALS AND METHODS
2.1 |Study site
In Mozambique, M. alfredi is most commonly encountered in the
coastal waters of the Inhambane Province, particularly from the
2CARPENTER ET AL.
FISH
Bazaruto Archipelago in the north to Závora in the south (Figure 1).
This 350 km stretch of coastline joins a narrow continental shelf that
experiences regular upwelling events, resulting in productive waters
that attract several planktivorous species, including whale sharks,
Rhincodon typus, giant manta rays, Mobula birostris, and shortfin devil
rays, Mobula kuhlii (Rohner et al., 2013; Quartly & Srokosz, 2004).
M. alfredi has been monitored in Závora since 2010; nonetheless,
because of limited resources, the remoteness of this location and min-
imal tourism/recreational diving, a comprehensive sampling design
was not initiated until 2016. Red Sands (RS) is a rocky reef with scat-
tered corals and sponges, at 1218 m depth, c. 3 km offshore. The site
is characterised by variability in environmental conditions: with hori-
zontal visibility ranging 120 m, various levels of current and surge,
and sea temperatures ranging from 16 to 23C in the winter and up to
27C in the summer (Cullain, unpubl. data).
2.2 |Sampling effort and design
Photographic sampling by trained researchers was conducted at RS
on SCUBA diving between 2010 and 2021. Weather, logistical limita-
tions and COVID-19 restrictions prohibited consistent, daily sampling
effort throughout and between years. During each survey, teams of
two to eight divers swam a transect that covered all monitored clean-
ing stations that make up RS. Upon encountering an individual
M. alfredi, a photo of the unique markings on the ventral surface was
taken. Sex was determined by the presence of external claspers for
males and absence for females (Marshall & Bennett, 2010a). The
authors assessed male maturity by the size of reproductive organs,
individuals being considered adult once the claspers extended past
the posterior edge of the pectoral fins (Marshall & Bennett, 2010a).
Female maturity was determined by the observation of pregnancy
(when the abdomen was clearly expanded), or the presence of repro-
ductive scars usually on the left pectoral fin (99% lateralisation:
Marshall & Bennett, 2010a). A female that was not noticeably preg-
nant, nor had mating scars, was recorded as unknown maturity. The
total number of identified and unidentified individuals were pooled
for each day of sampling. Resightings of an individual were recorded
when identified more than 24 h after the last sighting.
The authors used photo-ID data collected from 2010 to 2021
to assess population demographics. Mean counts of individuals or
time periods between sightings were calculated to assess the num-
ber and resightings of males and females standard deviation).
Mark-recapture models require consistent survey effort and because
of the nature of the current data set, only the most recent 6 years
fit these criteria (82% of the total identifications). Therefore, data
collected at RS during the 5 month peak season (JulyNovember) of
20162021 were included in the PRD modelling, resulting in six pri-
mary periods (years) and 29 secondary periods (months; Table 1).
Of the total 583 M. alfredi individuals catalogued for Závora,
401 were photo-identified at RS between 2016 and 2021 and
were included in the PRD. Of these, individuals of undetermined
FIGURE 1 Study map of Závora Bay showing the location of the Red Sands cleaning station and the bathymetry of the bay
CARPENTER ET AL.3
FISH
sex (n=18) were removed for the final PRD analysis that included
sex as a covariate.
2.3 |Mark-recapture analysis
The authors used a PRD with Hugginsestimator to analyse 6 year
photographic mark-recapture data of M. alfredi at RS, Závora
(Huggins, 1989; Pollock et al., 1990). Models were assembled using
package RMark(Laake, 2013) in R Version 4.1.2 (R Core Team,
2021), the R interface to programme MARK (Cooch & White, 2006;
White & Burnham, 1999). The six peak seasons were selected as pri-
mary periods because of higher M. alfredi sightings; each winter sea-
son had five monthly secondary periods (JulyNovember), except
2016 which had four (JulyOctober), because of no survey effort in
November 2016. Mobula alfredi sightings were lower in December
June; therefore, these months were excluded to allow adequate time
between primary periods to detect fluctuations in the population
(Kendall, 1999; Silva et al., 2009). Few M. alfredi individuals (n=21)
were sighted at other reefs in Závora, but never encountered at RS;
therefore, these individuals were excluded from the present study
(Supporting Information Table S1).
PRD models have the following assumptions: all ventral markings
on M. alfredi individuals were unique and remained stable over time,
the population was open to immigration, emigration, natality and mor-
tality between years, full closure within the aggregation months and
equal survival probability on all individuals (Cooch & White, 2006;
Kendall et al., 1995; Smith et al., 2013; Williams et al., 2002). Closure
was not assumed at RS specifically; rather the authors of this study
assumed that the individuals encountered at RS remained in the
Závora Bay region during these time periods and were thus recap-
tured at RS.
The authors evaluated apparent survival between primary periods
as time-constant φ(), time varying φ(t) and with a group effect for sex
φ(sex). Models including time-varying survival consistently yielded
inestimable parameters. The authors deemed it appropriate to exclude
time-varying survival from the final model set due to the longevity of
M. alfredi once mature; previous studies on M. alfredi populations
found survival estimates close to 1.0 between years (Couturier et al.,
2014; Kitchen-Wheeler et al., 2012). The temporary emigration
parameter represents the probability of present individuals in the pop-
ulation being absent for capture in a specific period (Kendall et al.,
1997). This was assessed as Markovian (γ0and γ00), random γ(γ0=γ00 )
or none (γ0,γ00 =0). Capture pand recapture cprobabilities were
modelled as time-constant p(), time-varying per year p(y) and with
effects of sampling effort p(s). Equal capture and recapture probability
(p=c) was excluded from the final candidate model set due to inesti-
mable parameters resulting from the variability of encounters per sec-
ondary period. Parameter estimates were model averaged based on
the model weight. The authors evaluated the confidence interval (C.I.)
and standard error (S.E.) of each estimated parameter. The PRD analy-
sis was subsequently conducted on the same data with pooled sexes
to yield numbers for total population abundance across the primary
periods. AICc was used to evaluate the best model that fitted the
data, determined by the smallest AICc value (Burnham & Anderson,
2004). A MannWhitney U-test was conducted using the exac-
tRankTestsR package to analyse the effect of sex on the total num-
ber of recaptures during the study period, with individuals of
undetermined sex excluded from the analysis (Hothorn & Hornik
2021). Significance was accepted at P< 0.05.
2.4 |Lagged identification rates
The authors used lagged identification rates (LIR), the probability of
resighting an individual after a given time lag, to estimate site use of
M. alfredi at RS (Whitehead, 2001). The SOCPROG 2.9 programme
(Whitehead, 2009), specifically the movement analysismodule, was
used. The authors compared observed individual sighting data from
2016 to 2021, when there was consistent survey effort, to several
exponential mathematical models that represented various habitat
use scenarios, including permanent residency, emigration and mortal-
ity, emigration and reimmigration, emigration and reimmigration with
mortality and a cyclical pattern of appearance. The quasi-AIC values
were used to select the best supported model due to the overdisper-
sion of the data (Whitehead, 2007). Data were bootstrapped
100 times, with 1000 maximum evaluations, to estimate the standard
error and parameter precision (Buckland & Garthwaite, 1991;
Whitehead, 2001).
3|RESULTS
3.1 |Population demographics
Sampling effort at RS ranged from 0 to 37 dives per month, with one
or two dives of 4472 min duration conducted per day, resulting in a
monthly sampling effort of between 44 and 1252 min (Table 1). The
TABLE 1 Sampling effort (minutes)
during primary periods (years) and
secondary periods (months) used for
Pollock's robust design of Mobula alfredi
at Red Sands, Závora, Mozambique
Secondary period 2016 2017 2018 2019 2020 2021
July 68 459 584 833 182 420
August 318 445 570 1252 569 790
September 194 771 725 615 609 828
October 95 481 632 469 44 471
November 0; omitted for PRD 696 456 369 72 333
4CARPENTER ET AL.
