Global

Vision International,
Seychelles - Curieuse Report No. 161-164

GVI Seychelles Curieuse

Island Conservation Expedition

Annual Report

2016
(January 2016 January 2017)




Submitted in whole to
Global Vision International
Seychelles National Parks Authority (SNPA)




Produced by
Peter Kowalski | Science Coordinator

And
Christophe Mason-Parker | Country Director
Alan Grant | Base Manager
Clare Atkinson| Science Officer
Amy Bradley |Science Officer
James McClelland | Science Officer
Bridgette Rademakers |Science Officer
Rebecca Hodgkiss | Science Coordinator

Special thanks
To all volunteers from January 2016 December 2016 for assisting with data collection.

GVI Seychelles Curieuse Island Conservation Expedition

Address: GVI c/o SNPA, PO Box 1240, Victoria, Mah, Seychelles

Email: seychelles@gviworld.com
Web page: http://www.gvi.co.uk and http://www.gviusa.com

2



Executive Summary
This report summarises the science programmes conducted by the Global Vision International (GVI)
Seychelles, Island Conservation Expedition on Curieuse Island, between January 2016 and January
2017.

The total rainfall for this time period was 1757.1mm (compared to the 2015 rainfall of 2748.3mm).

Bird monitoring has been put on hold as a priority in order to divert greater effort and manpower
into other current and upcoming projects. Until the end of 2015, 37 different species have been
recorded over 9,120 observations for coastal and mangrove sightings. Two years of data collection
has established a baseline understanding of seasonal visits by bird species to Curieuse. Over the
coming year, bird surveys will ideally be conducted on a quarterly basis by consecutive interns
expressing the desire to be trained in bird identification and monitoring methodologies. This data
will follow previously used methodologies in order to directly contribute to the existing data set.

The Coco de Mer growth survey followed on from the island-wide census completed in 2014. It has
now produced 33 months of growth and reproduction data, and some noteworthy trends are
becoming apparent. While populations on Curieuse and Praslin show many differences, similarities
have been observed in trends between leaf length and trunk height. The mean number of nuts per
female tree is steadily increasing as the study continues, which is promising in terms of the
reproductive output of the Curieuse population and its capacity to survive and recover to historical
levels. No trends are apparent in the number of catkins observed per male tree. This is likely due to a
low sample size, which can be remedied by further studies specifically geared towards male
reproduction in Coco de Mer. Some discrepancies were observed in the data, and a review of all
previously collected information resolved anomalous results. As the project continues, it is evident
that only a long-term study will provide data from which robust conclusions regarding the duration
of life stages, as well as trends in growth and reproduction, can be drawn.

The fourth year of the annual Aldabra giant tortoise census was completed in September 2016, with
a total of 114 individuals being successfully located throughout the island. The majority of tortoises
(n=92, 80.70%) were located at the Ranger Station, with the remainder dispersed throughout the
island. Only one individual was found dead this year (108) and overall the population appears to be
stable. Some tortoises have been found to have puncture wounds, peeling on their carapaces and

3



split scutes, all of which is consistent with the population on Aldabra. Two free-ranging juveniles
were newly located (127 and 130), adding to the previous two (126 and 128) being housed in the
nursery. Until July 2016, the giant tortoise nursery also housed 26 hatchlings. However, 25 of these
were stolen. From that point, 33 more hatchlings have been captured and placed in the nursery
under strict supervision in order to preserve the future of this small population. With increased
security measures, the nursery not only offers protection from poachers, but also reduces the risk
from introduced predators on the island, such as feral brown rats (Rattus norvegicus). Captive
tortoise hatchlings will continue to be measured and weighed biannually.

Mangrove tree growth rates continue to be monitored by measuring the girth at breast height (GBH)
of one tree at each of the mangrove waypoint poles. Bruguiera gymnorrhiza exhibited the fastest
mean growth rate per year. Permanent quadrats were established at five locations within the Baie
Laraie mangrove forest with the aim of investigating seedling recruitment and mortality and further
determining species distribution across the mangroves. The two most dominant species in the
mangrove forest are Rhizophora mucronata and Bruguiera gymnorrhiza. In order to determine the
extent of mangrove loss due to sedimentation at the forest fringe, three more quadrats will be
installed in 2017 in near shore areas.

The 2016-2017 hawksbill nesting season is currently taking place. This year saw a slow start to the
nesting season, which seems to have been not only Seychelles-wide but perhaps the same for other
parts of the Western Indian Ocean. Nesting activities peaked in December, and Grande Anse remains
the most heavily utilised nesting beach. Encounter rates also peaked in December, with a higher
number of encounters than the previous season. A relatively high number of untagged hawksbills
had tags applied, and a slightly higher number of previously tagged hawksbills were also recorded.
The few excavations undertaken thus far point to high reproductive success, in line with last season.
Photo identification commenced in 2010 and metal flipper tagging began during the 2013-2014
season. Both methods are being continued in the 2016-2017 season. The number of green turtle
activities and nests has been much lower compared to previous years.

September 2016 marked the end of season two and the beginning of season three of the juvenile
sicklefin lemon shark monitoring. The project has been very successful in providing baseline data to
inform management of this important marine predator. Since the onset of the project, a total of 313
individuals have been caught and tagged, and a total of 81 recaptures have been made, which has
provided a wealth of information on changes in size, weight, and body condition. Population

4



estimates have greatly exceeded preliminary estimates, and population size appears to be stable.
Improved methodologies and equipment have greatly benefitted the project and the sampling
mortality during season two has been reduced to 0% during the 2015/16 season and 0.69% during
the current season. Our team has commenced active acoustic tracking in 2017 in order to establish
the movement patterns and home range of this population, which should help inform park
management regarding critical habitats within Curieuse Marine National Park.

