ArticlePDF Available

Abstract and Figures

Despite the distribution of dugongs Dugong dugon ranging across nearshore waters of the tropical and subtropical regions of the Indian Ocean and western Pacific Ocean, their distribution in the western Indian Ocean is highly fragmented and appears to be declining. The population of the Bazaruto Archipelago is believed to comprise the only viable population in the region. In all, 27 surveys were flown over the Bazaruto Bay area to define the distribution and estimate the abundance of the species in the area. A total of 9 052 nautical miles of survey effort was flown during the surveys, from which there were 355 sightings of 760 dugongs. Two core areas of distribution were apparent within the surveyed area; a northern core area spread within the 10 m isobath between the Save River mouth and Ponta Bartolomeu Dias (21°24′S), and a southern core area aligned with the shallow sandbanks to the north and south of Santa Carolina Island. Group sizes recorded in the Bazaruto Archipelago were comparable to group sizes recorded in other regions where dugongs occur, although few large (>20) groups of dugongs were seen in this study. Line transect analyses of each survey showed dugong densities were considerably lower than densities recorded in surveys in Australian waters or in the Arabian Gulf, with a population estimate of 247 dugongs (CV = 34.1) when all surveys were considered, and 359 dugongs (CV = 38.2) when only the surveys that were carried out under adequate sighting conditions were included.
Content may be subject to copyright.
African Journal of Marine Science 2011, 33(3): xxx–xxx
Printed in South Africa — All rights reserved
Copyright © NISC (Pty) Ltd
ISSN 1814–232X EISSN 1814–2338
doi: 10.2989/1814232X.2011.637347
African Journal of Marine Science is co-published by NISC (Pty) Ltd and Taylor & Francis
Dugong abundance and distribution in the Bazaruto Archipelago, Mozambique
KP Findlay1*, VG Cockcroft2 and AT Guissamulo3
1 Department of Oceanography, University of Cape Town, Private Bag, Rondebosch 7701, South Africa
2 Centre for Dolphin Studies, Department of Zoology, PO Box 77000, Nelson Mandela Metropolitan University, Port Elizabeth
6031, South Africa
3 Universidade Eduardo Mondlane, Museu de História Natural, PO Box 257, Maputo, Mozambique
* Corresponding author, e-mail:
Manuscript received February 2011; accepted May 2011
Despite the distribution of dugongs Dugong dugon ranging across nearshore waters of the tropical
and subtropical regions of the Indian Ocean and western Pacific Ocean, their distribution in the
western Indian Ocean is highly fragmented and appears to be declining. The population of the Bazaruto
Archipelago is believed to comprise the only viable population in the region. In all, 27 surveys were
flown over the Bazaruto Bay area to define the distribution and estimate the abundance of the species
in the area. A total of 9 052 nautical miles of survey effort was flown during the surveys, from which
there were 355 sightings of 760 dugongs. Two core areas of distribution were apparent within the
surveyed area; a northern core area spread within the 10 m isobath between the Save River mouth
and Ponta Bartolomeu Dias (21°24 S), and a southern core area aligned with the shallow sandbanks to
the north and south of Santa Carolina Island. Group sizes recorded in the Bazaruto Archipelago were
comparable to group sizes recorded in other regions where dugongs occur although few large (>20)
groups of dugongs were seen in this study. Line transect analyses of each survey showed dugong
densities were considerably lower than densities recorded in surveys in Australian waters or in the
Arabian Gulf, and a population of 359 dugongs (CV = 38.2) was estimated from surveys that were
carried out under adequate sighting conditions.
Keywords: abundance, Bazaruto, distribution, Dugong dugon, Mozambique, western Indian Ocean
Dugongs Dugong dugon have an extensive range in
nearshore tropical and subtropical coastal and island waters
of the Indo-Pacific from southern Mozambique in the west to
between Vanuatu and Japan in the east (Marsh et al. 2002).
This geographical distribution extends across an estimated
140 000 km of coastline from approximately 27° N to 27° S,
and may be limited by water temperatures of less than
about 18 °C (Preen 1992, Marsh et al. 1994). The dugong is
currently listed as Vulnerable by IUCN, although quantitative
population estimates have been made in only three regions,
namely Australia, the eastern Red Sea, and the Arabian Gulf
(Bayliss and Freeland 1989, Preen 1989, Marsh and Saalfeld
1990, Marsh et al. 1994, Preen et al. 1997, Marsh and Lawler
2001, Marsh et al. 2002).
Marsh et al. (2002) noted that the dugong is declining or
extinct across a third of its range but that the status of the
species over a large proportion of its range is unknown,
particularly within the western Indian Ocean. Historically,
the geographical distribution of dugongs within the
western Indian Ocean extended from Somalia in the north,
through Kenya, Tanzania, Mozambique and farther east
off the Comoros Islands, the Seychelles, Madagascar and
Mauritius. Information from both qualitative and quantita-
tive surveys (Marsh et al. 2002) show that current dugong
distribution may be summarised as patchy across the
western Indian Ocean region, including Kenya, Tanzania,
Madagascar, the Seychelles, Mayotte and Mozambique
(Cockcroft 1993, 1995, Muir et al. 2003). Dugongs possibly
still occur in the Comoros (at Moheli Island) and off the
Somalia coast, but their current status is unknown. Marsh et
al. (2002) reported that dugongs have become extinct from
Mauritius and the Maldives and appear to have vagrant
status in the Seychelles, although more recently Hermans
and Pistorius (2008) confirmed that there are resident
animals at Aldabra atoll. According to Marsh et al. (2002),
extinction of the dugong in the western Indian Ocean region
is inevitable without immediate and effective conservation
measures, but although dugongs are protected across the
range of all the western Indian Ocean states, enforcement
is currently limited by both capacity and resources.
Findlay, Cockcroft and Guissamulo
Dugong distribution within Mozambique
Based on a survey along the Mozambican coastline in
1969, Hughes (1971) reported that dugongs were common
at Maputo Bay, Chidenguele, Inhambane Bay, Bazaruto
Bay, Mozambique Island and Pemba Bay (see Figure 1).
The author also speculated (based on the existence of
suitable habitat) that dugongs might occur in the Quirimbas
Archipelago, but their status in this area was unknown
(Hughes 1971, Smithers and Lobão Tello 1976). Smithers
and Lobão Tello (1976) observed dugongs in Maputo Bay
(8–10 animals), Inhambane Bay (2–4 animals), Ponta
Bartolomeu Dias (20 animals) and along the coast between
Bazaruto Bay and Save River where they were reported to
be common. Elsewhere, smaller groups were observed at
Angoche, Mozambique Island, Matimbane Bay and Pemba
where they were thought to be abundant until 1970. These
reports suggested a hiatus in the distribution of dugongs
between the Save River mouth and Angoche (Hughes
and Oxley-Oxland 1971), although the lack of survey effort
within this area (coupled with the generally high water
turbidity resulting from river discharge in this area, and the
consequent difficulty in visual surveys) should be noted.
The status and distribution of dugongs along the
Mozambique coast are believed to have altered signifi-
cantly since Hughes’s (1969) survey, with results of recent
aerial and vessel-based surveys indicating a considerable
decline in dugong abundance across Mozambique waters.
In the mid-1970s, dugong herd sizes of 8–10 individuals
were reported for Inhaca Island (Guissamulo and Cockcroft
1997). Guissamulo and Cockcroft (1997) further assessed
the distribution and relative abundance of dugongs in
Maputo Bay during 1992, when dugongs were sighted in the
eastern quarter of the bay, in the vicinity of Inhaca Island
(Guissamulo 1993). This area is now thought to support
only 2 or 3 individuals (Cockcroft and Young 1998). Two
individuals were sighted in Maputo Bay in 2007 (N Rabe,
in litt.) and four individuals were seen (ATG pers. obs.) in
2008. Two (possibly three) individuals were taken from this
isolated population by an artisanal fisher in 2010.
During a boat survey in Inhambane Bay in October
1994, dugongs were observed throughout the bay (ATG
unpublished data), although an aerial survey in 2001
recorded only a single dugong observed outside the bay,
during a low spring tide (Mackie 2001). No dugongs were
recorded during a survey of 2007 (AGT unpublished data).
The Bazaruto Archipelago area is reported to support
the largest dugong population along the East African coast
(Dutton 1994). An aerial census of this area (including
the Bazaruto National Park [Bazaruto, Santa Carolina,
Benguerra, Magaruque and Bangue islands]) conducted by
WWF in May 2001 found dugongs distributed throughout the
northern, central and south central parts of the Archipelago
between Bazaruto Island and the mainland (Marsh et al.
2002). Estimates based on strip-transect sightings in 1992
suggested a local population of 130 dugongs in Bazaruto
Bay (Guissamulo and Cockcroft 1997). However, when
maximum counts of 21 animals per aerial survey were
obtained during surveys in the 1990s, it was suggested that
the population was declining (Dutton 1998). It should be
noted that the past series of surveys carried out across the
Bazaruto Archipelago in the 1990s and early 2000s were
inconsistent both in terms of effort and methodology, so that
resulting estimates cannot provide population trend indices.