FISH
number of M. alfredi individuals in the photo-ID database increased
throughout the study period with large numbers of new identifica-
tions between 20102011 and 20172018 (Figure 2). Until 2016, the
number of newly identified M. alfredi surpassed resights, and after
2017, the number of resighted individuals exceeded new IDs
(Figure 3). An average of three individuals (±4.29) and up to 61 individ-
uals (10% of the photographed population) in a single day were identi-
fied visiting RS during peak season (20162021; n=274 total
identifications in one JulyNovember season).
Between 2010 and 2021, the authors recorded 1509 encounters
of 583 individual M. alfredi in Závora Bay. More than half, 54% (n=
312), of these were resighted at least once; 57% (n=331) of individ-
uals were seen only within a single year and 43% (n=252) across
multiple years. The mean time interval between initial and subsequent
sightings was 455 days (±694), with 10 individuals recorded with a
resighting interval of 1000 days or more, and a maximum of
10.9 years (3996 days) between resightings. The population exhibited
an even sex ratio, whereby 44% (n=255) were females, 50% were
males (n=295) and sex could not be determined for 6% (n=33).
There was no significant difference in the mean number of sightings
between females and males 2.82 (±2.34) and 2.53 (±2.34), respec-
tively (MannWhitney U-test; P=0.7981). Although more males than
females were resighted (males, n=171; females, n=139) in Závora,
individuals in the database that were sighted six or fewer times con-
sisted of mostly males, whereas individuals sighted seven times or
more during the study period were almost all females (Figure 4). Only
mature females had more than 10 sightings during the study period,
with the most resighted individual identified 18 times.
The authors recorded 44 pregnancies across 36 females during
the 11 year study period. Five individuals were observed to be preg-
nant on more than one occasion, with a mean postpartum interval of
33.4 (±8.8 months; Supporting Information Figure S1). About 56%
0
0
100
100
200
200
300
300
400
400
500
500
600
600
700
700
Number of idenfied individualsNumber of idenfied individuals
Time
Time
2010
2010 2011
2011 2012
2012 2013
2013 2014
2014 2015
2015 2016
2016 2017
2017 2018
2018 2019
2019 2020
2020 2021
2021
FIGURE 2 Discovery curve of
identified Mobula alfredi individuals
from 2010 to 2021 in Závora,
Mozambique
0
1
2
3
4
5
6
0
50
100
150
200
250
300
350
1 2 3 4 5 6
2017
Rao of resights/new idenficaons
deifitnedislaudividniforebmuN
Year of study
2018
2019
2016 2020 2021
FIGURE 3 Total number of newly
identified Mobula alfredi () and the
total number of resights ( ) in each
primary period of the study, and the
ratio of resights/new at Red Sands
(), in Závora, Mozambique
CARPENTER ET AL.5
FISH
(n=326) of M. alfredi were defined as mature, 49% (n=288) being
males and 7% females (n=38), although for most females maturity
could not be determined (n=250).
3.2 |Robust design
Eighteen candidate models were evaluated in the PRD analysis.
Models that integrated Markovian emigration, with capture probabil-
ity varying by sampling effort, were best supported (Table 2). The best
supported PRD model consisted of sex-dependent survival,
Markovian temporary emigration and an effect of sampling effort on
capture probability (Table 2). Annual apparent survival was estimated
higher for males than females, at 0.848 (0.09; 95% C.I. [0.597, 0.954])
and 0.823 (0.08; 95% C.I. [0.602, 0.935]), respectively (Supporting
Information Table S2). Capture probability dependent on sampling
effort fluctuated between primary periods, with the highest in 2020
(0.69; 95% C.I. [0.60, 0.76]) and lowest in 2016 (0.16; 95% C.I. [0.14,
0.18]) (Supporting Information Table S2). Overall annual abundances
ranged from 35 (95% C.I. [30, 45]) in 2016 to 233 (95% C.I. [224, 249])
FIGURE 4 Total number of Mobula alfredi identified at Red Sands in Závora, Mozambique: female (black ), male (grey ) and undetermined
(white )
TABLE 2 Selection of Pollock's
robust design (n=18) candidate models
for estimations of population size (N),
survival (φ; constant or sex varying),
temporary emigration (γ00 and γ0;
Markovian, random or none), capture (p)
and recapture (c) probabilities (constant,
with response to capture, varying by year
or varying by sampling effort) of Mobula
alfredi individuals that use Red Sands in
Závora, Mozambique
Model Rank npar AICc ΔAICc Model weight Deviance
φ
Sex
γ00
M
γ0
M
ρ
s
=c() 1 15 4137.92 0.00 0.846 5047.43
φ.γ00
M
γ0
M
ρ
s
=c() 2 13 4141.32 3.40 0.154 5039.88
φ
Sex
γ00
R
=γ0
R
ρ
s
=c() 3 7 4195.45 57.53 0.000 5113.87
φ.γ00
R
=γ0
R
ρ
s
=c() 4 5 4198.89 60.97 0.000 5121.37
φ
Sex
γ00
0
=γ0
0
ρ
y
=c() 5 9 4236.52 98.60 0.000 5150.85
φ
Sex
γ00
R
=γ0
R
ρ
y
=c() 6 10 4238.52 100.65 0.000 5150.85
φ.γ00
0
=γ0
0
ρ
y
=c() 7 7 4240.40 102.49 0.000 5158.82
φ.γ00
R
=γ0
R
ρ
y
=c() 8 8 4242.44 104.53 0.000 5158.82
φ
Sex
γ00
M
=γ0
M
ρ
y
=c() 9 16 4242.52 104.60 0.000 5142.40
φ.γ00
M
γ0
M
ρ
y
=c() 10 14 4245.91 107.99 0.000 5149.95
φ
Sex
γ00
0
=γ0
0
ρ
s
=c() 11 6 4295.44 157.52 0.000 5215.90
φ.γ00
0
=γ0
0
ρ
s
=c() 12 4 4299.05 161.13 0.000 5223.56
φ
Sex
γ00
R
=γ0
R
ρ=c() 13 5 4352.64 214.73 0.000 5275.13
φ.γ00
R
=γ0
R
ρ=c() 14 3 4356.10 218.18 0.000 5282.63
φ
Sex
γ00
M
γ0
M
ρ=c() 15 11 4359.07 221.16 0.000 5269.31
φ.γ00
M
=γ0
M
ρ=c() 16 9 4362.09 224.17 0.000 5276.43
φ.γ00
0
=γ0
0
ρ=c() 17 4 4364.08 226.16 0.000 5288.59
φ.γ00
0
=γ0
0
ρ
s
=c() 18 2 4367.73 229.81 0.000 5296.27
6CARPENTER ET AL.
FISH
in 2017 (Table 3). Differences in annual abundance estimates were
marginal for males and females, at 20115 and 13110, respectively
(Figure 5).
3.3 |Lagged identification rates
The best fit LIR models were F and H, which were practically equiva-
lent at <2 based on the ΔQAIC values (Supporting Information Table
S3). Nonetheless, model H made biological sense for the data, which
incorporated immigration, reimmigration and permanent emigration
and/or mortality (Table 4). Approximately 58 individuals (S.E.=16.19,
95% C.I. [35.37, 95.71]) were estimated to be present in the study area
on a given day. Mobula alfredi individuals had a mean residence time
of 4 days (S.E.=27.82, 95% C.I. [1.53, 80.02] days), with 10 days
(S.E.=415.81, 95% C.I. [5.22, 155.49] days) away from the study area.
Permanent emigration and/or mortality was estimated at 0.00029
(S.E.=0.00029, 95% C.I.[0.00024, 0.00070]). The plotted LIR curve
decreased rapidly from the date after identification indicating that
most individuals were transient to RS, with temporal annual use of RS
TABLE 3 Population size (N) for
males, females and overall Mobula alfredi
at Red Sands in Závora, Mozambique,
from the weighted average of the best-fit
models, and the number of uniquely
photo-identified individuals between July
2016 and November 2021
Sex Method Year Weighted average S.E. 95% C.I.