In 2015, beach profiling surveys began on Curieuse Island. Each year, substantial changes to the
profile of the beaches are observed between the Northwest and Southeast monsoon seasons, but
this had not been quantitatively measured before. Data collected by beach profile monitoring
provides a greater understanding of changes in profile, specifically regarding rates of erosion and
sand accretion, which affects the beach area, slope, and sediment volume. Profiling is conducted on
six beaches split into two sections, with each section being profiled every two months. Based on the
data collected thus far, some initial trends are beginning to emerge, which appear to be strongest on
the beaches along the southern coast of Curieuse (Anse Jose, Anse Cimitiere, and Anse Caiman).
Here, sediment movement in respect to both beach area and width is observed to generally be
shifting north-westerly during the Southeast monsoon season and south-easterly when the wind
direction changes in the Northwest monsoon. This is observed to a lesser extent with those beaches
oriented in a more westerly/easterly direction (Anse Laraie, Anse Papaie, and Grand Anse), though
certain results were observed which, if seen to persist in future years may offer some support to
these trends. Although there are still areas of this project which may be improved upon in the
coming year to improve data quality, the data from 2016 has begun to demonstrate the effects that
seasonal changes are having on the beaches of Curieuse.

5



Table of Contents

EXECUTIVE SUMMARY .............................................................................................................. 3

TABLE OF CONTENTS ................................................................................................................ 6
INTRODUCTION ........................................................................................................................ 8
ISLAND CONSERVATION EXPEDITION ....................................................................................... 9
STUDY SITES ............................................................................................................................ 10
TRAINING ............................................................................................................................... 10
COCO DE MER ......................................................................................................................... 12
INTRODUCTION ........................................................................................................................ 12
AIMS ..................................................................................................................................... 13
METHODOLOGY ....................................................................................................................... 13
RESULTS ................................................................................................................................. 15
DISCUSSION ............................................................................................................................ 18
CONCLUSION ........................................................................................................................... 20
GIANT TORTOISES ................................................................................................................... 21
INTRODUCTION ........................................................................................................................ 21
AIMS ..................................................................................................................................... 22
METHODOLOGY ....................................................................................................................... 23
RESULTS ................................................................................................................................. 27
DISCUSSION ............................................................................................................................ 33
CONCLUSION ........................................................................................................................... 37
MANGROVES .......................................................................................................................... 39
INTRODUCTION ........................................................................................................................ 39
AIMS ..................................................................................................................................... 39
METHODOLOGY ....................................................................................................................... 39
RESULTS ................................................................................................................................. 41
DISCUSSION ............................................................................................................................ 44
CONCLUSION ........................................................................................................................... 47
SEA TURTLES ........................................................................................................................... 49
INTRODUCTION ........................................................................................................................ 49
AIMS ..................................................................................................................................... 50
METHODOLOGY ....................................................................................................................... 50
RESULTS ................................................................................................................................. 53
DISCUSSION ............................................................................................................................ 59
CONCLUSION ........................................................................................................................... 61
LEMON SHARKS ...................................................................................................................... 63
INTRODUCTION ........................................................................................................................ 63
AIMS ..................................................................................................................................... 67
METHODOLOGY ....................................................................................................................... 67
RESULTS ................................................................................................................................. 71
DISCUSSION ............................................................................................................................ 83
CONCLUSION ........................................................................................................................... 88
BEACH PROFILING ................................................................................................................... 90
INTRODUCTION ........................................................................................................................ 90
AIMS ..................................................................................................................................... 90

6



METHODOLOGY ....................................................................................................................... 91
RESULTS ................................................................................................................................. 92
DISCUSSION .......................................................................................................................... 103
CONCLUSION ......................................................................................................................... 107
REFERENCES ......................................................................................................................... 108

7



Introduction

Global Vision International (GVI) Seychelles comprises two expeditions based on separate granitic
islands. The Island Conservation Expedition is based on a small granitic island called Curieuse,
located to the north of Praslin. Base camp is located at Anse St. Jose within the Curieuse Marine
National Park. This marine national park has been established since 1979, and represents an area of
14.7km2.

All of GVIs scientific work in Seychelles is conducted on behalf of and at the request of local
partners, using their chosen methodology. GVI supplies experienced staff, trained volunteers, and
equipment to conduct research in support of their on-going work. GVIs key partner in Seychelles is
the Seychelles National Parks Authority (SNPA).

Seychelles National Parks Authority (SNPA): A parastatal organisation partly funded by the
government, responsible for management and research relevant to the protection of the national
parks within Seychelles.

Seychelles Island Foundation (SIF): This organisation manages and protects the UNESCO World
Heritage Sites of Aldabra and Valle de Mai. Their work with the Coco de Mer forests on Praslin and
other endemic plants and animals is closely linked with the flora, fauna and Coco de Mer population
on Curieuse.

8



Island Conservation Expedition

The Seychelles archipelago represents the only mid-oceanic granitic islands on earth. Isolated for 75
million years, Seychelles now hosts a unique assemblage of flora and fauna, many of them extremely
primitive. Such ancient species include endemic palm trees such as the Coco de Mer (Lodoicea
maldivica) and Aldabra Giant tortoises (Aldabrachelys gigantea). However, 200 years of human
settlement has exerted a negative influence on the native biota of these islands. Habitat loss and
fragmentation, as well as invasive species, have resulted in several extinctions and reduced
populations of many species to perilous levels. Natural resource exploitation continues to pose a
serious threat to Seychelles native flora and fauna (Hill 2002).