No dugongs were recorded in aerial surveys of the area
between Pemba and Mtwara in northern Mozambique
(Cabo Delgado Province) in 2007 (ATG unpublished data).
A local fisher reported seeing a lone dugong in 2001 near
Quilalea Island in the Quirimbas National Park (Motta
2001). Although no dugongs appear to have been reported
from this area since an individual drowned in a net in the
Quirimbas National Park in November 2003 (WWF 2004),
an incidental sighting of a lone individual was reported
from there to KPF in 2009. The intensification of large-
mesh gillnetting from 1976 onwards (sometimes directed
at dugongs), coupled with lack of law enforcement, is
thought to have been the principal cause of the perceived
decline of the dugong population in Mozambique. Such
fishing pressure is further compounded by seine-netting,
commercial trawl operations and palisade fish traps. WWF
(2004) suggested that habitat destruction of seagrass beds
(through increased levels of riverine sedimentation and from
natural cyclone and flood events), and increased anthro-
pogenic disturbance through exposure to vessel noise
(particularly tourism vessels), are further threats to dugong
populations in Mozambique.
This paper reports on a series of aerial surveys carried
out in 2006 and 2007 to define the current distribution and
abundance of dugongs within the Bazaruto Archipelago and
surrounding waters.
Material and methods
Study area
The Bazaruto Archipelago is a series of five islands (Bazaruto,
Benguerra, Magaruque, Bangue and Santa Carolina) situated
in a general north–south linear orientation in the vicinity of
21° S on the central coastline of Mozambique, although Santa
Carolina lies to the west of the others (Figure 1). Bazaruto
Bay is a shallow (generally <30 m deep) protected bay of
some 1 000 km2, lying between the four outer islands of the
Bazaruto Archipelago and the mainland. This large bay has
extensive seagrass beds and consequently provides suitable
habitat for dugongs. There are two distinct basins in this bay,
one to the north of Santa Carolina Island (maximum depth 33
m) and one in the centre section of the bay (maximum depth
24 m). The southern section of the bay comprises vast areas
of tidal flats, which often dry out during spring low tides.
The main feature of water circulation within the bay is the
strong tidal currents (mean spring tidal range is approxi-
mately 3 m during normal spring tides) during the flood and
ebb phases, and wave action is largely restricted to the
seaward side of the islands. Such strong tidal flows maintain
the deep channels on the landward side of the islands. The
physical and chemical characteristics of the water masses
of Bazaruto Bay exhibit spatial and temporal variability by
season as rainfall is highly variable both within and between
years. In the dry season, the bay has a marine character,
with a uniform salinity ranging from 35 to 36. However, in
the wet season the bay becomes more estuarine, exhibiting
a lower overall average salinity (33–35) compared with
the dry season. Water temperatures for the Bazaruto area
range annually between 24 and 28 °C (McClanahan et al.
African Journal of Marine Science 2011, 33(3): xxx–xxx 3
2000). The Bazaruto Archipelago lies within a high-risk
region for tropical cyclones.
As a shallow tropical bay, Bazaruto Bay contains a number
of important seagrass meadows or beds. Prior to this project,
the information on seagrass species composition, extent
and distribution was only known for the southern extent
of the bay. Within the Inhassoro and Cabo São Sebastião
areas, seagrasses cover an area of approximately 88 km2
of the shallow intertidal and subtidal waters inside the 5 m
isobath. Nine species of seagrass were recorded there (ATG
unpublished data), namely: Thalassondendron ciliatum,
Cymodocea rotundata, Cymodocea serrulata, Thalassia
hemprinchii, Halophila ovalis, Nanozostera capensis,
Halodule uninervis, Halodule wrightii and Siringodium isoeti-
folium. Seagrass meadows also occur north of Inhassoro and
off the Govuro River estuary and westward of the Bartolomeu
Dias area, where it is suspected that N. capensis, H.
uninvervis and C. rotundanta dominate the meadows.
Two major marine protected areas occur in the vicinity
of Bazaruto Archipelago, namely the Bazaruto Archipelago
National Park and the total protection zone of the Cabo de
São Sebastião.
Aerial sighting surveys
Field methods
Aerial surveys between March 2006 and October 2007
(Table 1) were flown over the Bazaruto Bay area from Cabo
São Sebastião in the south to north of the Save River mouth
in the north and between the coast and the 30 m isobath
(extending seawards of the Bazaruto Archipelago islands in
the south). Surveys were flown as a series of east to west
or west to east transects in a progressive north to south
40° E35° E
25° S
20° S
15° S
Cabo São Sebastião
Bazaruto Is.
Govuro River
Dias spit
Bazaruto Archipelago
Enlarged area
Benguerra Is.
Bangue Is.
Magaruque Is.
Pomene Estuary
Quirimbas Archipelago
Matimbane Bay
Inhaca Island
Save River
Mozambique Island
Inhambane Bay
Bazaruto Bay
Pemba Bay
Maputo Bay
Carolina Is.
Figure 1: (a) Coast of Mozambique and (b) Bazaruto Archipelago showing positions of locations referred to in the text
Findlay, Cockcroft and Guissamulo
direction. All transit time en route from Vilankulos airport
to the start of survey in the north, between transects and
en route from the end of survey to Vilankulos airport were
not regarded as effort. Sightings made while in transit were
recorded as secondary sightings, whereas sightings made
during effort transects were recorded as primary sightings.
Secondary sightings were excluded from abundance
estimation or density analyses. Transects were flown at an
altitude of 450 ft (137 m) and a speed of approximately 80
knots, in two types of aircraft; a Cessna 210 with six seats
and high-wing configuration (ZS SPV) was used in the flight
of 18 November 2006, whereas four different single-engine
Cessna 182 or 185 aircraft (ZS IGR, ZS EPO, D-EOVC and
C9 JSH) were used for the remainder of the surveys. The
number of transects flown per survey was not constant,
being influenced by fuel capacity, available observer time or
weather conditions. Because surveys were generally flown
from north to south, such limitations prevented the southern
end of the survey area from being covered on all occasions.
However, all survey transects that were flown over this
southern area of the bay (shaded area near the bottom of
Figure 2) were eventually excluded from the analyses (see
Results for explanation).
All survey transects were flown following the aircraft’s
Global Positioning System (GPS). Line spacing between
transects was 2 nautical miles. The lengths of the transects
were variable depending on the coastal orientation and
water depth; in the northern area, between Inhassoro and
the Save River, the eastern limit of the survey was limited
by the 20 m isobath, whereas within the Bazaruto National
Park, surveys were carried out between the islands and the
mainland, although surveys extended beyond the islands in
the southern area where shallow banks extended to the east
of the islands. The greater part of each survey was carried
out in passing mode. However, when large groups were
observed or suspected, or in cases in which identification
or group size estimation was not certain, the aircraft would
be diverted from the survey track line for confirmation of the
sighting. Other sightings of dugongs made during confirma-
tion were considered secondary sightings and recorded, but
not used for estimation of dugong numbers.
The survey crew consisted of a pilot, two observers and
a data recorder. A single observer searched the entire
visual observation area (outwards from the most vertical
perspective that could be gained) from the rear seats on
each side of the aircraft. The data recorder, seated in the
forward starboard position, logged flight paths, altitude,
survey effort, sighting parameters and weather conditions,
and also directed the pilot. Positions of effort and sightings
were recorded using a handheld GPS. On making a
Date Transects
(nautical miles)
Sightings of
Sightings of
Sightings of
calves Comment
31 March 2006 22 282.36 12 22 3
1 April 2006 22 282.36 11/2 16/2 2
2 April 2006 22 282.36 12/3 41/4 1
18 November 2006 17 225.07 8 18 4
16 December 2006 20 222.15 8/2 9/9 0
13 January 2007 25 333.85 15/3 61/4 5
14 January 2007 3 5.33 2 2 0 Survey aborted due to
poor weather conditions
28 January 2007 36 455.39 14/1 22/2 2
6 February 2007 36 424.14 31/1 53/1 4
15 February 2007 32 366.34 13/2 28/7 5
15 April 2007 26 337.75 21 30 2
26 April 2007 30 420.35 21 40 0
17 May 2007 31 432.33 28/1 52/1 5
9 June 2007 32 408.41 19/6 34/8 5
19 June 2007 32 405.01 12 20 2
26 June 2007 32 410.03 15/1 19/2 0
8 July 2007 34 420.73 12 42 1
17 July 2007 32 402.39 11 30 2
4 August 2007 32 407.98 3 3 0
20 August 2007 32 408.05 14/1 68/1 5
26 August 2007 0 Survey aborted due to
poor weather conditions
17 September 2007 26 129.95 2 3 0 Survey aborted due to
poor weather conditions
20 September 2007 30 385.65 8/1 24/1 3
6 October 2007 32 402.71 5 8 1
14 October 2007 32 398.33 8 33 3
29 October 2007 31 406.15 11 16 3 Twin-platform survey
29 October 2007 30 396.83 15 24 4 Twin-platform survey
Total 729 9 052.01 331/24 718/42 62
Table 1: Survey effort and sightings of dugongs made during aerial surveys carried out during the study
African Journal of Marine Science 2011, 33(3): xxx–xxx 5
(a) (b) (c)
Figure 2: Distribution of (a) aerial survey search effort (transects) and of sightings of dugongs for (b) all groups and (c) groups containing a calf observed during surveys of the Bazaruto
Archipelago. The dark shaded areas were excluded from analyses due to shallow depths precluding dugong distribution
Findlay, Cockcroft and Guissamulo
sighting, the observer measured the perpendicular distance
from the survey track line to the dugong sighting using an
inclinometer to measure the dip azimuth angle to the group
(surveys after 16 December 2006), or by assigning sightings
to distance bins marked by tape on the aircraft wing-strut
(surveys prior to and including 16 December 2006). All
information on a dugong sighting (including the number of
animals, occurrence of calves in groups, and the sighting
angles) was immediately relayed by the observer to the
data recorder, who operated the GPS and recorded sighting
data. Data and associated GPS positions were immedi-
ately downloaded on completion of each of the surveys, and
entered into spreadsheets for initial verification analyses.