Male PRD 2016 20 2.71 1728
2017 115 3.79 110126
2018 61 2.55 5869
2019 59 2.50 5666
2020 106 3.60 102117
2021 27 1.61 2633
Photo-ID 20162021 215
Female PRD 2016 13 2.19 1121
2017 110 3.67 105120
2018 36 1.88 3442
2019 54 2.38 5262
2020 74 2.87 7083
2021 18 1.31 1724
Photo-ID 20162021 168
Overall PRD 2016 35 3.63 3045
2017 233 6.16 224249
2018 102 3.51 98112
2019 114 3.77 109125
2020 185 5.22 178199
2021 49 2.24 4656
Photo-ID 20162021 401
0
20
40
60
80
100
120
140
2016 2017 2018 2019 2020 2021
Annual abundance
FIGURE 5 Estimates of yearly
abundance of Mobula alfredi (N) ± 95%
C.I. at Red Sands in Závora,
Mozambique from 2016 to 2021,
estimated from the best-fit model
(φSexγ00Mγ0Mρs=c()) with sex as a
covariate (a; grey =male,
black =female)
CARPENTER ET AL.7
FISH
(Figure 6). The plotted LIR curve then levels and decreases until the
end of the study period suggesting emigration and subsequent return
and/or return to the area each season. The declining rate of the LIR
displays individual dispersal and the shape of the curve indicates a
short residency period at the aggregation site, with reimmigration at a
later stage by a proportion of the individuals (Whitehead, 2009).
4|DISCUSSION
In this study, the authors describe a seasonal, site-specific aggregation
of M. alfredi at RS, an aggregation site in Závora Bay, showing this site
to be an important regional habitat for the species. Half of the
583 identified individuals displayed affinity to this location (54%
resighting rate) despite belonging to a larger, wide-ranging population
(Venables et al., 2021). A total of 312 individuals returned to RS on
multiple occasions, up to a maximum of 18 times during the 11 year
study period, with no difference in site use between the sexes.
Estimated annual abundance at RS ranged from 35 to 233 individuals
during the JulyNovember season. This seasonal peak in abundance
combined with the resighting rate of individuals reflects the seasonal
importance of the Závora region for a proportion of the larger
M. alfredi population in southern Mozambique. Compared with other
monitored M. alfredi populations that found constant survival to be
0.9 (Deakos et al., 2011; Couturier et al., 2014), there was a lower
apparent survival (males, 0.848; females, 0.823) at RS, suggesting
transience during periods when conditions are not favourable for visi-
tation to Závora Bay. This was further supported by an average resi-
dence time of 4 days estimated using LIR.
Mobula alfredi habitat use varies spatially and temporally at other
well-studied locations (Armstrong et al., 2020; Dewar et al., 2008),
which is also evident in the findings. Site affinity has been globally
reported for M. alfredi, with individuals consistently returning to clean-
ing station reefs over long periods of time, up to 30 years (Couturier
et al., 2014; Couturier et al., 2018; Dewar et al., 2008; Venables et al.,
2020). The results show that M. alfredi presently aggregate at one
shallow reef in Závora, rather than a collection of deeper reefs (25
30 m) as documented in Tofo (Marshall et al., 2011; Venables, 2020).
Higher resighting rates were found at RS (54%) compared with clean-
ing stations in eastern Australia (Couturier et al., 2011) and Tofo
(Venables, 2020), but less than Indonesia, Hawaii and the Maldives
(Couturier et al., 2011; Deakos et al., 2011; Germanov et al., 2019;
Harris & Stevens, 2021). The authors found seasonal peaks in sight-
ings at Závora, as opposed to year-round sightings at other identified
M. alfredi aggregations in the Inhambane Province (Venables, 2020).
Island populations of M. alfredi in the western Indian Ocean also
exhibit year-round site use, with seasonally driven peaks related to
monsoon winds (Peel et al., 2019; Stevens, 2016). Although oceanic
processes such as monsoonal shifts, seasonal-driven currents and
tides affect M. alfredi site use in the Komodo National Park, Indonesia
and eastern Australia, ontogenetic patterns were found to influence
habitat use in Nusa Penida, Indonesia, the Gulf of Mexico and Hawaii,
and the authors suggest these could be potential drivers of M. alfredi
use of RS (Armstrong et al., 2020; Axworthy et al., 2019; Dewar et al.,
2008; Germanov et al., 2019; Harris et al., 2021; Stewart et al., 2018).
Further telemetry studies and direct assessments of zooplankton
abundance and composition may be needed to identify the drivers of
M. alfredi visitation to Závora and improve our understanding of them.
Mozambique and Australia are among few places in the world
where M. alfredi live along an extended continental coastline, which
may explain the observed transience of individuals at these locations
(Armstrong et al., 2020; Venables, 2020). Long-term monitoring of
both of these M. alfredi populations show habitat use of an entire
coastline where movement patterns may result from temporal shifts
in productivity, as opposed to island habitats, which may have more
TABLE 4 Model selection for lagged identification rate of reef
manta rays in Závora Bay, Mozambique (20162021)
Model Model description ΔQAIC
A Closed (1/a1=N) 89.60
B Closed (a1=N) 89.60
C Emigration/mortality (a1=emigration rate;
1/a2=N)
46.05
D Emigration/mortality (a1=N;a2=mean
residence time)
46.05
E Emigration +reimmigration (a1=emigration
rate; a2/(a2+a3) =proportion of population
in study area at any time)
28.15
F Emigration +reimmigration (a1=N;a2=res
time in; a3=res time out
0.95
G Emigration +reimmigration +mortality 9.93
H Emigration +reimmigration +mortality a1=N;
a2=res time in; a3=res time out;
a4=mortality
0.00
I Cyclical a1cos (a2td)+a3 93.56
FIGURE 6 Empirical data (mean ± S.E.) for the lagged
identification rate, the probability of re-identifying Mobula alfredi in
Závora Bay, Mozambique, over increasing time periods, with fitted
emigration plus reimmigration plus mortality model
8CARPENTER ET AL.
FISH
reliable food sources in the area (Armstrong et al., 2020; Peel et al.,
2019, Peel et al., 2020; Rohner et al., 2013; Venables, 2020). The lon-
gest point-to-point migration reported for an individual M. alfredi was
1150 km in Australia (Armstrong et al., 2019), whereas telemetry stud-
ies in Mozambique found rapid movements of up to 90 km in a single
day (Venables et al., 2020). The Inhambane coastline consists of a nar-
row continental shelf with mesoscale, eddy-driven upwelling in the
Mozambican channel, which contributes to productivity, thus its fluc-
tuation may drive M. alfredi movements up and down the coast
(Quartly & Srokosz, 2004; Rohner et al., 2013). The support of
Markovian emigration in the PRD models further implies that some
M. alfredi individuals leave for multiple seasons and eventually return.
Variations in movement and visitation patterns between the years
could be attributed to oceanic processes that affect zooplankton
patchiness and distribution including El Niño Southern Oscillation
(ENSO) and/or dipole effects (Beale et al., 2019; Folt & Burns, 1999;
Whitney & Crow, 2007).
The annual abundance of M. alfredi identified at this single reef in
Závora is high when compared to aggregations in Hawaii (Axworthy
et al., 2019; Deakos et al., 2011), the Seychelles (Peel, 2019) and
Japan (Kashiwagi, 2014), and lower when compared to the seasonal,
site-specific, aggregation at Lady Elliot Island, Australia (Couturier
et al., 2014). Previously published abundance estimates (20032012)
from the Tofo region of the Inhambane Province (Marshall et al.,
2011; Venables, 2020) were larger than our estimates for Závora, but
with fewer overall resightings. Nonetheless, after 2013, Venables
(2020) found <100 M. alfredi individuals to be using the reefs around
Tofo, whereas in the present study abundances at RS were consis-
tently >100 from 2017 to 2020. Abundance estimates between 2016
and 2021 varied noticeably, with 2017 and 2020 having greater cap-
ture rates compared to other primary periods. Such variation each
year may be attributed to productivity shifts or ontogenetic factors,
although further study is required to confirm this.