Curieuse Island is a small granitic island (2.86km2) in the Seychelles, approximately 1km north of the
island of Praslin. Curieuse is notable for its bare red earth intermingled with the unique Coco de Mer
palms, one of the cultural icons of Seychelles - present only in three main populations on Praslin and
Curieuse.

In 1979, Curieuse and its surrounding waters were declared Curieuse Marine National Park in order
to protect the native wildlife. Today, it is home to approximately 130 free-ranging Aldabra giant
tortoises (Aldabrachelys gigantea), found primarily at the Ranger Station but also in smaller numbers
throughout the island. Sea turtles are often found in the surrounding seagrass and reef habitats, and
several of Curieuses beaches represent important nesting sites for female green and hawksbill
turtles, particularly during their nesting season (between October and February). Another key
component of the Curieuse marine ecosystem is the mangrove forest. Mangrove trees are found
most extensively around the lagoon area at Baie Laraie, and bridge the gap between the marine and
terrestrial environments, playing a key role in maintaining optimum reef building conditions for
corals (Obura and Abdulla 2005) as well as providing a vital habitat for birds and fish, including the
sicklefin lemon shark.

The objectives of the Island Conservation Expedition on Curieuse for 2016 focused on the
continuation of the Coco de Mer growth survey, an on-going mangrove monitoring project, the
fourth annual giant tortoise census, on-going sea turtle monitoring, beach profiling, and the second
and third seasons of the sicklefin lemon shark monitoring programme. The fundamental goal behind
all fieldwork is to ensure data collected is relevant and valuable to our project partners. The
information collected by GVI Seychelles is available through SNPA to help inform management
decisions and for use as a baseline for future study.

9



Study Sites

Map of Curieuse Island, showing all current survey sites as undertaken by GVI. Sea turtle nesting
beaches: 1) Anse Caiman/Cimitiere, 2) Anse St. Jose, 3) Anse Mandarin, 4) Anse Laraie, 5) Anse
Papaie, 6) Grande Anse, 7) Anse Badamier. Beaches currently being profiled are 1-2 and 4-6.

Training
Island Conservation Health and Safety
All Expedition Members on the Island Conservation expedition are educated through safety
inductions to work in all survey areas and walk off-path to study sites. Risk assessments have been
carried out for all surveys undertaken. Volunteers are provided with first aid training through the
Emergency First Response course, which is taught on-site.

Terrestrial & Marine Species identification and Field Techniques
GVI relies heavily on volunteers to carry out all of its fieldwork. These volunteers stay for periods of
between two and 12 weeks. To ensure precision and continuity, all volunteers are intensively trained
and have a fully trained staff member accompany them on all field surveys. All expedition volunteers
are required to identify the various life-stages of Coco De Mer palms, understand appropriate
handling and measurements for giant tortoises, sea turtles and lemon shark pups and learn the six
species of mangrove tree present on Curieuse. They are also trained in how to operate equipment
used for each survey, which includes a GPS, PIT tag scanner, refractometer, Abney level and fishing
gear. Training is initially provided in the form of presentations, classroom sessions and informal

10



discussions with the expedition staff, followed by in-field training in practical field techniques. Exams
are given in mangrove and bird species ID. Self-study materials are also available in the form of
textbooks, field guides, journal articles and flashcards. Volunteer progress is monitored and staff
supervision remains vigilant until each volunteer demonstrates a grasp of all procedures and is able
to identify key species.

11



Coco de Mer
Introduction

Coco de Mer (Lodoicea maldivica) despite referencing the Maldives in its Latin name is a palm
endemic to Seychelles, which carries the largest seedpod of any plant on earth. This species is a
classic example of island gigantism, holding records in leaf length, fruit size and weight, and largest
female flowers of any palm (Edwards et al. 2003, summarised in Fleischer-Dogley 2006). The Coco de
Mer (CdM) forest on Curieuse is one of three remaining main global populations of this species, with
the other two found on Praslin Island at Fond Ferdinand (FF) and the Valle de Mai (VM), a UNESCO
World Heritage site. An iconic species of Seychelles, the CdM is also classified as endangered
according to IUCN criteria (Fleischer-Dogley et al. 2011a), which in itself makes it an important study
species. As a renowned flagship species for Seychelles, this palm provides revenue through tourism,
and nut harvesting and sale.

The CdM seed pods (known as nuts) are popular tourist souvenirs, and the suggestive shape adds
to their appeal. Nuts are harvested legally on Praslin and Curieuse, in numbers thought to be
sustainable. They are sold by licensed vendors for 150-400, and come with an individual
identification card to verify their origin. The Seychelles government keeps strict control over the
trees in order to protect the genetic heritage of the islands, and it is illegal to collect or sell
unlicensed nuts. However, CdM nuts are in high demand on the black market, and poaching
represents a significant concern in all populations. Unfortunately, these trees have certain life
history traits that make it difficult for the species to rebound if it becomes vulnerable, such as a late
age at reproduction (20-40 years to reach sexually maturity) and long development period of nuts
(6-7 years) (Edwards et al. 2003).

The Coco de Mer has been the subject of numerous studies; however, few researchers have
investigated the Curieuse population. GVI Seychelles and SNPA conducted a 5 year census of the
islands CdM population (2009-2014). The census produced a population count and basic life stage
information (Sanchez et al 2014; Dunn et al 2015). However, relatively little is known regarding the
growth rate of these palms and how long they take to transition between life stages.