Data analyses
Perpendicular distances of dugong groups from the track
line were calculated for sightings as d.tan(
), where d was
the aircraft altitude at the time of sighting and
was the
horizon minus the dip azimuth angle of the sighting from
the horizon (measured at the time of sighting abeam by
handheld inclinometer).
The Distance software programme (Thomas et al. 2006)
was used to fit a hazard-rate model (Buckland 1985), to
the perpendicular distances grouped into 0.05 nautical mile
intervals to give the sighting probability density function f(0)
and its variance V[f(0)]. All the perpendicular distances of
sightings, left-truncated at 0.1 nautical miles due to the low
detection probabilities directly under the aircraft, were used.
This pooled sighting probability density function was initially
applied to all surveys combined, to estimate abundance
over the entire series of surveys. Given that the application
of a generic effective search width will bias results of the
individual surveys (upwards under good sighting conditions
and downwards under poor sighting conditions), a further set
of analyses was performed on the data collected during the
five surveys carried out under optimal sighting conditions.
The entire area that was surveyed was calculated as the
extent of the area between the coastline and the eastern
limit of the survey transects, excluding the land area of
islands or the area shaded in Figure 2 (see Results for
Assumptions of the abundance analyses
Estimates of animal abundance across the survey area
require a number of assumptions to be met including:
The area surveyed represents the entire range of the
population. The current range limits of the Bazaruto
Archipelago dugong population are unknown. Anecdotal
reports of individuals within the Pomene Estuary and
published historic records of individuals in Inhambane
Bay suggest a possibly broader range than the area
surveyed. However, the association of these animals
with the Bazaruto Archipelago population is unknown
and it is assumed for the purpose of this study that such
groups are discrete and no migration occurs in and out
of the surveyed area. It is further assumed that the entire
Bazaruto Archipelago population is found within the area
surveyed. This assumption may bias the abundance
estimate downwards if animals are present outside of the
area surveyed (e.g. to the south of Cabo São Sebastião
or to the north of the Save River mouth).
The probability of detection of groups of animals on the
track line g(0) is assumed to be 1 (that all groups distrib-
uted on the track line will be detected with certainty) and
the probability of detection decreases with distance from
the survey track line.
A number of authors (Marsh and Sinclair 1989, Gales
et al. 2004, Preen 2004) have reported on detection bias
of animals that are missed by observers in strip transect
surveys. Such bias may result from (a) availability bias
where animals are not near the surface and consequently
not available to observers as potential sightings, and (b)
perception bias that results from visible animals being
missed by the observers. The assumption that all groups
on the track line will be detected with certainty is unlikely to
be met and will bias the abundance estimate downwards.
An estimate of the number of groups available as
potential sightings may be carried out by analysis of the
sightings from independent sighting platforms using a
capture–recapture Petersen model. A twin-platform survey
was carried out on 29 October 2007, when two aircraft
surveyed the area independently. The standard survey
aircraft (the primary aircraft) completed the survey followed
by an independent second survey aircraft flying the same
transects some 5–7 minutes behind the primary aircraft.
The available groups were determined using Chapman’s
modified Petersen model (Chapman 1951, in Seber 1982).
N = (n1 +1) (n2+1)/(m +1) – 1 (1)
where n1 is the number of sightings made from the primary
aircraft, n2 is the number of sightings made independently
from the second aircraft and m is the number of sightings
made by both aircraft. However, a number of assumptions
are required to be met for this model to be valid including:
(a) the population of groups is closed (that no immigra-
tion or emigration of groups, or splitting or merging
of groups occurs, between the two independent
(b) all groups are equally likely to be sighted in each
survey, and;
(c) all groups, and only groups sighted by both aircraft,
are recognised as such.
The entire surveyed area represents dugong habitat and
dugongs are randomly distributed throughout this habitat
with respect to survey transects. Two aspects of this
assumption require consideration:
uneven distribution of survey effort in relation to animal
distribution is likely to bias results if density estimates
from high or low density areas are spread across the
entire survey area; and secondly,
dugong habitat within the survey area is likely to be
fragmented by shallow sandbanks, with such
fragmentation varying by tidal cycle.
Aerial sighting surveys
In all, 27 surveys were flown during this study. Particulars
of each survey are presented in Table 1 and the distribu-
tion of total survey effort and sightings are shown in Figure
2. No survey was flown between 15 February and 15 April
African Journal of Marine Science 2011, 33(3): xxx–xxx 7
2007 due to the extensive damage received to the chartered
aircraft (ZS IGR) during Cyclone Favio on 22 February 2007,
and delays encountered in sourcing a replacement aircraft.
No surveys were attempted on 5 May, 25 May, 25 July, 26
August and 5 September of 2007 due to inclement weather,
and on 14 January, 26 August and 17 September, surveys
were abandoned after initiation due to high winds and sea
state and poor sighting conditions. A total of 9 052 nautical
miles of survey effort was flown during the 27 surveys. There
were 331 primary and 24 secondary sightings comprising
718 and 42 individuals respectively (Table 1).
Distribution, size and composition of dugong groups
The distribution of sighted dugong groups by depth interval is
shown in Figure 3. Although not adjusted for survey effort or
by tidal levels at the time of the survey, the results suggest
that dugongs occur out to the 20 m isobath. However, the
inshore or shallow-water distribution limit remains unknown
and is likely to vary with the tidal cycle. No dugongs were
sighted to the south of the line between Vilankulos and Cabo
São Sebastião (Figure 2). This area appears to comprise a
number of tidally inundated sandbanks and is possibly too
shallow to be utilised by dugongs. Consequently, it was
excluded from analyses of abundance. Within the surveyed
area, two core areas of distribution are apparent from
Figure 2. The northern core area of distribution was spread
across the inshore and offshore area to approximately the
10 m isobath between the Save River mouth and 21°24 S.
The southern core area, situated within Bazaruto Bay and
inshore of Bazaruto, Benguerra and Magaruque islands,
appeared to be aligned with the shallow sandbanks to the
north and south of Santa Carolina Island. The prominence
of these core areas was increased when the distribution of
individuals rather than groups was reviewed, in that such
core areas often contained large groups.
The mean group size (primary sightings made during full
survey effort) was 2.22 (SE = 0.191) (Figure 4). No seasonal
pattern in group size was evident (single-factor ANOVA; F =
0.4895, p = 0.6897, df = 354). A total of 62 (8.1%) of the 760
sighted dugongs in the Bazaruto Archipelago were calves
(Figure 2). The proportion of calves per survey ranged between
0 (December, April, June and August) and 22.2% (November)
(Table 1). Despite higher proportions of calves being observed
in spring (September–November), no significant difference
in the proportion of calves in the population were observed
between seasons (single-factor ANOVA; F = 2.11634, p =
0.128597, df = 24). As with the distribution of all groups, that of
calves appears to be centred on the two core areas.
Abundance estimation
Data from 23 survey flights were used for dugong
abundance estimation. The two aborted flights on 14
January 2007 and 17 September 2007 were excluded due to
incomplete survey coverage, and the survey carried out on
16 December 2006 was excluded due to uncertainty in the
recorded positions of the track lines (which were recorded to
the aircraft GPS and appeared to be in error). Furthermore,
only the results from the primary survey aircraft were used
from the 29 October 2007 twin-platform survey. All of the
remaining flights have been assumed to be of equal-sighting
probability with respect to environmental conditions. A total
of 8 824 nautical miles of survey effort were flown during
the remaining 23 survey flights, during which 321 primary
sightings were made (see Table 1). Nine sightings had no
associated distance data and were excluded from analyses.