Both sexes displayed similar use of RS in contrast to many moni-
tored locations where females are more frequently resighted
(Marshall et al., 2011; Setyawan et al., 2018). The observed even sex
ratio in this study supported preliminary findings by Venables (2020),
but in contrast to the 61% female-bias found in Tofo (Venables,
2020). Often when a greater geographic area is monitored with
increased information on the metapopulation, even sex ratios have
been reported, including in the Maldives and French Polynesia, or
more uncommonly at a single site (Carpentier et al., 2019; Perryman
et al., 2019; Stevens, 2016; Venables, 2020). Male M. alfredi were pri-
marily mature at RS, with several juveniles that later returned as
mature over the course of the study. The main difference in site use
by the sexes was that specific mature females (n=13) were resighted
on 10 or more occasions, with some of these individuals encountered
at RS over a duration of almost 11 years. Our findings of an aggrega-
tion of M. alfredi returning to this exact reef may reflect the impor-
tance of this site for sociality and/or courtship ritual (Perryman et al.,
2021; Stevens et al., 2018; Thorburn et al., 2019).
An estimated average residence time of 4 days from the LIR anal-
ysis suggests M. alfredi individuals to visit the study site for a short
period of time in peak season and then leave. The large ranges in stan-
dard error and 95% C.I. in the LIR analysis are likely due to the individ-
ual variability in sightings from the empirical data, and the variation in
sightings year to year, which was also apparent in the PRD analysis.
The residence time to RS was lower than M. alfredi populations
around islands in French Polynesia (range 66130 days) and Coral
Bay, Australia (56 days); nonetheless, residence time out was lower
than French Polynesia (range 59117) and Coral Bay, Australia
(92 days), suggesting that in Závora, individuals are more likely to
move in and out of the study area even during peak season
(Armstrong et al., 2020; Carpentier et al., 2019).
Challenging weather conditions, the logistics of operating in a
remote location and resource availability contributed to uneven sam-
pling effort throughout the study period. The authors accounted for
this in the PRD analysis by modelling capture probability with an
effect of sampling effort. Further limitation in sampling for the PRD
included times when an individual was present at the aggregation but
not photographed. Such limitations are characteristic of M. alfredi
photo-ID studies, including potential violations of model assumptions
(i.e., survival probability being the same for all individuals) (Deakos
et al., 2011; Couturier et al., 2014; Venables, 2020). Given the level of
anthropogenic impact (Venables, 2020) and predation pressure
(Marshall & Bennett, 2010b) affecting southern Mozambique may
result in similar survivorship of this specific aggregation. Considering
their longevity, this 6 year analysis is brief relative to the life span of
M. alfredi. Nevertheless, the PRD in this context provided baseline
estimations of the number of M. alfredi that use RS in Závora,
Mozambique, an area which is currently unprotected.
More than 20 years of research along the Inhambane Province
has identified the largest known M. alfredi population in Africa, yet
with drastic declines in sightings, of up to 88% (Marshall et al., 2011;
Rohner et al., 2013; Venables, 2020). This population is now stated to
be of immediate conservation concern by local and international sci-
entists (Peel, 2019; Rohner et al., 2013; Tibiriçá et al., 2011; Venables,
2020). Mobula alfredi is listed in Appendix II (2013) of the Convention
for International Trade in Endangered Species (CITES) and in Appendi-
ces I and II (2014) of the Conservation of Migratory Species of Wild
Animals (CMS). Nationally, manta species were protected under
Mozambican law in 2017 (Law 5/2017) which banned fishing of
CITES-listed species; nonetheless, little was enforced (Boletim da
Republica May 2017; Venables, 2020). As a vulnerable (Marshall et al.,
2018) and economically important species (Venables et al., 2016),
M. alfredi officially received national protection in 2021 (Boletim da
República, 2020); nonetheless, along the Inhambane coastline, they
remain under threat from indiscriminate netting and longlining, partic-
ularly in the south of the province (Marshall et al., 2011; Temple et al.,
2018). To increase protection of this mobile species in Mozambique, it
is essential to focus on priority habitats where they might be at risk,
such as RS, where a seasonal inshore aggregation occurs every year.
At present, the majority of protected critical habitat is concentrated in
the north of the province in the Bazaruto Archipelago (Pelegrín et al.,
2015). Although part of a single breeding population, photo-ID and
acoustic telemetry have indicated preferential habitat use to different
CARPENTER ET AL.9
FISH
sites, meaning that M. alfredi individuals using the northern regions do
not show equal visitation to the southern regions of Tofo and Závora
(Venables et al., 2020). Anthropogenic pressures from fishing continue
to impact the southern M. alfredi in most of their home range, includ-
ing Závora, which is at the southern extent of the area where they are
most commonly encountered in Mozambique. Therefore, the authors
recommend immediate, site-specific protection of key habitats in the
south, such as RS, as an essential step for conservation management.
They also advocate for the design and implementation of a standalone
marine-protected area in Závora Bay to protect the larger critical habi-
tat for elasmobranchs in this southern region of the Inhambane
Province (O'Connor & Cullain, 2021). The Inhambane coast was
declared a Mission Blue Hope Spot in 2022 in recognition of its diver-
sity of threatened species, and the government of Mozambique has
proposed to implement a large seascape-type environmental protec-
tion area (EPA) from the Bazaruto Archipelago southwards towards
Závora (Administraç˜ao Nacional das
´
Areas de Conservaç˜ao and
Conservation International, 2020). Given the trajectory of the decline
of the M. alfredi population along this coastline and the seasonal
importance of this habitat, the authors advise the protection of
Závora Bay be prioritised during this process.
AUTHOR CONTRIBUTIONS
M.C. conceived the central idea of the article. Data collected by
N.R.C. and Y.T. with some contributions by all authors. Analysis by
M.C. with input by S.K.V. Written by M.C. with input from all other
authors.
ACKNOWLEDGEMENTS
We thank Závora Lab and MAR Expeditions for starting the research and
management of the data. We thank Ms. Linda Cooke from Závora Lodge,
Mr. Paul Vukovich, the University of Cape Town Department of Biological
Sciences and the Marine Megafauna Foundation for supporting this work.
We thank Dr. Simon Pierce for his advice on the LIR analysis.
ORCID
Michelle Carpenter https://orcid.org/0000-0001-7562-924X
REFERENCES
Administraç˜ao Nacional das
´
Areas de Conservaç˜ao & Conservation Inter-
national (2020, November 25). Visioning the Save River - Pomene
Environmental Protection Area: Workshop report. Stakeholder work-
shop held at Vilanculos Beach Lodge, Vilanculos, Mozambique.
Armstrong, A. J., Armstrong, A. O., Bennett, M. B., McGregor, F.,
Abrantes, K. G., Barnett, A., Dudgeon, C. L. (2020). The geographic
distribution of reef and oceanic manta rays (Mobula alfredi and Mobula
birostris) in Australian coastal waters. Journal of Fish Biology,96(3),
835840.
Armstrong, A. J., Armstrong, A. O., McGregor, F., Richardson, A. J.,
Bennett, M. B., Townsend, K. A., Dudgeon, C. L. (2020). Satellite tag-
ging and photographic identification reveal connectivity between two
UNESCO world heritage areas for reef manta rays. Frontiers in Marine
Science,7, 725. https://doi.org/10.3389/fmars.2020.00725.
Armstrong, A. O., Armstrong, A. J., Bennett, M. B., Richardson, A. J.,
Townsend, K. A., & Dudgeon, C. L. (2019). Photographic identification
and citizen science combine to reveal long distance movements of
individual reef manta rays Mobula alfredi along Australia's east coast.
Marine Biodiversity Records,12(1), 14.
Axworthy, J. B., Smith, J. M., Wing, M. S., & Quinn, T. P. (2019). Sex biased
individual variation in movement patterns of a highly mobile, near-
shore marine planktivore, the reef manta ray Mobula alfredi.Journal of
Fish Biology,95(6), 13991406.
Beale, C. S., Stewart, J. D., Setyawan, E., Sianipar, A. B., & Erdmann, M. V.