Koch and Kaiser-Bunbury (2010) conducted a CdM growth rate study on Praslin, though Coco de Mer
trees on Curieuse exhibit distinct differences to populations on Praslin (Fleischer-Dogley et al.
2011b) and the results from studies on Praslins populations are not necessarily applicable to the
Curieuse population. In order to harmonise data collection between the three populations, GVI

12



Seychelles initiated a long-term growth rate study, following the same methodology as Koch and
Kaiser-Bunbury (2010), in April 2014. The initial set-up and first monitoring phase highlighted
variation between Curieuse and Praslin (Sanchez et al 2015) but provided only limited growth data
as the project was in its early stages. Monitoring has continued throughout 2016, and the 33 months
of study to date have now produced some more informative data. Some concerns regarding the
methodology have been addressed, while others are currently in review.

Aims
The primary goal of the growth survey is to assess survivorship by documenting the time spent by
CdM at each life stage. By determining leaf and trunk growth rates, we can compare the difference
between life stages and populations. Additionally, assessing inflorescence production among female
trees allows for the assessment of variation in nut production between populations. Also, studying
inflorescences in male trees will allow us to determine variation in catkin production. Understanding
the amount of time trees remain in each life stage improves our understanding of this population,
informing management and ultimately resulting in improved protection for this unique population of
endangered plant.

Methodology
75 trees (15 of each life stage - male, female, immature, juvenile, and seedling) were selected in the
area overlooking Baie Laraie (Figure 1). Seedlings are young plants displaying three or fewer leaves,
while juveniles have more than three leaves but no trunk, and immature trees have more than three
leaves and a visible trunk or defined swell (base of the crown). Adult trees possess trunks, and
produce sexual characteristics (female nuts or male catkins). The methodology and life stage
classification used in this study mirrors that used on Coco de Mer trees studied at VM (see Koch and
Kaiser-Bunbury 2010).

13



Figure 1. Map of tree distribution for the Coco de Mer growth survey by life stage and sex.

Trees were selected for this study based on tree groupings; each grouping has a mixture of life
stages. Another influence on tree selection was whether it would be accessible for researchers.
Some trees were not used as part of the survey simply because they were too tall and researchers
could not access the crown of the palm. Each tree was given a unique code, and their exact position
was recorded using a GPS device.

Between the onset of the study (April 2014) and June 2016, each tree was revisited every monitoring
period (approximately every three months), as per the methodology used at VM. However,
beginning in July 2016, the monitoring period became longer (approximately every six months) due
to the slower growth rates exhibited by the Curieuse population when compared to the population
at VM. Additionally, with fluctuating volunteer numbers and an increasing number of projects, it
became logistically difficult to adhere to the previous three month monitoring periods. Since trees
are not measured exactly every three months to the day, growth is calculated as a daily growth rate.
Therefore, extending the duration of the monitoring period should not affect the usability of and
inferences made from the data collected. In addition, following analysis of growth rate estimates
from the first two years of the study, it was determined that a six month study period was sufficient
for measuring growth in the individual components of the plant, with leaf growth being the critical
measurement in that regard.

14



During the initial setup, the three youngest (most central) leaves of each tree were identified and
labelled L1 L3 according to age (oldest-youngest respectively). The total length of each leaf was
recorded, and a mark was painted at 40cm above the base of the leaf. Referred to as mark A, this
mark was re-measured on each visit to determine leaf growth. The total length of bayonets (early
stage leaves growing from the centre of the crown that have not yet opened into full leaves) was
also recorded at each visit. Once open, a bayonet was then considered a leaf and was set up (using
the aforementioned technique) and given a leaf code (sequential from the previous youngest leaf
(e.g. L4, L5, etc). Measurements for a specific leaf were discontinued when it stopped growing, i.e.
when no change was seen in leaf length after three visits (Koch and Kaiser-Bunbury 2010).

Additional data were collected for each tree, including the number of green leaves, trunk height and
Girth at Breast Height (GBH) of the trunk. During setup, or once they reached an appropriate size,
immature and adult trees were marked with paint 10cm below the swell at the most accessible
point, and the GBH of the trunk was marked and measured at a height of 150cm above the ground
level. With the exception of trunk height, all these additional measurements were re-taken on
subsequent visits to determine growth.

Reproductive information was also collected for adult trees. The number of flowering catkins (male
trees) and the number and classification (primary, maturing, or ripe) of nuts per inflorescence
(female trees) was also recorded.

Results
To date, nine phases of post-setup monitoring have been completed (most recently, period nine was
completed in January 2017). These phases ranged from 63-182 days in duration (mean 98 36 SD),
the variation in duration being largely due to fluctuations in volunteer numbers and available
manpower.

A number of issues have been identified with the methodology, and work is currently on-going to
remedy issues with the existing data set. For example, a review of the existing data has highlighted a
number of measurements that indicate minor negative growth in leaves, trunk height, and trunk
girth. For the purpose of this report, these values have largely been excluded from analysis, and the
data set is currently being standardised over the current longer study periods (six month intervals) in
order to avoid inherent human error in the measurement of small changes in size over short time

15



periods. Also, the mean total leaf length measurements now include leaves that have been added
after the initial setup.

In Period 9 (P9), females exhibited taller trunks (352.3 102.2 SD) with a greater girth (89.0cm
10.0 SD) than males (trunk height: 199.5cm 75.1 SD; girth: 89cm 10 SD). Throughout the study,
immature plants exhibited the smallest mean trunk height (109.4cm 64.5 SD) and girth (75.0cm
12.0 SD) (Table 1).

Table 1 Means standard deviation of growth parameters for male, female, immature, juvenile,
and seedling Coco de Mer trees between Period 6 (P6) and Period 9 (P9) (encompassing 2016). Total
leaf length includes bayonets.