The total survey area defined by the limits of each of the
surveys was calculated at 1 211 square nautical miles. This
area was defined by the coast in the west and the outer limits
of the transect lines in the east, with the shallow sandbank
area to the south-east of Vilankulos excluded (Figure 2).
Dugong distribution across the total surveyed area is shown
in Figure 2. In order to increase the available sample size,
the estimation of the sighting probability density function f(0)
was carried out using distance data pooled from all surveys.
The distribution of sightings by perpendicular distance from
the track line is shown in Figure 5. The inability of observers
to detect dugong groups below the aircraft is clearly apparent
in these data. Consequently, two analyses of sighting
probability density function were carried out: the first using
all sightings to define the sighting probability density function,
>205–10 11–200–5
All groups
Calf groups
Figure 3: Distribution of dugong groups (all groups and groups
containing a calf) recorded by depth interval during aerial surveys
of the Bazaruto Archipelago
Figure 4: Frequency of dugong group sizes recorded during aerial
surveys of the Bazaruto Archipelago
Findlay, Cockcroft and Guissamulo
and the second using a left truncation at 0.1 nautical miles so
that only sightings seen outside of a distance of 0.1 nautical
miles were used. Figure 6 shows the hazard rate modelled
f(0) to each of these options respectively. Given the paucity
of sightings below the aircraft, the truncated distance model
was selected as the most robust. Table 2 lists parameters
of the sighting probability density functions calculated using
the hazard rate model in the Distance analysis. No significant
correlation was found between sighting distance and group
size (r2 < 0.001, p = 0.93), which suggests detection was
independent of group size.
The density of dugongs recorded on the aerial surveys
across the Bazaruto Archipelago, using the effective strip
width calculated with the truncated distance data, was
0.058 per km2 for all surveys, and 0.123 per km2 when only
the five optimal surveys were considered. During the only
twin-platform survey that was carried out, the primary aircraft
(D-EOVC) made 11 primary sightings of dugongs, whereas
the secondary aircraft (ZS EPO) made 15 primary sightings
(Table 3). Six sightings were matched by eye as common to
both platforms on the basis of their positions and their time of
sighting. The capture–recapture Petersen model estimated a
total ‘population’ of 31 groups available for sighting within
the search width during this survey. Sightings of 11 groups
by the primary observation platform and 15 groups by the
secondary observation platform suggest that 35.5% and
48.4% of dugong groups were sighted by these platforms
respectively. The high proportions of missed sightings during
the twin-platform survey are presumed to have resulted from
the poor sighting conditions under which it was conducted.
This suggests that surveys conducted under suboptimal
conditions will underestimate abundance. The estimate
pooled across the five optimal sighting condition surveys of
359.0 (CV = 38.2) is therefore considered to be the most
robust estimate of dugong abundance from this study.
The aerial surveys identified that dugongs occur throughout
the shallow area inshore of the 20 m isobath from the Save
River mouth to Cabo São Sebastião of the Bazaruto Bay
area. Two distinct core areas of distributional abundance
were apparent from the sighting records, a northern
core area offshore of the Govuro River mouth and the
Bartolomeu Dias spit, and a core area in the vicinity of
Santa Carolina Island, where a number of large dugong
groups were recorded. The fact that the second core area
has little seagrass cover suggests that animals utilise this
area for reasons other than feeding. Dugongs use distinct
habitats for various activities, and shallow waters such
as tidal sandbanks and estuaries have been reported as
sites for calving (Hughes and Oxley-Oxland 1971, Marsh
et al. 1984). The use of these areas may be a strategy to
DISTANCE (nautical miles)
Figure 5: Distribution of all primary sightings of dugongs with
perpendicular distance from the track line observed during surveys
of the Bazaruto Archipelago
DISTANCE (nautical miles)
Figure 6: Detection probability function of dugong groups
sighted during aerial surveys of the Bazaruto Archipelago fitted
to (a) the frequencies of perpendicular distances of all primary
sightings in 0.05 nautical mile distance bins, (b) the frequencies of
perpendicular distances of all primary sightings in 0.05 nautical mile
distance bins outside of 0.01 nautical miles, and (c) the frequencies
of perpendicular distances of primary sightings made in the five
optimal sighting condition surveys in 0.05 nautical mile distance
bins outside of 0.01 nautical miles
African Journal of Marine Science 2011, 33(3): xxx–xxx 9
minimise the risk of shark predation to calves (Anderson
1981) and more time might be allocated to safe but lower-
quality feeding microhabitats when the likelihood of encoun-
tering sharks is increased (Wirsing et al. 2007).
The group sizes recorded in the Bazaruto Archipelago
were comparable to group sizes recorded in other regions
of dugong abundance, although few very large groups of
dugongs (of >20 individuals) were seen in the Bazaruto
Archipelago compared to other regions. For example, Preen
(2004) recorded a maximum group size of 674 dugongs in
the Arabian Gulf. The proportion of groups with calves was
also lower within the Bazaruto Bay region. Preen (2004)
reported proportions of dugong calves in the western and
southern Arabian Gulf in 1986 as 14.5%, compared with
18.7% in the southern Gulf in 1999, and noted that these
figures were typical of the proportions recorded on other
surveys (Marsh et al. 1994).
Of the two core areas of calf distribution, the high density
within Bazaruto Bay is more or less expected in terms of
water clarity, decreased exposure to predation and water
movement. However, the high density of calf sightings in the
northern core area is unexpected given the poor water clarity
often encountered in this area. The reason for the hiatus in
calf distribution between these two areas is unknown, but
may be related to water movement in and out of the bay,
the distribution of suitable calf habitat or to anthropogenic
influences of seine-net fishing in the region (and the impact
this may have on habitat). Surprisingly, two calf groups were
sighted offshore of the islands, where both wave action and
exposure to potential shark predation could be high.
The population numbers of dugongs in the Bazaruto
Archipelago estimated in this study is probably the
most robust population estimate of this population to
date. Abundance estimates from previous surveys have
suggested a smaller population than that recorded in
this study, although it should be noted that the different
estimates are not directly comparable on account of
different survey methods and survey area limits. Guissamulo
and Cockcroft (1997) estimated a local population of 130
dugongs in the bay based on strip transect sightings,
although their survey area was smaller than that covered
in our study and consequently a smaller proportion of the
local population was likely to be surveyed. Whereas Dutton
(1998) suggested that the population was declining because
no more than 21 animals were counted per survey during
aerial counts conducted during the 1990s, these count
data were not subjected to abundance analyses and were
most likely underestimates of abundance. Furthermore,
the extent of the survey area covered by Dutton (1998) is
unknown. An aerial census in May 2001 found dugongs
distributed throughout the northern, central and south
central areas of the Archipelago between Bazaruto Island
and the mainland and based on aerial counts (of between
25 and 130 individuals) between 1990 and 2002, WWF
(2004) suggested that this population was declining. In
1990, a single survey revealed a count of 92 dugongs in
the area between Bazaruto, Benguerra and Magaruque
islands and the mainland, from the northern tip of Bazaruto
to Vilankulos (VGC and ATG, unpublished data). Although
these data remain unanalysed and the extent of the survey
areas may not be directly comparable, the high raw counts
Length of
(nautical miles)
(square nautical
search width
group size
of groups
of animals
All surveys pooled
247 8 824.56 6.537 (0.447) 0.153 (0.010) 2.227 (0.22) 0.091 (0.009) 0.204 (0.281) 247 (34.1)
Five optimal sighting condition surveys
5 398.33 845.98 9.115 (2.578) 0.110 (0.31) 5 (2.05) 0.057 (0.029) 0.286 (0.186) 242 (65.6)
5 402.71 855.84 9.115 (2.578) 0.110 (0.31) 1.6 (0.24) 0.566 (0.021) 0.091 (0.491) 77 (54.18)
8 333.85 861.20 9.115 (2.578) 0.110 (0.31) 6.75 (3.45) 0.109 (0.049) 0.737 (0.503) 635 (68.29)
8 408.85 867.75 9.115 (2.578) 0.110 (0.31) 5.87 (4.17) 0.089 (0.039) 0.525 (0.438) 456 (83.40)
26 432.33 938.74 9.115 (2.578) 0.110 (0.31) 1.92 (0.24) 0.274 (0.098) 0.527 (0.199) 495 (37.78)
Mean 359 (38.20)
Table 2: Parameters used in the population estimates for the survey region with left truncation at a perpendicular distance of 0.1 nautical miles
Findlay, Cockcroft and Guissamulo
recorded on the 1990 survey suggest higher densities than
those recorded during the current surveys.