(2019). Population dynamics of oceanic manta rays (Mobula birostris)in
the Raja Ampat archipelago, West Papua, Indonesia, and the impacts
of the El Niñosouthern oscillation on their movement ecology. Diver-
sity and Distributions,25(9), 14721487.
Buckland, S. T., & Garthwaite, P. H. (1991). Quantifying precision of mark-
recapture estimates using the bootstrap and related methods. Biomet-
rics,47, 255268.
Burnham, K. P., & Anderson, D. R. (2004). Multimodel inference: Under-
standing AIC and BIC in model selection. Sociological Methods and
Research,33(2), 261304.
Carpentier, A. S., Berthe, C., Ender, I., Jaine, F. R., Mourier, J., Stevens, G.,
Clua, E. (2019). Preliminary insights into the population characteris-
tics and distribution of reef (Mobula alfredi) and oceanic (M. birostris)
manta rays in French Polynesia. Coral Reefs,38(6), 11971210.
Clark, T. B. (2010). Abundance, home range, and movement patterns of
manta rays (Manta alfredi,M. birostris) in Hawaii. PhD thesis, University
of Hawaii at Manoa, School of Ocean and Earth Sciences.
Cooch, E. G., & White, G. C. (2006). Program MARK: A gentle introduction. Fort
Collins, Colorado: Colorado State University Available at http://www.
phidot.org/software/mark/docs/book (accessed January 2, 2022).
Couturier, L. I., Dudgeon, C. L., Pollock, K. H., Jaine, F. R. A.,
Bennett, M. B., Townsend, K. A., Richardson, A. J. (2014). Population
dynamics of the reef manta ray, Manta alfredi in eastern Australia.
Coral Reefs,33(2), 329342.
Couturier, L. I., Jaine, F. R., Townsend, K. A., Weeks, S. J.,
Richardson, A. J., & Bennett, M. B. (2011). Distribution, site affinity
and regional movements of the manta ray, Manta alfredi (Krefft, 1868),
along the east coast of Australia. Marine and Freshwater Research,
62(6), 628637.
Couturier, L. I. E., Marshall, A. D., Jaine, F. R. A., Kashiwagi, T., Pierce, S. J.,
Townsend, K. A., Richardson, A. J. (2012). Biology, ecology and con-
servation of the mobulidae. Journal of Fish Biology,80(5), 10751119.
Couturier, L. I. E., Newman, P., Jaine, F. R. A., Bennett, M. B., Venables, W.
N., Cagua, E. F., Richardson, A. J. (2018). Variation in occupancy and
habitat use of Mobula alfredi at a major aggregation site. Marine Ecol-
ogy Progress Series,599, 125145.
Croll, D. A., Dewar, H., Dulvy, N. K., Fernando, D., Francis, M. P., Galván-
Magaña, F., Newton, K. M. (2016). Vulnerabilities and fisheries
impacts: The uncertain future of manta and devil rays. Aquatic Conser-
vation: Marine and Freshwater Ecosystems,26(3), 562575.
Boletim da República. Decreto no. 5/2017. Aprovado o Regulamento da
CITES, Maputo: República de Moçambique; 2017.
Boletim da República. Decreto no. 89/2020. Aprovado o Regulamento da
Pesca Marítimo (REPMAR), Maputo: República de Moçambique; 2020.
Dala-Corte, R. B., Moschetta, J. B., & Becker, F. G. (2016). Photo-
identification as a technique for recognition of individual fish: A test
with the freshwater armored catfish Rineloricaria aequalicuspis Reis &
Cardoso, 2001 (Siluriformes: Loricariidae). Neotropical Ichthyology,
14(1), e150074.
Deakos, M. H., Baker, J. D., & Bejder, L. (2011). Characteristics of a manta
ray Manta alfredi population off Maui, Hawaii, and implications for
management. Marine Ecology Progress Series,429(1), 245260.
Dewar, H., Mous, P., Domeier, M., Muljadi, A., Pet, J., & Whitty, J. (2008).
Movements and site fidelity of the giant manta ray, Manta birostris,in
the komodo Marine Park, Indonesia. Marine Biology,155(2), 121133.
Fernando, D., & Stewart, J. D. (2021). High bycatch rates of manta and
devil rays in the small-scaleartisanal fisheries of Sri Lanka. PeerJ,9,
e11994.
10 CARPENTER ET AL.
FISH
Folt, C. L., & Burns, C. W. (1999). Biological drivers of zooplankton patchi-
ness. Trends in Ecology and Evolution,14(8), 300305.
Germanov, E. S., Bejder, L., Chabanne, D. B. H., Dharmadi, D.,
Hendrawan, I. G., Marshall, A. D., Loneragan, N. R. (2019). Contrast-
ing habitat use and population dynamics of reef manta rays within the
Nusa Penida marine protected area, Indonesia. Frontiers in Marine Sci-
ence,6(215), 1-17. https://doi.org/10.3389/fmars.2019.00215.
Germanov, E. S., & Marshall, A. D. (2014). Running the gauntlet: Regional
movement patterns of Manta alfredi through a complex of parks and
fisheries. PLoS One,9(10), e110071.
Grusd, S. P., Moloney, C. L., Distiller, G., Watson, R. G. A., Cowley, P. D., &
Gennari, E. (2019). Using mark- recapture methods to estimate popula-
tion size and survival of pyjama sharks Poroderma africanum in Mossel
Bay, South Africa. African Journal of Marine Science,41(4), 361372.
Harris, J. L., Hosegood, P., Robinson, E., Embling, C. B., Hilbourne, S., &
Stevens, G. M. W. (2021). Fine- scale oceanographic drivers of reef
manta ray (Mobula alfredi) visitation patterns at a feeding aggregation
site. Ecology and Evolution,11(9), 45884604. https://doi.org/10.
1002/ece3.7357.
Harris, J. L., McGregor, P. K., Oates, Y., & Stevens, G. M. (2020). Gone with
the wind: Seasonal distribution and habitat use by the reef manta ray
(Mobula alfredi) in the Maldives, implications for conservation. Aquatic
Conservation: Marine and Freshwater Ecosystems,30(8), 16491664.
Harris, J. L., & Stevens, G. M. (2021). Environmental drivers of reef manta
ray (Mobula alfredi) visitation patterns to key aggregation habitats in
the Maldives. PLoS One,16(6), e0252470.
Homma, K., Maruyama, T., Itoh, T., Ishihara, H., & Uchida, S. (1999). Biol-
ogy of the manta ray, Manta birostris Walbaum, in the indo-Pacific. In
B. Seret & J. Y. Sire (Eds.), Indo-Pacific fish biology: Procurement 5th
international conference indo-Pacific fishes, Noumea, 1997 (pp. 209
216). Paris: Ichthyological Society of France.
Hothorn, T., & Hornik, K. (2021). exactRankTests: Exact Distributions for
Rank and Permutation Tests. R package version 0.8-34. https://CRAN.
R-project.org/package=exactRankTests
Huggins, R. M. (1989). On the statistical analysis of capture experiments.
Biometrics,76(1), 133140.
Karanth, K. U. (1995). Estimating tiger Panthera tigris populations from
camera-trap data using capture- recapture models. Biological Conserva-
tion,71(3), 333338.
Kashiwagi, T. (2014). Conservation biology and genetics of the largest liv-
ing rays: Manta rays. PhD thesis, University of Queensland, Molecular
Fisheries Laboratory.
Kashiwagi, T., Marshall, A. D., Bennett, M. B., & Ovenden, J. R. (2011).
Habitat segregation and mosaic sympatry of the two species of manta
ray in the Indian and Pacific oceans: Manta alfredi and M. birostris.
Marine Biodiversity Records,4(e53), 1-8. https://doi.org/10.1017/
S1755267211000479.
Kendall, W. L. (1999). Robustness of closed capturerecapture methods to
violations of the closure assumption. Ecology,80(8), 25172525.
Kendall, W. L., Nichols, J. D., & Hines, J. E. (1997). Estimating temporary
emigration using capturerecapture data with Pollock's robust design.
Ecology,78(2), 563578.
Kendall, W. L., Pollock, K. H., & Brownie, C. (1995). A likelihood-based
approach to capture-recapture estimation of demographic parameters
under the robust design. Biometrics,51(1), 293308.