P6 P9 Annual
Parameters
Change
Male trees
Trunk height (cm) 197.6 75.2 199.5 75.1 +1.9
Girth at breast height (GBH) (cm) 84.6 8.0 84.7 8.2 +0.1
Total leaf length (cm) 378 138.9 368.7 143.4 -9.3
Green leaves per tree 13.6 1.9 13.4 2.6 -0.2
Female trees
Trunk height (cm) 351 101.4 352.3 102.2 +0.4
GBH (cm) 87.4 11.4 89.0 10 +1.6
Total leaf length (cm) 361.8 110.8 362 116.3 +0.2
Green leaves per tree 15.5 2.93 15.8 3 +0.4
Immature plants
Trunk height (cm) 105.1 64.3 109.4 64.5 +4.3
GBH (cm) 75.5 12 75 12 -0.5
Total leaf length (cm) 428.8 166.3 430.4 151.2 +1.6
Green leaves per tree 10.9 2.5 11.1 2.4 +0.3
Juvenile plants
Trunk height (cm) - - -
GBH (cm) - - -
Total leaf length (cm) 444.2 167.7 431.3 182.5 -12.9
Green leaves per tree 4.92 1.7 5.3 1.7 -0.4
Seedlings
Trunk height (cm) - - -
GBH (cm) - - -
Total leaf length (cm) 228.9 90.2 209.6 90.7 -19.3
Green leaves per tree 2.4 0.5 2.5 0.7 0

However, over 2016, females exhibited a slower average increase in trunk height (0.4 cm) than
males (1.9 cm). Annual increase in trunk height for immature trees exceeded that of male and

16



female trees (4.3), though immature trees exhibited negative growth in GBH (-0.5), which is
attributed to human measurement error (Table 1).

Juvenile trees produce the longest average leaf length (431.3cm 182.5 SD) followed by immature
(430.4cm 151.2 SD), male (368.7 cm 143.4 SD), female trees (362.0cm 116.3 SD), then seedlings
(209.6cm 90.7 SD) (Figure 2). With regard to leaf growth rate, the leaves of immature and male
trees grew fastest (0.29cm/day 0.48 SD and 0.29cm/day 0.36 SD, respectively). Female trees
grew at a slightly slower rate (0.27cm/day 0.37 SD), followed by juveniles (0.22cm/day 0.32 SD)
and seedlings (0.15cm/day 0.2 SD) (Figure 2).

600 0.35

500 0.3
Mean total leaf length (cm)

Mean growth/day (cm)

0.25
400
0.2
300
0.15
200
0.1

100 0.05

0 0
Male Female Immature Juvenile Seedling

Figure 2. Mean total leaf length (mean SE; blue bars) and mean growth rate (red line) for the
various life stages and sexes of Coco de Mer trees for periods 6 to 9 (encompassing 2016).

Overall, relatively little variation exists in the number of green leaves per tree over 2016, while
certain individuals appear more productive than others. On average, females produce the most
leaves (15.8 3 SD), then males (13.4 2.6 SD), immature (11.1 2.4 SD), juvenile (5.3 1.7 SD) and
seedlings (2.5 0.7 SD) (Table 1). Male inflorescence production (number of flowering catkins) did
not exhibit any clear trends. The maximum number of flowering catkins recorded on a single tree
was two and the minimum was zero (Table 2).

17



Table 2. Mean SD and maximum number of flowering catkins recorded per tree initial setup (April
2014) until the end of period (P) 9 (January 2017).

Setup P1 P2 P3 P4 P5 P6 P7 P8 P9
No. of 0.87 0.60 0.67 0.87 0.87 0.43 0.33 0.73 0.76 0.67
catkins 0.64 0.63 0.72 0.64 0.74 0.51 0.49 0.46 0.73 0.49
Maximum 2.00 2.00 2.00 2.00 2.00 1.00 1.00 1.00 2.00 1.00

Female nut production appears to be increasing over time (Figure 3), with the average number of
nuts per tree rising from 1.53 1.5 SD to 2.07 1.6 SD over 2016.


2.5

2
Mean #nuts per tree

1.5

0.5

0
Setup P1 P2 P3 P4 P5 P6 P7 P8 P9

Figure 3. Mean number of nuts per Coco de Mer tree from initial setup (April 2014) until the end of
P9 (January 2017).

Discussion
A comparison of Curieuse data against the two main Praslin populations, Fond Ferdinand and Valle
de Mai suggests distinct differences (summarised in Sanchez et al. 2015), which can be attributed to
varying environmental conditions among sites and the high phenotypic plasticity of L. maldivica
(Fleischer-Dogley et al. 2011).

Growth trends over this year appear to be different from the previous, namely in that females no
longer exhibit the fastest mean leaf growth rates among the life stages in our study area. Instead,
males and immature trees experience similar mean growth rates higher than that of females. And
while mean growth in male trees has decreased slightly over 2016, the mean growth rate of
immature trees has increased (Figure 2). For immature trees, this increase is logical since our data
suggests that both mean leaf length and growth rates increase in sub-adult trees over time. Savage
and Ashton (1984) reported that juvenile petiole length was positively correlated to canopy height,

18



indicating that the leaves increase in growth rate in order to compete for canopy space. This could
partially explain why growth rates are lower in adults, as they may reduce leaf growth upon
becoming established in the canopy and divert more energy into trunk growth and reproduction.
Moreover, the observed reduction in total leaf length in males, juveniles, and seedlings is likely due
to the addition of new leaves, which, being smaller in size, will result in a lower average total leaf
length value.