Little is known about the human-induced changes in
the habitat of dugongs in the Bazaruto Archipelago over
the past 40 years, although increased human pressure,
including increased vessel traffic and associated noise and
potential vessel strike effects, increased beach seine-net
fishing which may heavily alter seagrass structure through
disturbance, and increased pollution loading, suggest that
habitat quality may have declined over this period. The
greatest current impact on adult survivorship appears to
be the commercial fishery for shark fins in the Bazaruto
Archipelago. This fishery uses 40 cm stretch size gillnets
set for extended periods (unattended overnight) in known
dugong habitats and has resulted in many dugong mortali-
ties (Dutton 1994, Cockcroft et al. 1994, Guissamulo and
Cockcroft 1997). There is also evidence that the catch of
dugongs has developed into a directed fishery in Bazaruto
Bay (Guissamulo and Cockcroft 1997, Cockcroft and
Young 1998), where nets are set at night. Fishers do not
openly admit to taking dugongs; however, its meat is prized
(Cockcroft et al. 1994). Due to the illegal nature of the
activity, the extent of mortality to the Bazaruto Archipelago
dugong population from hunting remains unknown.
However, it is believed to be in the order of 4–6 individuals
per year (VGC pers. obs.). Other anthropogenic and natural
stressors to the population include vessel activity (including
potential strikes and vessel noise). Such stressors are of
particular concern due to both the inherently low reproduc-
tive rates of dugongs and the role played by stochastic or
episodic events such as the impacts of tropical cyclones or
floods on seagrass bed habitat (Heinsohn and Spain 1974).
The densities of sightings recorded in the Bazaruto
Archipelago area, although high compared with the rest of
East Africa, are low compared with populations in Australia
and the Arabian Gulf, where densities observed from aerial
strip surveys ranged from 0.21 km–2 (± 0.05) in 1986 in the
southern Arabian Gulf, 0.08 km–2 in the eastern Red Sea
(Preen 1989) to 0.71 km–2 in Shark Bay, Australia (Preen et
al. 1997). Given that the local Bazaruto Archipelago dugong
population might be viewed as the only viable population
within the western Indian Ocean metapopulation (which
has clearly declined over the past 30–40 years), it should
be afforded the highest possible conservation efforts.
Preen (2004) has suggested that the long-term survival of
the dugong in the western Indian Ocean will depend on the
establishment of an adequate network of protected areas
where the impacts of human activities can be minimised.
Furthermore, the Bazaruto Archipelago dugong population
must be viewed in the context of a seed or source population
for the western Indian Ocean metapopulation. Whereas daily
movements of dugongs are dependent on tidal amplitude,
dugongs may move considerable distances, as detected
during tracking studies using VHF and satellite telemetry
equipment in Australian waters (Marsh and Rathbun 1990,
Gales et al. 2004, Sheppard et al. 2006). Several studies
have shown that dugongs appear to move seasonally (or at
least during winter) in response to water temperature thresh-
olds of 17–19 °C (Anderson 1986, Preen 1992, 2004, Marsh
et al. 1994, Sheppard et al. 2006). Movements of dugongs
into or out of the Bazaruto area are largely unknown. North
of the Chiloane Islands, the water visibility is often extremely
poor, even in the ‘dry’ season when river discharge is
reduced. This poor visibility extends out to at least 15 nautical
miles offshore, as far as the aircraft used in our study was
permitted to fly. Although areas north of Sofala were not
explored, a preliminary examination of satellite photographs
suggests that there is little dugong habitat between the
Chiloane Islands and about 600 km to the north, which may
provide a natural barrier to northward dispersion by Bazaruto
dugongs. South of Cabo São Sebastião, the continental shelf
is narrow and probably does not provide adequate habitat
for dugongs, other than for transient movements. Inhambane,
which about 10 years ago was known to accommodate at
least 16 dugongs, now appears to be overexploited by
human activity. Dugongs were not seen during an exploratory
survey in 2007 and the density of fishing boats and fish traps
observed during this survey suggests that dugongs may well
have been all but extirpated from the area.
Despite the reported decline in dugongs in the Bazaruto
Bay area (Dutton 1994), the importance of this area to
dugongs is particularly evident in the context of the limited
dugong populations within the western Indian Ocean. The
Bazaruto population probably represents the most viable
(and possibly the only viable) population of dugongs in
the western Indian Ocean south of the Arabian Gulf. The
long-term conservation and recovery of this species within
the western Indian Ocean may well be dependent on the
‘seeding’ of the region by individuals from the Bazaruto
population (and their subsequent survival after ‘reseeding’
within areas that have shown recent population declines).
Consequently, it is recommended that the dugongs of
the Bazaruto Archipelago area be afforded the strongest
conservation priorities and that a dedicated and integrated
management plan for dugong conservation in the area is
implemented immediately. Such a management plan should
address dugong bycatch concerns as a matter of priority.
Acknowledgements — We are deeply indebted to the field team
of Aaron Banks, Nadia Sitoe and Santos Luis Mucave for their
dedication and effort in the field on this project, often under
difficult working conditions. The owners of Dugong Investments
are thanked for assistance with accommodation in Inhasorro. In
particular, Martin and Caron are thanked for their friendship and
outstanding hospitality at Dugong Lodge and for assistance with the
project. We thank Peter Farquar of Pelican Air/Bird Dog Aviation for
his contribution to aerial surveys prior to Cyclone Favio, and for his
continued interest in the project thereafter. Peter Ragg and Darren
Potgieter of Conservation Air Patrol are acknowledged for their
assistance in, and piloting of, the flown aerial surveys after Cyclone
Platform Model parameter Value
D-EOVC Capture (n1)11
ZS EPO Recapture (n2)15
D-EOVC and ZS EPO Common sightings (m)5
D-EOVC Proportion sighted 35.4%
ZS EPO Proportion sighted 48.4%
Table 3: Results of the independent twin-platform aerial survey
carried out on 29 October 2007 as per the modified Petersen model
(Equation 1)
African Journal of Marine Science 2011, 33(3): xxx–xxx 11
Favio. Sasol Petroleum International is gratefully acknowledged for
their support of this project at all levels.
Anderson PK. 1981. The behaviour of the dugong (Dugong dugon)
in relation to conservation and management. Bulletin of Marine
Science 31: 640–647.
Anderson PK. 1986. Dugongs of Shark Bay, Australia – seasonal
migration, water temperature, and forage. National Geographic
Research 2: 473–490.
Bayliss P, Freeland WJ. 1989. Seasonal distribution and abundance
of dugongs in the western Gulf of Carpentaria. Australian Wildlife
Research 16: 141–149.
Buckland ST. 1985. Perpendicular distance models for line transect
sampling. Biometrics 41: 177–195.
Cockcroft VG. 1993. A preliminary assessment of the status of
Madagascar’s coastal resources, with special reference to
marine mammals and reptiles. Final report for United Nations
Environment Programme, South African Department of Foreign
Affairs and African Wildlife. December 1993.
Cockcroft VG. 1995. Aerial survey in Kenya finds few dugongs.
Sirenews 24 [newsletter of the IUCN/SSC Sirenian Specialist
Cockcroft VG, Salm RV, Dutton TP. 1994. The status of dugongs
in the western Indian Ocean. In: First International Manatee and
Dugong Research Conference, 11–13 March 1994, Gainesville,
Cockcroft VC, Young DD. 1998. An investigation of the status of
coastal marine resources along the west coast of Madagascar.
Unpublished report, Worldwide Fund for Nature (WWF), Gland,
Dutton P. 1994. Past and present status of dugong Dugong dugon
in the Bazaruto Archipelago and other known habitats on the
Mozambique Coast. In: First International Manatee and Dugong
Research Conference, 11–13 March 1994, Gainesville, Florida.
Dutton P. 1998. East African dugongs disappearing. Sirenews 29
[newsletter of the IUCN/SSC Sirenia Specialist Group].
Gales NJ, McCauley RD, Lanyon J, Holley DK. 2004. Changes in
abundance of dugongs in Shark Bay, Ningaloo and Exmouth
Gulf, Western Australia: evidence for large-scale migration.
Wildlife Research 31: 283–290.
Guissamulo AT. 1993. Dolphins and dugongs distribution in Maputo
and Bazaruto Bays: an assessment of interaction with artisanal
and semi industrial trawling fisheries. Unpublished report to
UNEP, July 1993.
Guissamulo AT, Cockcroft VG. 1997. Dolphin and dugong
occurrence and distribution and fisheries interactions in Maputo
and Bazaruto Bays, Mozambique. Paper SC/49/SM24 presented
at the 49th meeting of the International Whaling Commission,
London, September 1998.
Heinsohn GE, Spain AV. 1974. Effects of a tropical cyclone on littoral
and sub-littoral biotic communities and on a population of dugongs
(Dugong dugon (Muller)). Biological Conservation 6: 143–152.
Hermans A, Pistorius PA. 2008. Marine mammal diversity in the
remote waters of Aldabra Atoll, Southern Seychelles. Atoll
Research Bulletin 564: 1–7.
Hughes GR. 1969. Dugong status survey in Mozambique. In: World
Wildlife Yearbook. Switzerland: World Wildlife Fund. pp 137–139.