Kitchen-Wheeler, A. M., Ari, C., & Edwards, A. J. (2012). Population esti-
mates of Alfred mantas (Manta alfredi) in Central Maldives atolls:
North male, Ari and Baa. Environmental Biology of Fishes,93(4),
557575.
Kreft, G. (1868). Deratoptera alfredi (Prince Alfred's ray). Illustrated Sydney
News,5(50), 116.
Kumli, K. R., & Rubin, R. D. (2008). Photo-identification of the manta ray,
Manta birostris, in the Revillagigedo Islands, Mexico. In Joint meeting of
ichthyologists and herpetologists (JMIH) abstract book (p. 23
´
A28). Mon-
treal: American Elasmobranch Society Devil Ray Symposium.
Laake, J. L. (2013). RMark: an R interface for analysis of capture-recapture
data with MARK. AFSC Processed Report 2013-01, prepared for the
U.S. Department of Commerce. Seattle, Washington: Alaska Fisheries
Science Center, NOAA, National Marine Fisheries Service.
Lawson, J. M., Fordham, S. V., O'Malley, M. P., Davidson, L. N.,
Walls, R. H., Heupel, M. R., Ender, I. (2017). Sympathy for the devil:
A conservation strategy for devil and manta rays. PeerJ,5, e3027.
Limbaugh, C. (1961). Cleaning symbiosis. Scientific American,205(2), 4249.
Marshall, A., Kashiwagi, T., Bennett, M. B., Deakos, M., Stevens, G.,
McGregor, F., Clark, T., Ishihara, H., & Sato, K. (2018). Mobula alfredi
The IUCN Red List of Threatened Species 2018 (amended version of
2011 assessment) e.T195459A126665723. https://doi.org/10.2305/
IUCN.UK.20112.RLTS.T195459A126665723.en.
Marshall, A. D., & Bennett, M. B. (2010a). Reproductive ecology of the reef
manta ray Manta alfredi in southern Mozambique. Journal of Fish Biol-
ogy,77(1), 169190.
Marshall, A. D., & Bennett, M. B. (2010b). The frequency and effect of
shark-inflicted bite injuries to the reef manta ray Manta alfredi.African
Journal of Marine Science,32(3), 573580.
Marshall, A. D., Dudgeon, C. L., & Bennett, M. B. (2011). Size and structure
of a photographically identified population of manta rays Manta alfredi
in southern Mozambique. Marine Biology,158(5), 11111124.
Marshall, A. D., & Pierce, S. J. (2012). The use and abuse of photographic
identification in sharks and rays. Journal of Fish Biology,80(5), 1361
1379.
McConkey, S. D. (1999). Photographic identification of the New Zealand
Sea lion: A new technique. New Zealand Journal of Marine Freshwater
Research,1(33), 6366.
Murie, C., Spencer, M., & Oliver, S. P. (2020). Current strength, tempera-
ture, and bodyscape modulate cleaning services for giant manta rays.
Marine Biology,167(5), 111.
OConnor, B., & Cullain, N. (2021). Distribution and community structure
of at-risk and Data Deficient elasmobranchs in Zavora Bay, Mozam-
bique. African Journal of Marine Science,43(4), 521532.
O'Malley, M. P., Townsend, K. A., Hilton, P., Heinrichs, S., & Stewart, J. D.
(2016). Characterization of the trade in manta and devil ray gill plates
in China and Southeast Asia through trader surveys. Aquatic Conserva-
tion of Marine Freshwater Ecosystems,27(2017), 394414.
Peel, L. R. (2019). Movement patterns and feeding ecology of the reef
manta ray (Mobula alfredi) in Seychelles. PhD thesis, University of
Western Australia, School of Biological Sciences.
Peel, L. R., Daly, R., Keating Daly, C. A., Stevens, G. M., Collin, S. P., &
Meekan, M. G. (2019). Stable isotope analyses reveal unique trophic
role of reef manta rays (Mobula alfredi) at a remote coral reef. Royal
Society Open Science,6(9), 190599.
Peel, L. R., Stevens, G. M., Daly, R., Daly, C. A. K., Lea, J. S., Clarke, C. R.,
Meekan, M. G. (2019). Movement and residency patterns of reef
manta rays Mobula alfredi in the Amirante Islands, Seychelles. Marine
Ecology Progress Series,621, 169184.
Peel, L. R., Stevens, G. M., Daly, R., Keating Daly, C. A., Collin, S. P.,
Nogués, J., & Meekan, M. G. (2020). Regional movements of reef
manta rays (Mobula alfredi) in Seychelles waters. Frontiers in Marine Sci-
ence,558, 1-17.
Pelegrín, I. D., Luís, L. D., Nerantzoulis, I. C., & Uetimane, A. E. (2015). Par-
que Nacional do Arquipélago de Bazaruto (PNAB). Plano de Maneio
20162025 de uma
´
Area de Conservaç˜ao Marinha, Província de
Inhambane, Moçambique. EIA &, Lda. Administraç˜ao Nacio nal das
´
Areas de Conservaç˜ao, Ministério da Terra, Ambiente e Desenvolvi-
mento Rural, Maputo.
Perryman, R. J., Carpenter, M., Lie, E., Sofronov, G., Marshall, A. D., &
Brown, C. (2021). Reef manta ray cephalic lobe movements are modu-
lated during social interactions. Behavioral Ecology and Sociobiology,
75(3), 115.
Perryman, R. J., Venables, S. K., Tapilatu, R. F., Marshall, A. D.,
Brown, C., & Franks, D. W. (2019). Social preferences and network
CARPENTER ET AL.11
FISH
structure in a population of reef manta rays. Behavioral Ecology and
Sociobiology,73(8), 118.
Pollock, K. H. (1981). Capture-recapture models: A review of current
methods, assumptions, and experimental design. In C. J. Ralph & J. M.
Scott (Eds.), Estimating numbers of terrestrial birds. Studies in avian biol-
ogy (pp. 426435). Lawrence, Kansas: Allen Press Inc,
Pollock, K. H., Nichols, J. D., Brownie, C., & Hines, J. E. (1990). Statistical
inference for capture-recapture experiments. Wildlife Monographs,
204(1), 397.
Quartly, G. D., & Srokosz, M. A. (2004). Eddies in the southern
Mozambique Channel. Deep Sea Research Part II: Topical Studies in
Oceanography,51(13), 6983.
R Core Team. (2021). R: A language and environment for statistical computing.
Vienna, Austria: R Foundation for Statistical Computing VERSION 4.1.2.
Rohner, C. A., Pierce, S. J., Marshall, A. D., Weeks, S. J., Bennett, M. B., &
Richardson, A. J. (2013). Trends in sightings and environmental influ-
ences on a coastal aggregation of manta rays and whale sharks. Marine
Ecology Progress Series,482, 153168.
Rubin, R. D. (2002). Manta rays: Not all black and white. Shark Focus,15,45.
Schofield, G., Katselidis, K. A., Dimopoulos, P., & Pantis, J. D. (2008). Inves-
tigating the viability of photo- identification as a tool to study endan-
gered sea turtle populations. Journal of Experimental Marine Biology
and Ecology,360(2), 103108.
Setyawan, E., Sianipar, A. B., Erdmann, M. V., Fischer, A. M., Haddy, J. A.,
Beale, C. S., Mambrasar, R. (2018). Site fidelity and movement pat-
terns of reef manta rays (Mobula alfredi): Mobulidae using passive
acoustic telemetry in northern Raja Ampat, Indonesia. Nature Conser-
vation Research,3(4), 115.
Silva, M. A., Magalh˜aes, S., Prieto, R., Santos, R. S., & Hammond, P. S.
(2009). Estimating survival and abundance in a bottlenose dolphin
population taking into account transience and temporary emigration.
Marine Ecology Progress Series,392, 263276.
Smith, H. C., Pollock, K., Waples, K., Bradley, S., & Bejder, L. (2013). Use of
the robust design to estimate seasonal abundance and demographic
parameters of a coastal bottlenose dolphin (Tursiops aduncus) popula-
tion. PLoS One,8(10), e76574.