A number of factors could result in the observed decrease in leaf growth rates by adult trees,
particularly females. Firstly, the average number of nuts per tree has been steadily increasing since
the onset of this study; in fact, the mean number of nuts per tree has more than doubled since early
2014. Moreover, Dunn et al. (2014) observed an even lower mean number of nuts per tree (0.797,
N=546) during the island-wide CdM census, which preceded the current growth study. Silverton
(1987) suggested that the energy expended in seed production could explain the reduced size of
female trees on Praslin, in comparison to males. As such, this observed increase in reproductive
output on Curieuse is potentially reducing available energy for leaf growth. Although mean nut
production on Curieuse is far lower than that of females in FF & VM (8.86 and 6.38 nuts per tree,
respectively; Fleisher-Dogley 2006), growth could also be limited by a combination of increased nut
production coupled with environmental factors such as the relatively poor soil conditions on
Curieuse.

Hodgkiss et al. (2016) highlighted a correlation between trunk height and nut production, suggesting
that as trees grow and age, greater numbers of nuts per tree should be expected. At FF and VM on
Praslin, female trunk height was 7.79m and 9.26m respectively (Fleisher-Dogley 2006) compared to
approximately 3.5m on Curieuse, and if tree growth continued over time, females on Curieuse could
potentially increase nut production to higher levels than at present.

To date, only one tree has been observed to change life stage, developing from immature to male
with the production of its first recorded catkin. However, this tree has a trunk height of
approximately 250cm, which is taller than the male mean, and it produced its first catkin following
the initial monitoring phase. Therefore, it is possible that this individual either did not have a catkin
present or was incorrectly identified during set-up. Additionally, immature trees display negative
GBH growth across the year, which is likely due to human error during surveying.

19



As noted by Hodgkiss et al. (2016), there is no strong correlation to monthly rainfall in the number or
frequency of flowering catkins produced by male trees. Moreover, no monthly or seasonal variation
is evident either. While flowering catkins may be recorded on some trees, the number of male trees
surveyed in one month, for example, can easily be less than five. Therefore, the sample size or
sampling frequency would need to be increased in order to detect any patterns in male or female
reproduction with any level of certainty. For example, consecutive interns could contribute to a
robust, long-term data set related to nut and/or catkin production as part of their independent
project.

It is also suggested that the trees in our study are not completely representative of the range of
individuals and environmental conditions on the island. For example, Hodgkiss et al. (2016) noted
that there are much larger and presumably more productive trees on Curieuse than those sampled
in our study, which has the potential to result in bias. This is particularly evident in the fact that the
trees studied were selected in part due to their accessibility for researchers lacking advanced
climbing equipment and specialised training in tree climbing. Therefore, while this study provides an
important comparison against other CdM populations as well as insight regarding the baseline
biology of this species on Curieuse, care should be taken when making generalisations regarding this
population based on only 75 smaller trees in a relatively small area out of a population of over 6000
individuals over the entire island.

Conclusion
Following 33 months of study, a strong set of baseline data regarding growth, leaf size, trunk size,
and reproduction has now been established and should provide a basis for comparison well into the
future. A clear and promising trend is becoming evident in female nut production, and this
information, alongside growth patterns, is critical to the effective management of the species on
Curieuse. A limited harvest of CdM nuts has been conducted on Curieuse in recent years, and further
information regarding island-wide nut production should aid Seychelles National Parks Authority in
determining a level of sustainable harvest. As such, it may be extremely useful to recruit consecutive
interns to collect further information regarding nut and catkin production outside of the trees in our
growth study. This project must continue as a long-term research project in order to answer
questions regarding the duration of CdM life stages, growth and reproduction, and ensure that this
population of rare and endangered plant receives management considerations based on the best
available science.

20



Giant Tortoises
Introduction
Currently, the only natural wild population of the Aldabra giant tortoise (Aldabrachelys gigantea) is
believed to be found on the Aldabra atoll (Bourne and Coe 1978). Most islands in the Western Indian
Ocean, including the inner granitic islands of Seychelles, once hosted wild populations of giant
tortoises in the past (Stoddart et al. 1979). However, populations declined during the 1700s and
1800s as settlers exploited the giant tortoises for food and trade and exported them aboard ships
(reviewed by Gerlach et al. 2013). All members of the species remaining in the Inner Granitic islands
are thought to have been relocated from Aldabra.

Between 1978 and 1982, The Curieuse Experiment saw approximately 250 Aldabra giant tortoises
transferred from Aldabra to Curieuse Island (Stoddart et al. 1982) in an attempt to boost tourism,
encourage scientific studies, and protect a species currently listed as Vulnerable (Tortoise and
Freshwater Turtle Specialist Group 1996). The first stage of introductions took place in 1978, with 95
giant tortoises brought to Curieuse (Stoddart et al. 1982). Hatchlings were discovered in 1980,
indicating successful breeding in the introduced population. An additional 78 tortoises were
introduced in 1980, followed by 74 more in 1982. Three other tortoises (of unknown origin) were
also released on Curieuse in 1983 and 2000 (Gerlach et al. 2013), and a small number of other
individuals have since been introduced to the island (former pets).

The giant tortoises were initially released on Curieuse near the Ranger Station at Baie Laraie. While
the majority of the tortoise population remains near the Ranger Station, some have migrated and
individuals can now be found throughout the island (Sanchez et al. 2015; Samour et al. 1987).