Hughes G[R]. 1971. Preliminary report on the sea turtles and
dugongs of Mozambique. Veterinária Moçambicana 4: 45–62.
Hughes GR, Oxley-Oxland R. 1971. A survey of dugongs (Dugong
dugon) in and around António Enes, northern Moçambique.
Biological Conservation 3: 299–301.
Mackie CS. 1999. Aerial census of dugongs, dolphins and turtles in
Bazaruto National Park, Mozambique: May 1999. WWF Project
Mackie C[S]. 2001. Aerial census of dugongs, dolphins and turtles
in the Bazaruto National Park, WWF Eastern African Marine
Ecoregion Programme. Report to the Directorate of Forestry and
Wildlife, Ministry of Agriculture, Mozambique.
Marsh H, Corkeron P, Morissette N, Preen AR. 1995. The status of
the dugong in the southern Great Barrier Reef Marine Park. Great
Barrier Reef Marine Park Authority Research Publication 41.
Marsh H, Heinsohn GE, Marsh LM. 1984. Breeding cycle, life history
and population dynamics of the dugong, Dugong dugon (Sirenia:
Dugongidae). Australian Journal of Zoology 32: 767–788.
Marsh H, Lawler I. 2001. Dugong distribution and abundance in
the southern Great Barrier Reef Marine Park and Hervey Bay:
results of an aerial survey in October–December 1999. Research
Publication No. 70. Great Barrier Reef Marine Park Authority,
Marsh H, Penrose H, Eros C, Hugues J. 2002. Dugong status
reports and actions plans for countries and territories. Nairobi:
Marsh H, Prince RIT, Saalfeld WK, Shepherd R. 1994. The
distribution and abundance of the dugong in Shark Bay, Western
Australia. Wildlife Research 21: 149–161.
Marsh H, Rathbun GB. 1990. Development and application of
conventional and satellite radio tracking techniques for studying
dugong movements and habitat use. Australian Wildlife Research
17: 83–100.
Marsh HD, Saalfeld WK. 1990. The distribution and abundance of
dugongs in the Great Barrier Reef Marine Park south of Cape
Bedford. Australian Wildlife Research 17: 511–524.
Marsh H, Sinclair DF. 1989. Correcting for visibility bias in strip
transect aerial surveys of aquatic fauna. Journal of Wildlife
Management 53: 1017–1024.
McClanahan TR, Sheppard CRC, Obura DO (eds). 2000. Coral
reefs of the Indian Ocean: their ecology and conservation. USA:
Oxford University Press.
Motta H. 2001. Análise do FOFA (forças e fraquezas) de
Implementação do Parque Nacional das Quirimbas. Draft
Document. World Wildlife Fund.
Muir CE, Sallema A, Abdallah O, De Luca DW, Davenport
TRB. 2003. The dugong (Dugong dugon) in Tanzania: a
national assessment of status, distribution and threat. Wildlife
Conservation Society.
Preen A. 1989. Technical report, dugongs, Volume 1: the status and
conservation of dugongs in the Arabian Region, Saudi Arabia.
Meteorological and Environmental Protection Administration
[MEPA] Coastal and Marine Management Series, Saudi Arabia.
Preen A. 1992. Interactions between dugongs and seagrasses in a
subtropical environment. PhD thesis, James Cook University of
North Queensland, Townsville, Australia.
Preen AR. 2004. Distribution, abundance and conservation status
of dugongs and dolphins in the southern and western Arabian
Gulf. Biological Conservation 118: 205–218.
Preen AR, Marsh H, Lawler IR, Prince RIT, Shepherd R. 1997.
Distribution and abundance of dugongs, turtles, dolphins and
other megafauna in Shark Bay, Ningaloo Reef and Exmouth Gulf,
Western Australia. Wildlife Research 24: 185–209.
Seber GAF (ed.). 1982. The estimation of animal abundance. New
York: Macmillan Publishing.
Sheppard JK, Preen AR, Marsh H, Lawler I, Whiting SD, Jones
R. 2006. Movement heterogeneity of dugongs, Dugong dugon
(Müller), over large spatial scales. Journal of Experimental
Marine Biology and Ecology 334: 64–83.
Smithers RHN, Lobão Tello JLP. 1976. Checklist and atlas of the
mammals of Mozambique. Museum Memoir No. 8, Trustees of
the National Museums and Monuments of Rhodesia, Salisbury.
Thomas L, Laake JL, Strindberg S. Marques FFC, Buckland ST,
Borchers DL, Anderson DR, Burnham KP, Hedley SL, Pollard
JH, Bishop JRB, Marques TA. 2006. Distance 5.0. Release “x”1.
Findlay, Cockcroft and Guissamulo
Research Unit for Wildlife Population Assessment, University of
St. Andrews, UK. Available at
Wirsing AA, Heithaus MR, Dill LM. 2007. Living on the edge: dugongs
prefer to forage in microhabitats that allow escape from rather than
avoidance of predators. Animal Behaviour 74: 93–101.
WWF (World Wildlife Fund). 2004. Towards a Western Indian
Ocean dugong conservation strategy: the status of dugongs
in the Western Indian Ocean region and priority conservation
actions. WWF Eastern African Marine Ecoregion programme,
Dar es Salaam, Tanzania.
... The Bazaruto seascape in Mozambique, East Africa, which includes two marine protected areas (MPAs), is home to a dugong subpopulation that has survived despite these pressures (Figure 1; Cockcroft et al., 1994;Dutton, 1994;Findlay et al., 2011). These animals, henceforth referred to as the Bazaruto subpopulation, are the last known viable group of dugongs along the East African coast and are thus of significant biological, economic and cultural value (Findlay et al., 2011). ...
... The Bazaruto seascape in Mozambique, East Africa, which includes two marine protected areas (MPAs), is home to a dugong subpopulation that has survived despite these pressures (Figure 1; Cockcroft et al., 1994;Dutton, 1994;Findlay et al., 2011). These animals, henceforth referred to as the Bazaruto subpopulation, are the last known viable group of dugongs along the East African coast and are thus of significant biological, economic and cultural value (Findlay et al., 2011). Indeed, the East Africa dugong subpopulation, which includes the Bazaruto subpopulation, is geographically isolated from the nearby Red Sea subpopulation and is genetically distinct from subpopulations around western Indian Ocean islands, such as Madagascar (Marsh et al., 2011;Plön et al., 2019). ...
... Indeed, the East Africa dugong subpopulation, which includes the Bazaruto subpopulation, is geographically isolated from the nearby Red Sea subpopulation and is genetically distinct from subpopulations around western Indian Ocean islands, such as Madagascar (Marsh et al., 2011;Plön et al., 2019). Despite its enduring presence, the Bazaruto subpopulation remains threatened by the aforementioned pressures that were responsible for declines elsewhere in East Africa (Findlay et al., 2011). ...
Full-text available
Although the dugongs of Mozambique's Bazaruto Archipelago are the last known viable subpopulation along the East African coast, they remain threatened by a variety of anthropogenic and natural pressures that must be addressed to ensure their persistence. We aimed to establish recent trends in dugong abundance through a series of aerial surveys carried out between 2017 and 2021 over the Bazaruto seascape. We also assessed hotspots of dugong distribution in this region where targeted conservation measures may prove most effective. Finally, we modeled dugong population growth and mortality rates under varying scenarios to better understand the level of risk to anthropogenic mortality. We derived a total abundance of 325 SD 145 dugongs. While the estimates from this study and those from surveys in 2006–2007 (359 SD 137) suggest that dugong abundance has not changed significantly over the past 15 years, the confidence intervals of these estimates are too wide to detect potentially small changes relative to the subpopulation's size. The distribution of dugong sightings within the Bazaruto seascape over the 5‐year period indicates two core zones, one of which occurs outside the limits of established conservation areas. Population viability analyses demonstrated variability around rates of unnatural mortality that would cause long‐term decline in abundance. Together, these results provide strong motivations for higher level conservation actions such as the expansion and management of formal marine protected areas in the region and the listing of the East African dugong subpopulation within the IUCN Red List of Threatened Species. Our study also highlights the importance of developing alternative livelihoods and sustainably managing small‐scale fisheries in collaboration with local communities and other stakeholders in order to reduce the prevalence of fishing equipment and anthropogenic activities that directly or indirectly threaten dugongs in the Bazaruto seascape.
... In the remote Palau Archipelago of Micronesia, mean group size were 2.0 dugongs in 1978 (34 sightings, Brownell et al. 1981Brownell et al. ), 1.4 in 1983Rathbun et al. 1988;Marsh et al. 1995), and1.4 in 1991 (26 sightings;Marsh et al. 1995). In the Bazaruto Archipelago, Mozambique, Findlay et al. (2011) made 760 sightings during aerial surveys, with a mean group size of 2.2, a mode of one, and maxima of 21-50 observed. Surveys in New Caledonia, at the eastern fringe of the distribution of dugongs in the South Pacific, during 2003 resulted in mean group size of 1.5 dugongs with 72% of sightings solitary and 16% in pairs (mostly cows and calves) on transect surveys (total 125 dugongs; incidental sightings off survey tracks totaled an additional 99 sightings of 198 dugongs); the largest herd sighted had 45 dugongs (Garrigue et al. 2008). ...