Stevens, G. M., Hawkins, J. P., & Roberts, C. M. (2018). Courtship and mat-
ing behaviour of manta rays Mobula alfredi and M. birostris in the
Maldives. Journal of Fish Biology,93(2), 344359.
Stevens, G. M. W. (2016). Conservation and population ecology of manta
rays in the Maldives. PhD thesis, University of York, School of
Environment.
Stewart, J. D., Nuttall, M., Hickerson, E. L., & Johnston, M. A. (2018). Important
juvenile manta ray habitat at flower garden banks National Marine Sanctu-
ary in the northwestern Gulf of Mexico. Marine Biology,165(7), 18.
Temple, A. J., Kiszka, J. J., Stead, S. M., Wambiji, N., Brito, A.,
Poonian, C. N. S., Berggren, P. (2018). Marine megafauna interac-
tions with small-scale fisheries in the southwestern Indian Ocean: A
review of status and challenges for research and management. Reviews
in Fish Biology and Fisheries,28(1), 89115.
Thorburn, J., Neat, F., Burrett, I., Henry, L. A., Bailey, D. M., Jones, C. S., &
Noble, L. R. (2019). Ontogenetic variation in movements and depth
use, and evidence of partial migration in a benthopelagic elasmo-
branch. Frontiers in Ecology and Evolution,7, 353.
Tibiriçá, Y., Birtles, A., Valentine, P., & Miller, D. K. (2011). Diving tourism
in Mozambique: An opportunity at risk? Tourism in Marine Environ-
ments,7(34), 141151.
Towner, A. V., Wcisel, M. A., Reisinger, R. R., Edwards, D., & Jewell, O. J.
(2013). Gauging the threat: The first population estimate white sharks
in South Africa using photo identification and automated software.
PLoS One,8(6), e66035.
Venables, S., Winstanley, G., Bowles, L., & Marshall, A. D. (2016). A giant
opportunity: The economic impact of manta rays on the Mozambican
tourism industryAn incentive for increased management and protec-
tion. Tourism in Marine Environments,12(1), 5168.
Venables, S. K. (2020). Ecology and conservation of a threatened reef
manta ray (Mobula alfredi) population in southern Mozambique. PhD
thesis, University of Western Australia, School of Biological Sciences.
Venables, S. K., Marshall, A. D., Armstrong, A. J., Tomkins, J. L., &
Kennington, W. J. (2021). Genome-wide SNPs detect no evidence of
genetic population structure for reef manta rays (Mobula alfredi)in
southern Mozambique. Heredity,126(2), 308319.
Venables, S. K., van Duinkerken, D. I., Rohner, C. A., & Marshall, A. D.
(2020). Habitat use and movement patterns of reef manta rays, Mobula
alfredi in southern Mozambique. Marine Ecology Progress Series,634,
99114.
White, G. C., & Burnham, K. P. (1999). Program MARK: Survival estimation
from populations of marked animals. Bird Study,46, S120S139.
Whitehead, H. (2001). Analysis of animal movement using opportunistic
individual identifications: Application to sperm whales. Ecology,82(5),
14171432.
Whitehead, H. (2007). Selection of models of lagged identification rates
and lagged association rates using AIC and QAIC. Communications in
Statistics: Simulation and Computation,36(6), 12331246.
Whitehead, H. (2009). SOCPROG programs: Analysing animal social struc-
tures. Behavioral Ecology and Sociobiology,63(5), 765778.
Whitney, N. M., & Crow, G. L. (2007). Reproductive biology of the tiger
shark (Galeocerdo cuvier) in Hawaii. Marine Biology,151(1), 6370.
Wiirsig, B., & Jefferson, T. A. (1990). Methods of photoidentification for
small cetaceans. Representative for the International Whaling Commis-
sion,2,4352.
Williams, B. K., Nichols, J. D., & Conroy, M. J. (2002). Analysis and manage-
ment of animal populations. San Diego, California: Academic Press.
SUPPORTING INFORMATION
Additional supporting information can be found online in the Support-
ing Information section at the end of this article.
How to cite this article: Carpenter, M., Cullain, N., Venables,
S. K., Tibiriçá, Y., Griffiths, C., & Marshall, A. D. (2022).
Evidence of Závora Bay as a critical site for reef manta rays,
Mobula alfredi, in southern Mozambique. Journal of Fish
Biology,112. https://doi.org/10.1111/jfb.15132
12 CARPENTER ET AL.
FISH
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Persistent threats from fishing pressure and increasing habitat degradation, as well as slow recovery rates resulting from K-selected life histories, are putting increasing numbers of elasmobranch species (sharks, rays and skates) at risk of extinction worldwide. Global declines in elasmobranch populations merit evidence-based conservation measures, but establishment of such measures has been hindered by a paucity of population-level data. Mozambique supports a high number of threatened (Vulnerable, Endangered or Critically Endangered) and Data Deficient elasmobranch species; however, most of these populations are poorly understood. We collected baseline data on the presence, species richness, diversity, and habitat use of elasmobranchs within Zavora Bay, Mozambique, using baited remote underwater video and underwater visual census surveys, at six sites, from January 2017 to March 2020. Our results show that Zavora Bay is used year-round by 20 elasmobranch species, all of which are listed on the IUCN Red List of Threatened Species as either Data Deficient, threatened, or Near Threatened with declining population trends. Many of these species are heavily targeted by fisheries or caught as bycatch throughout their range, and some do not encounter protected areas anywhere within their natural range, highlighting the urgency to establish adequate conservation strategies to prevent further population declines.
Article
Full-text available
Background Expanding fisheries in developing nations like Sri Lanka have a significant impact on threatened marine species such as elasmobranchs. Manta and devil (mobulid) rays have some of the most conservative life history strategies of any elasmobranch, and even low to moderate levels of bycatch from gillnet fisheries may lead to significant population declines. A lack of information on life history, demographics, population trends, and fisheries impacts hinders effective management measures for these species. Method We report on mobulid fishery landings over nine years between 2011 and 2020 across 38 landing sites in Sri Lanka. We collected data on catch numbers, body sizes, sex, and maturity status for five mobulid species. We used a Bayesian state-space model to estimate monthly country-wide catch rates and total annual landings of mobulid rays. We used catch curve analyses to estimate total mortality for Mobula mobular , and evaluated trends in recorded body sizes across the study period for M. mobular, M. birostris, M. tarapacana and M. thurstoni . Results We find that catch rates have declined an order of magnitude for all species across the study period, and that total annual captures of mobulid rays by the Sri Lankan artisanal fishing fleet exceed the estimated annual captures of mobulids in all global, industrial purse seine fisheries combined. Catch curve analyses suggest that M. mobular is being fished at rates far above the species’ intrinsic population growth rate, and the average sizes of all mobulids in the fishery except for M. birostris are declining. Collectively, these findings suggest overfishing of mobulid ray populations in the northern Indian Ocean by Sri Lankan artisanal fisheries. We recommend strengthening the management of these species through improved implementation of CITES, CMS, and regional fisheries management actions. In addition, we report on the demographic characteristics of mobulids landed in Sri Lanka and provide the first record of M. eregoodoo in the country.