Since the giant tortoises were relocated to Curieuse, several population censuses have been
completed with varying results. In 1986, the Zoological Society of London found 144 individuals,
though a few months later a survey by the warden located only 102, and attributed the decline in
numbers to theft and rat predation (Samour et al. 1987). In 1990, 117 tortoises were re-spotted
during a census by Hambler (1994), and a less complete census in 1997 coordinated by Mortimer
(1998) found 110 tortoises. There is evidence that tortoises are reproducing on Curieuse, as
hatchlings have been found by SNPA Rangers each year, yet poaching and predation by rats could be
mortality factors in the population. Heightened efforts to increase recruitment into the population
and hatchling survival have been made by SNPA. A nursery was established to protect hatchlings

21



found by Rangers and GVI Seychelles staff from any predators, poaching, and handling by tourists. At
the age of approximately five years old, at which point they are large enough to not be threatened
by predation from rats, the tortoises are PIT tagged and released into the free-ranging population.

In 2013, the first annual census was undertaken by GVI Seychelles and SNPA. Tortoises were tagged
with passive integrated transponder (PIT) tags as a permanent means of identification, and various
techniques were used to assist in identifying each individual. Specifics of those methods can be
found in Dunn et al. (2014). In 2013, a total of 125 tortoises were found, many of which were
identified from previous surveys. However, unless tortoises have gone unnoticed, this total of 125 is
much less than the 250 that were originally released on the island over 30 years ago. The original
discovery of the decrease in population size had been alarming, and stresses the importance of
conducting an annual census and monitoring the population.

Aims
The primary aim of the annual census is to reveal how many of the original tortoises brought over
from Aldabra remain on Curieuse, along with their basic movements across the island. Over time,
this census is designed to determine tortoise growth rates, home range, age (when followed from
hatchling size), and the size at which tortoises begin to display sexual characteristics. Aldabra
tortoises have been researched on Aldabra Atoll; however, climatic differences between the atoll
and the inner granitic islands likely have an impact on the habits and growth rates of the tortoises.
This census will also provide baseline data that could be expanded upon to further investigate areas
such as food preferences and activity cycles. The lack of an increase in the population size raises
questions related to population recruitment and hatchling survival. The census will hopefully
increase the chances of discovering any hatchlings that have successfully hatched in the wild. An
additional aim is to locate as many tortoise nests as possible, and subsequently conduct excavations
in an attempt to shed light on rates of clutch size and hatching success.

In addition to the yearly census of the free-ranging tortoises, there is also a biannual census of
hatchlings in the nursery, where similar growth measurements are taken to allow us to track growth
of the hatchlings.

22



Methodology
This census conducted by GVI is carried out in two parts, with the majority of the individuals on
Curieuse being surveyed with one methodology and all the hatchlings in the nursery being
monitored in a slightly different way.

Free-Ranging Giant Tortoise Census
In order to locate tortoises to include in the census, GVI staff and volunteers surveyed the island
searching for individuals using a map (Figure 4) marked with location codes from a previous tortoise
study at Curieuse (Lewis et al. 1991). Those areas known to be favoured by the animals from
previous studies and include the Ranger Station, Anse Papaie, Grand Anse, Fond Blanc, Point Rouge,
the North and South mangroves, Anse Badamier, and the North Coast. Time was spent searching all
areas that the teams could access. However, it has been assumed that if GVI personnel could not
traverse certain terrain, then neither could the tortoises. Signs of tortoise activity (e.g. droppings)
were also used to indicate their presence. If none were found in the more difficult to access areas, a
search may be abandoned in that location for the rest of the season.

Figure 4. The map of Curieuse Island used in the annual census, originally used by the Oxford
Expedition team in 1990 (Lewis et al 1991). The map is organised into areas depending on the
accessibility for the giant tortoises.

Each time a tortoise was encountered, it first needed to be identified to determine whether or not it
had been encountered previously that year. There are a number of ways to identify previously
encountered individuals. This includes a unique ID number which was applied to the carapace of the
tortoise using a yellow marking stick (Sharpie MeanStreak) on the 4thor 5th dorsal scute (Figure 5).

23



This allowed for quick identification of individual tortoises without the need to scan for PIT tags.
However, this mark is not permanent and lasts only weeks or months. Therefore, if a tortoise was
unmarked, it was scanned for an existing PIT tag using a Trovan scanner.

Figure 5. Diagram of a tortoise carapace (upper shell). Dotted line a. demonstrates the over-the-
curve carapace length (OCCL) measurement, and the width of the 3rd dorsal scute is measured along
dotted line b. Dotted line c. demonstrates the point on marginal scute 11 on either side past which
the tail is considered to be long.

When tortoises were initially relocated to Curieuse, and again during a census on Curieuse in 1997, a
metal disc was attached to the 4th dorsal (D4) scute. A plastic disc was also attached to D4 to a
majority of the tortoises in 2013. If any discs were still present, the numbers were recorded. If it was
obvious there once was a tag due to left over glue even though the disk was missing MD for
missing disc would be recorded. If neither the tag nor glue from the Aldabra and/or the Curieuse
census discs was present, then an N for never present was recorded.

If it was determined that a tortoise had not yet been encountered that year, then the date and time
was recorded. In order to aid future surveys, and to monitor the movement of PIT tags throughout
the tortoises body, the PIT Tag Location (where the PIT tag was detected with the scanner, e.g. left
rear hip or tail) and Scan Method (how the reading was obtained, e.g. through the carapace or from
underneath the plastron) was also recorded. The location of each tortoise encounter was recorded
using a GPS. Additionally, the location was matched to an Area Number on the map being used
(Figure 4) in order to allow for current data to be collated with historical data.

24



Figure 6. The width measurement as recorded over the carapace (from Gaymer 1968).

Figure 7. For the toenail measurement, the second nail from the outside on either rear leg is used.