Sirenian social and reproductive behaviors lack much complexity or diversity. Whereas sirenians are usually sighted as solitary, or as cows with single calves, aggregationsAggregation of many individuals can occur. Persistent social groupings are unknown. Home ranges are widely overlapping. Mating systems of dugongsDugong (Dugong dugon)(Dugong dugonDugong (Dugong dugon)) have been variously described as leks or as scramble promiscuityScramble promiscuity (mating herdsMatingherd) and lone mating pairs have been observed in areas of low density, but further research into the hypothesized leks is needed (especially because scramble promiscuityScramble promiscuity has been observed in the same region). DugongsDugong (Dugong dugon) and all manateesManatee (Trichechus) show scramble promiscuityScramble promiscuity, wherein malesMales form groups that escort single females with much physical contact for many days. The strongest social bonds are between females and nursingNursingcalves. Florida manateesFlorida manatee (Trichechus Manatus Latirostris) (Trichechus manatus latirostris) show natal philopatryNatal philopatry for years after weaningWeaning. Socially transmitted knowledgeSocially transmitted knowledge(traditionTradition) appears important to Florida manateesFlorida manatee (Trichechus Manatus Latirostris) and perhaps all species of sirenians, particularly in regions where seasonal movementsSeason/ seasonalmovements during winter are necessary for survival, such as in winter for Florida manateesFlorida manatee (Trichechus Manatus Latirostris), and dugongsDugong (Dugong dugon) at the high latitude limits of their range. Some populations of Antillean, Amazonian, and African manateesAfrican manatee (Trichechus senegalensis) have regular movements in response to seasonal flooding and access to foodFood, which also may be learned through traditionTradition. DugongsDugong (Dugong dugon) may rely on group movements based on traditional knowledge in response to regional lossLoss, of foodFood supply from extreme weather eventsExtreme weather events. Communication is most obvious through vocalizationsVocalizations, which can show individual distinctiveness. Vocal communicationVocalcommunication is most prevalent between mothers and calves. Allomaternal careAllomaternal Careoccurs in Florida manateesFlorida manatee (Trichechus Manatus Latirostris) at shared aggregationAggregation sites. Florida manateesFlorida manatee (Trichechus Manatus Latirostris) occupying a given region can consist of multiple matrilines that develop through the early bonding of calves to mothers and subsequent natal philopatryNatal philopatry. Population geneticsPopulationgenetics research supports maleMales-biased dispersalDispersal and possible female-based philopatry in other trichechids, but perhaps not as strongly in dugongids. Considerable further research is needed on these and related topics to more comprehensively understand sirenian social and reproductive behavior.
... Some Kenyan and Tanzanian coastal communities use dugong meat, bones, and oil for food, medicine, and ornamental purposes (Wamukoya et al., 1995;Marsh et al., 2002Marsh et al., , 2011Muir and Kiszka, 2012;Said et al., 2020). In Bazaruto Bay, Mozambique, which supports the most important dugong population in the region (Marsh et al., 2011), the illegal directed harvest of dugongs is estimated at 4-6 dugongs per year (Findlay et al., 2011). Bycaught animals are sold, albeit illegally, because they represent a windfall for impoverished fishers (Marsh et al., 2011). ...
Full-text available
Wild animals are captured or taken opportunistically, and the meat, body parts, and/or eggs are consumed for local subsistence or used for traditional purposes to some extent across most of the world, particularly in the tropics and subtropics. The consumption of aquatic animals is widespread, in some places has been sustained for millennia, and can be an important source of nutrition, income, and cultural identity to communities. Yet, economic opportunities to exploit wildlife at higher levels have led to unsustainable exploitation of some species. In the literature, there has been limited focus on the exploitation of aquatic non-fish animals for food and other purposes. Understanding the scope and potential threat of aquatic wild meat exploitation is an important first step toward appropriate inclusion on the international policy and conservation management agenda. Here, we conduct a review of the literature, and present an overview of the contemporary use of aquatic megafauna (cetaceans, sirenians, chelonians, and crocodylians) in the global tropics and subtropics, for species listed on the Appendices of the Convention on the Conservation of Migratory Species of Wild Animals (CMS). We find that consumption of aquatic megafauna is widespread in coastal regions, although to varying degrees, and that some species are likely to be at risk from overexploitation, particularly riverine megafauna. Finally, we provide recommendations for CMS in the context of the mandate of the Aquatic Wild Meat Working Group.
... Indeed, dugong grazing influences the biomass, species composition and nutritional quality of seagrass beds [16][17][18] , as well as their resilience 19 and dispersal 20 . Dugong regional distribution has mainly been documented through dedicated large scale aerial surveys in Mozambique, the Arabian Gulf, Australia, New Caledonia, and Malaysia 7,13,[21][22][23][24] or through fishermen questionnaires 25,26 . While the former is a lot more sophisticated and accurate both techniques are restricted in time (they represent snap-shots of daytime distribution) and analyzed at a coarse scale that limits the understanding of regional habitat use and spatio-temporal variations. ...
Full-text available
Mobile marine species display complex and nonstationary habitat use patterns that require understanding to design effective management measures. In this study, the spatio-temporal habitat use dynamics of the vulnerable dugong (Dugong dugon) were modelled from 16 satellite-tagged individuals in the coral reef lagoonal ecosystems of New Caledonia, South Pacific. Dugong residence time was calculated along the interpolated tracks (9371 hourly positions) to estimate intensity of use in three contrasting ecoregions, previously identified through hierarchical clustering of lagoon topographic characteristics. Across ecoregions, differences were identified in dugong spatial intensity of use of shallow waters, deeper lagoon waters and the fore-reef shelf outside the barrier reef. Maps of dugong intensity of use were predicted from these ecological relationships and validated with spatial density estimates derived from aerial surveys conducted for population assessment. While high correlation was found between the two datasets, our study extended the spatial patterns of dugong distribution obtained from aerial surveys across the diel cycle, especially in shallow waters preferentially used by dugongs at night/dusk during high tide. This study has important implications for dugong conservation and illustrates the potential benefits of satellite tracking and dynamic habitat use modelling to inform spatial management of elusive and mobile marine mammals.
... The IUCN Red List considers the dugong's status as Vulnerable on a global scale (Marsh and Sobtzick 2019) with an estimated population decline of 20% in the last century (Marsh et al. 2002). However, this classification is based on a global average which includes significant populations in Australia (c. 85,000 individuals) (CITES 2000) and the Arabian Gulf (5800 individuals) (Preen 2004).In the western Indian Ocean D. dugon would be more appropriately classified as Critically Endangered, based on the fact that: 1) the largest population in the region is considered to be in Mozambique (200-300 individuals, Findlay et al. 2011) with only small fragmented and declining populations scattered across the coastlines of east Africa, Madagascar, Comoros, and the Seychelles and 2) the recent mitochondrial DNA analyses reveal that the Malagasy and Comorian populations are genetically distinct from the east African population (Plon et al. 2019). ...
Full-text available
The authors provide the most comprehensive account to date of the occurrence, distribution and conservation status of dugongs in Madagascar based on historical (since 1902) and second half of 20th century observational and anecdotal data, in depth surveys of coastal communities conducted since 2009, a scientific field survey in 2009 up to the most recent reported mortality in 2018. The report highlights the highly precarious state and genetic isolation of the Madagascar dugong population and makes recommendations to rectify and enforce protective legislation, study the effects of industrial shrimp trawling, conduct awareness campaigns (especially in MPAs and LMMAs containing dugong & seagrass), expand monitoring efforts (especially through MPAs) and emphasize the importance of dugong in Malagasy culture and traditions.
... Habitats include sandy beaches, coastal dunes and coastal lakes, mangroves, seagrass beds, coral and biogenic reefs, deep sea and offshore pelagic (including deep-sea canyons), and five islands. The most iconic species is the dugong, arguably the only viable population in the WIO (Findlay et al., 2011), marine turtles, whales and dolphins, billfish, and the sand oyster (Pinctada spp.). Initially proclaimed in 1971 (Legislative Decree 46/71 of 25 May), it was extended to its current limits in 2001 (Decree 39/2001 of 25 May). ...
... 16% of the world's seagrass species)(Bandeira, 2002;Bandeira and Gell, 2003) . The seagrass meadows in the Bazaruto Archipelago in Mozambique support one of the remaining viable dugong populations in the WIO(Findlay et al ., 2011) . Protection of this valuable habitat is critical for the survival of the species . ...
... 16% of the world's seagrass species)(Bandeira, 2002;Bandeira and Gell, 2003) . The seagrass meadows in the Bazaruto Archipelago in Mozambique support one of the remaining viable dugong populations in the WIO(Findlay et al ., 2011) . Protection of this valuable habitat is critical for the survival of the species . ...