Article
Full-text available
A detailed understanding of the dynamics of small-scale (10s km) habitat use by the reef manta ray (Mobula alfredi) in the Maldives Archipelago is required to develop an effective national conservation management plan for this wide-ranging species. Here, a combination of photo-ID sightings data and acoustic telemetry were used to investigate both long-term M. alfredi visitation trends and small-scale movement patterns to key habitats on the eastern side of Baa Atoll (Hanifaru Bay feeding area, Dhigu Thila multifunctional site, and Nelivaru Thila cleaning station). All tagged and most of the sighted M. alfredi exhibited high affinity to the eastern side of Baa Atoll, where 99% of detections occurred, and 69% of individuals were re-sighted in multiple years. Sightings data suggests that visitation patterns may be associated with differences in habitat use by sex and maturity status. Boosted regression trees indicated that tag detection probability at Hanifaru Bay increased with increased westerly wind speed (>5ms-1) during the day, close to a new and full moon just after high tide, and when the tidal range was low. Interaction effects between predictors suggest that wind-driven oceanographic processes, such as Langmuir Circulation, maybe working to increase zooplankton concentration at this location. Tag detection probability increased at Dhigu Thila under similar conditions. At Nelivaru Thila, it increased at lower wind speeds (
Article
Full-text available
Globally, reef manta rays (Mobula alfredi) are in decline and are particularly vulnerable to exploitation and disturbance at aggregation sites. Here, passive acoustic telemetry and a suite of advanced oceanographic technologies were used for the first time to investigate the fine-scale (5-min) influence of oceanographic drivers on the visitation patterns of 19 tagged M. alfredi to a feeding aggregation site at Egmont Atoll in the Chagos Archipelago. Boosted regression trees indicate that tag detection probability increased with the intrusion of cold-water bores propagating up the atoll slope through the narrow lagoon inlet during flood tide, potentially transporting zooplankton from the thermocline. Tag detection probability also increased with warmer near-surface temperature close to low tide, with near-surface currents flowing offshore, and with high levels of backscatter (a proxy of zooplankton biomass). These combinations of processes support the proposition that zooplankton carried from the thermocline into the lagoon during the flood may be pumped back out through the narrow inlet during an ebb tide. These conditions provide temporally limited feeding opportunities for M. alfredi, which are tied on the tides. Results also provide some evidence of the presence of Langmuir Circulation, which transports and concentrates zooplankton, and may partly explain why M. alfredi occasionally remained at the feeding location for longer than that two hours. Identification of these correlations provides unique insight into the dynamic synthesis of fine-scale oceanographic processes which are likely to influence the foraging ecology of M. alfredi at Egmont Atoll, and elsewhere throughout their range.
Article
Full-text available
Reef manta rays (Mobula alfredi) are social elasmobranchs that have structured societies and actively interact with preferred social partners. Being able to detect cues and signals produced by conspecifics is vital in enabling social behavior. Many elasmobranch species communicate via body and fin postures, but it is not yet known if or how mobulid rays communicate or respond to cues produced by conspecifics. These rays have specialized cephalic lobes that are highly flexible and used in feeding, but may have other functions such as gestural communication. In this study, we developed a standardized method to assess manta ray behavior in the wild via focal sampling and frame-by-frame video analysis. From observations recorded at cleaning stations in Raja Ampat, West Papua, we described various types of cephalic lobe positioning and movements made by free-ranging M. alfredi and investigated these in different behavioral contexts. We found that cephalic lobe curls were modulated when approaching both conspecifics and human divers, as well as during interactions with cleaner fish, suggesting that these lobes may be used in sensing water movements, olfactory sensing, and/or gestural communication. Cephalic lobes were moved independently of one another, but we found no evidence of individual laterality. The lack of chemosensory capacity on the lobes suggests that gestural communication is the most likely function, but further research is required to determine this. These results are informative in understanding the function of gestural communication in manta ray social interactions and add to our growing understanding of elasmobranchs’ sophisticated social behavior. Significance statement Social animals rely on communication and cues produced by conspecifics to respond to opportunities and threats. Many elasmobranch species use body and fin postures as gestural signals during competitive, courtship, and other social interactions. Reef manta rays (Mobula alfredi) are social elasmobranchs that form groups in near-shore, shallow water habitats. They possess paired, flexible cephalic lobes positioned on the side of the head which are specialized for feeding but may have a variety of functions. Here, we describe, quantify, and suggest likely explanations for the movement and positioning of cephalic lobes during various manta ray social behaviors, including conspecific, human, and cleaner fish interactions. Our results suggest that lobe movements may be used in detecting water movements, olfactory sensing, and/or gestural communication, and are an important aspect of social behavior in this species.
Article
Full-text available
Reef manta rays (Mobula alfredi) are capable of long-distance dispersal when habitat is continuous. In the Ningaloo Reef World Heritage Area located on Australia’s mid-west coast, M. alfredi is sighted year-round and is a focal species for ecotourism in the region. Despite value to local tourism and demographic vulnerability to exploitation, basic information regarding M. alfredi movements and area use in this region is lacking. Here we used satellite tags and a long-term photographic-identification database collected by researchers and citizen scientists to examine the species’ visitation patterns and regional space use. Tagged M. alfredi remained in coastal shelf waters, preferentially occupying shallow depths < 20 m (76% of depth observations). Comparison of real tracks against those produced by correlated random walk simulations revealed directional dispersal and connectivity between the Ningaloo and Shark Bay World Heritage Areas. The greatest linear distance an individual moved after tagging was ∼700 km. This movement range is comparable to seasonal M. alfredi migration observed at similar latitudes on Australia’s east coast. Photographic-identification of 1121 individual M. alfredi, from 5146 sighting events, revealed that some individuals exhibited long-term site affinity to locations within Ningaloo Marine Park, with the longest spanning 15.16 years (9.8% individuals > 10 years). Lagged Identification Rate analysis suggested site visitation was best described by a pattern of emigration and reimmigration. While our observations of movement and residency establishes basic information previously unknown for the species in this region, the overlap of core area use with existing UNESCO World Heritage Areas demonstrates the potential for well-positioned marine parks to provide protection of critical habitat for M. alfredi.
Article
Full-text available
The decline in numbers of reef manta rays (Mobula alfredi) throughout their range has highlighted the need for improved information on their spatial ecology in order to design effective conservation strategies for vulnerable populations. To understand their patterns of movement in Seychelles, we used three techniques—archival pop-up satellite tags, acoustic tags, and photo-identification—and focussed on the aggregation at D’Arros Island and St. Joseph Atoll within the Amirantes Group. M. alfredi were photographed within six of the seven Island Groups of Seychelles, with 64% of individuals being resighted at least once between July 2006 and December 2019 over timeframes of 1–3,462 days (9.5 years; median = 1,018 days). Only three individuals from D’Arros Island were resighted at a second aggregation site located more than 200 km away at St. François Atoll during photo-identification surveys. Satellite-tracked M. alfredi (n = 5 tracks; maximum 180 days) remained within the boundary of the Seychelles Exclusive Economic Zone, where they spent the majority of their time (87%) in the upper 50 m of the water column in close proximity to the Amirantes Bank. The inclusion of acoustic tagging data in the models of estimated satellite-track paths significantly reduced the errors associated with the geolocation positions derived from archived light level data. The insights gained into the patterns of horizontal and vertical movements of M. alfredi using this multi-technique approach highlight the significance of D’Arros Island and St. Joseph Atoll, and the wider Amirantes Group, to M. alfredi in Seychelles, and will benefit future conservation efforts for this species within Seychelles and the broader Western Indian Ocean.
Article
Little is known about the extent of genetic connectivity along continuous coastlines in manta rays, or whether site visitation is influenced by relatedness. Such information is pertinent to defining population boundaries and understanding localized dispersal patterns and behaviour. Here, we use 3057 genome-wide single-nucleotide polymorphisms (SNPs) to evaluate population genetic structure and assess the levels of relatedness at aggregation sites of reef manta rays (Mobula alfredi) in southern Mozambique (n = 114). Contrary to indications of limited dispersal along the southern Mozambican coastline inferred from photo-identification and telemetry studies, our results show no evidence of population structure (non-significant FST < 0.001) for M. alfredi along this coast. We also found no evidence that individuals sampled at the same site were more related than expected by chance for males, females or across both sexes, suggesting that kinship may not influence visitation patterns at these sites. We estimated the effective population size (Ne) of this population to be 375 (95% CI = 369–380). Comparison to a distant eastern Indian Ocean site (Western Australia, n = 15) revealed strong genetic differentiation between Mozambique and Western Australia (FST = 0.377), identifying the Indian Ocean basin as a barrier to dispersal. Our findings show that genetic connectivity in M. alfredi extends for several hundred kilometres along continuous coastlines. We therefore recommend that the population in Mozambique be considered a discrete management unit, and future conservation plans should prioritize integrated strategies along the entire southern coastline.