The seemingly unhurried nature of manatees and dugongs belies their great capacity for undertaking long-distance journeys, often repeatedly in the form of round-trip seasonal migrations, but sometimes as movements independent of seasonal influence. Unique attributes of sirenian biology that interact with features of their environment to mold patterns of movement and habitat use include herbivory, limited thermoregulatory physiology for coping with cold and, for manatees, an apparent need to ingest fresh water. Manatees and dugongs are remarkably adaptable in their large-scale movement behavior, as manifested by the considerable variation in the occurrence and extent of migrations across populations within species, and among individuals within populations. Some populations and individuals are relatively sedentary year-round, whereas others migrate hundreds of kilometers between seasonal ranges. The environmental selective pressures driving seasonal movements vary across species, climates, and ecosystems, but are most commonly generated by predictable fluctuations in water temperature (Florida manatee, some dugong populations), rainfall (coastal populations of Antillean and African manatees), or water level (inland populations of all 3 manatee species living in flood-pulse river systems) over the annual cycle. In each case there is a season (winter, dry, or low-water) of heightened environmental stress where the animals’ range is restricted to areas around a key limiting resource (warm water, fresh water, or deep water), and forage is therefore less available or of lower nutritional quality. Because dugongs are strictly marine and do not require fresh water, they experience fewer seasonally imposed constraints and are less likely to migrate than manatees. Consequently, the large-scale movements of dugongs seem more stochastic; assessing the status of forage over a wide area through occasional long-distance exploratory forays may represent a behavioral adaptation to periodic extensive declines in seagrass caused by extreme weather events. The available evidence for manatees indicates strong fidelity to seasonal or year-round ranges across years. A common finding from tracking studies is the existence of considerable variation in large-scale movement behavior among and within individuals, which should confer adaptability to environmental change in the short term and provide the raw material for evolutionary change over the long term.
Full-text available
Motivated by the need to estimate the abundance of marine mammal populations to inform conservation assessments, especially relating to fishery bycatch, this paper provides background on abundance estimation and reviews the various methods available for pinnipeds, cetaceans and sirenians. We first give an “entry-level” introduction to abundance estimation, including fundamental concepts and the importance of recognizing sources of bias and obtaining a measure of precision. Each of the primary methods available to estimate abundance of marine mammals is then described, including data collection and analysis, common challenges in implementation, and the assumptions made, violation of which can lead to bias. The main method for estimating pinniped abundance is extrapolation of counts of animals (pups or all-ages) on land or ice to the whole population. Cetacean and sirenian abundance is primarily estimated from transect surveys conducted from ships, small boats or aircraft. If individuals of a species can be recognized from natural markings, mark-recapture analysis of photo-identification data can be used to estimate the number of animals using the study area. Throughout, we cite example studies that illustrate the methods described. To estimate the abundance of a marine mammal population, key issues include: defining the population to be estimated, considering candidate methods based on strengths and weaknesses in relation to a range of logistical and practical issues, being aware of the resources required to collect and analyze the data, and understanding the assumptions made. We conclude with a discussion of some practical issues, given the various challenges that arise during implementation.
Full-text available
Study of the gonads of 108 dugongs from north Queensland indicates that neither females nor males are continuously in breeding condition. The gonadal activity of males in a population is not synchronized. Ovaries tend to be active in the latter half of the year. There is no evidence for females coming into oestrus soon after giving birth but conception can occur during lactation. In the Townsville-Cairns area, dugongs calve from August-September through December. Neonates are between 1.0 and 1.3 m long and weigh 20-35 kg. Dugongs of both sexes less than 2.2 m long are likely to be immature, those over 2.5 m long are probably mature, and the status of animals between 2.2 and 2.5 m long is uncertain. The pre-reproductive period seems to be very variable but is a minimum of 9-10 years for both sexes. The gestation period is about 1 year and lactation can last at least 1.5 years. The usual litter size is one. The secondary and tertiary sex ratios are 1:1. Estimates of the calving interval based on pregnancy rates, the rate of accumulation of placental scars, and calf counts from aerial surveys and photographs, range from 3 to 7 years for various populations. A simple population model has been used to calculate the relationship between calving interval and adult mortality rate for stationary populations with different pre-reproductive periods and juvenile mortality rates. Even the most optimistic schedule of reproduction and juvenile mortality demands an adult survivorship of about 90% per year for population maintenance.
Full-text available
In 1986 and 1987, Dugong dugon were counted from the air at an overall sampling intensity of 10.1% over 39 396 km2. The corrected population estimate (±SE) was 3479±459 dugongs at an overall density of 0.088±0.012 km-2, a precision of 13%. Highest densities were observed on inshore seagrass beds and in waters <5 m deep. Maps of density and distribution are given. -from Authors
Full-text available
Techniques were developed for tracking individual dugongs using buoyant, tethered, conventional and satellite radio transmitters, and applied to six dugongs caught off the North Queensland coast. The results support the Great Barrier Reef Marine Park Authority's policy of conserving dugongs by giving a high level of protection to some inshore seagrass beds that support large numbers of animals. The relative merits of conventional and satellite telemetry for tracking dugongs are discussed. -from Authors
Dungong dugon in this Western Australia bay concentrate near the E shore in summer and near the W shore in winter. Satellite imagery showing relative temperatures revealed close correspondence between these seasonal distributions and warmer water. On the winter range, correlation between local water temperature and dugong numbers implied that local movements were also influenced by temperature: dugongs abandoned feeding areas when temperatures dropped below 19oC. On the summer range dugongs foraged on beds of a tropical seagrass (Halodule sp.). In winter they were restricted to the temperate seagrass Amphibolis antarctica. The latter differs in growth form from seagrasses on which dugongs normally forage, and grazing appeared to be inefficient. Analyses for fiber, nutrients, and secondary compounds revealed significantly differences in nutritional quality. Data support the hypotheses that Shark Bay dugongs are forced to migrate from summer feeding grounds to winter refugia to minimize low temperature stress, and that migration to the summer range optimizes both diet and temprature. -from Author
We develop methodology for correcting for visibility bias by calculating and applying survey-specific correction factors in strip transect aerial surveys of aquatic fauna and incorporating their associated errors into the population estimate. The technique is applicable at all densities of the target species. Perception bias (the proportion of groups of the target species that are visible in the transect yet missed by observers) is corrected for using a modified Petersen estimate calculated for each of 2 teams of 2 observers with 1 team on either side of the aircraft. Within a team, each observer reports their uncolluded observations into a separate track of a 2-track tape recorder, so that after the survey, each group can be characterized as being seen by only 1 (specified) or both members of the team. A correction factor is also suggested to standardize for the proportion of animals that are unavailable to observers because of water turbidity.
Strip-transect aerial surveys of Shark Bay, Ningaloo Reef and Exmouth Gulf were conducted during the winters of 1989 and 1994. These surveys were designed primarily to estimate the abundance and distribution of dugongs, although they also allowed sea turtles and dolphins, and, to a lesser extent, whales, manta rays and whale sharks to be surveyed. Shark Bay contains a large population of dugongs that is of international significance. Estimates of approximately 10000 dugongs resulted from both surveys. The density of dugongs is the highest recorded in Australia and the Middle East, where these surveys have been conducted. Exmouth Gulf and Ningaloo Reef are also important dugong habitats, each supporting in the order of 1000 dugongs. The estimated number of turtles in Shark Bay is comparable to the number in Exmouth Gulf plus Ningaloo Reef (7000–9000). The density of turtles in Ningaloo Reef and, to a lesser extent, Exmouth Gulf is exceptionally high compared with most other areas that have been surveyed by the same technique. Shark Bay supports a substantial population of bottlenose dolphins (2000–3000 minimum estimate). Exmouth Gulf and Ningaloo Reef were not significant habitats for dolphins during the winter surveys. Substantial numbers of whales (primarily humpbacks) and manta rays occur in northern and western Shark Bay in winter. Ningaloo Reef is an important area for whale sharks and manta rays in autumn and winter. The Shark Bay Marine Park excludes much of the winter habitats of the large vertebrate fauna of Shark Bay. In 1989 and 1994, more than half of all the dugongs were seen outside the Marine Park (57·4 and 50·7%, respectively). Approximately one-third to one-half of turtles and dolphins were seen outside the Marine Park (in 1989 and 1994 respectively: turtles, 43 and 27%; dolphins, 47 and 32%). Almost all the whales and most of the manta rays were seen outside the Marine Park. Expansion of the Shark Bay Marine Park, to bring it into alignment with the marine section of the Shark Bay World Heritage Area, would facilitate the appropriate management of these populations. This would also simplify the State– Commonwealth collaboration necessary to meet the obligations of World Heritage listing. The coastal waters of Western Australia north of the surveyed area (over 6000 km of coastline) are relatively poorly known and surveys of their marine megafauna are required for wise planning and management.