Various measurements were taken for each individual tortoise encountered to allow for growth
studies. The carapace width and the over-the-curve carapace length (OCCL), as well as the width of
the 3rd dorsal scute were measured (Figures 5 and 6). Three categories (tail, plastron, toenails) can
be indicative of sex, therefore, a plastron can be defined as being "concave", "slightly concave" or
"flat". The length of the tail was recorded as being "long" or "short", with long tails being those that
extended past the midline of the 11th marginal (Table 3) and short tails being those that didnt. A
measurement of the second toenail from the rear of a back leg (Figure 7) was taken.

25



Table 3. The visual characteristics that could indicate the life stage or sex of a giant tortoise.

Potential Reproductive Unknown (immature
Full Male Juvenile
Male Female male or female)
Tail *L=long n/a S=short S=short S=short
SC = slightly
Plastron *C=concave F=flat F=flat F=flat
concave
OCCL-mid 70 cm 70 cm 70 cm *<70 cm 70 cm
White Line n/a n/a n/a 2-3 n/a
Growth since the
n/a No n/a n/a
1997 census?
Female Reproductive
n/a n/a *NB or CoCop n/a n/a
Activity

* Indicates a character that is definitive independent of other characters

Grey background -Indicates a supporting character
NB - Nesting behaviour is defined as nest digging or egg laying
CoCop - Cooperative copulation defined by observed penetration

A scale to determine the thickness of the white lines between scutes was developed, with the theory
that a thick white line indicates that a tortoise is not yet fully grown. These lines, if present were
measured to the nearest millimetre. After all data had been collected and the yellow ID number
repainted if needed, a photo of the ID number was taken to identify the tortoise along with a photo
of the 3rd dorsal scute and any distinguishing marks and injuries.

Sexually mature males may typically have long tails, concave plastrons and short toenails while
females have the opposite. The apparent sex of a tortoise was determined by the criteria defined in
Table 3, with this criteria being introduced to the methodology in 2015. Only sexually mature males
can be sexed using visual characteristics alone; a small tortoise with a short tail and flat plastron
could either be an immature male or a female. Only a tortoise that has been seen digging a nest,
laying eggs, or cooperatively engaging in copulation can be confidently sexed as female, though
these events are not seen often. Juveniles were classed as having an OCCL of less than 70cm.
Therefore, for the purpose of data analysis, tortoises are classed as: full male, potential male
(those starting to display male sexual characteristics, specifically a slightly concave plastron),
reproductive female, juvenile, and unknown (immature male or female).

Monitoring of Captive Hatchlings
Similar growth data was collected for the hatchlings kept in the nursery at the Rangers Station every
six months, in May and November each year. Hatchlings in the nursery at the end of 2015 had
previously been classed by age based on when they were found and whether they were obviously

26



new-born hatchlings from the most recent season (i.e. approximately six months of age in May 2014,
for example, indicates they were found as new-borns during the previous hatching season around
the end of 2013). Hatchlings were measured (width, OCCL, width of 3rd dorsal), weighed, and marked
with Sharpie MeanStreak on a specific marginal scale with a unique pattern for future identification.
Photographs were taken of the carapace, plastron and any distinguishing marks such as
extra/missing scutes (i.e. front, back, top of carapace, underneath of plastron, right side, and left
side).

Results
There have now been 130 tortoises included over the course of the GVI censuses on Curieuse since
2013, with four individuals (068, 109, 116, and 119) being known to already be dead from 2015. In
this years annual giant tortoise census, there were a total of 114 tortoises encountered and
included. One individual (number 108) was initially found alive, measured and included in the
survey, but then found dead several months later (this individual has therefore been included in the
analysis). This leaves 10 individuals that were not found at all in 2016, four of which were also not
encountered in 2015 (006, 038, 110, and 115), and 2 which have not been included in the census for
the past three years (121 and 123). This leaves a remaining four individuals that have evaded the
field teams this year. Tortoise number 129 was included in the 2015 census but not PIT tagged, this
was however able to be done this year. Also, an additional tortoise was discovered this year, and it
was also tagged and added to the census as number 130.

Table 4. Areas of Curieuse (based on sections in Figure 4) where free-ranging tortoises were found
in each census and the totals for each year.

Oxford Area Number Samour et al. Lewis et al. 2013 Initial 2014 Initial 2015 Initial 2016 Initial
and Section (1986) (1991) encounter encounter encounter encounter
1 - Grand Anse 8 0 9 8 7 11
2 - Anse Papaie 1 6 3 3 4 2
3+4 - Ranger Station 109 84 104 97 93 92
5 - North Mangroves 7 18 3 3 2 4
6 - South Mangroves 0 0 0 1 0 0
8 - Anse St. Jose 0 0 0 0 1 2
9 - Anse Badamier 0 3 0 1 1 1
11 - North of Grand Anse 6 0 1 1 2 1
12 - Mt. Libine 5 2 1 4 1
12
13 - Fond Blanc 0 2 3 1 0
14 - North Coast 0 1 0 0 0 0
17 - North Ridge 1 1 1 0 0 0
Total 144 118 125 118 115 114

27



Most of the tortoises that were found in this years census (Table 4) were encountered at the Ranger
Station (n=92, 80.70%), which has not changed much over the years since the animals were first
released there. It was at Grand Anse (n=11, 9.65%), the island's largest beach that the second largest
concentration of individuals was found. Anse Papaie (1.75%) and Anse St. Jose (1.75%) both had two
tortoises found on them, with three other locations (Anse Badamier, the area North of Grand Anse,
and Mt. Libine) all having one tortoise found in each area. It seems that the majority of tortoises
released on the island have not dispersed far from the release location. However those that have
dispersed do not seem to be prevented by terrain that may seem impassable, with some moving as
far as Anse Badamier in the North, and Anse St. Jose in the South.

For the purpose of this report, when considering the sex of tortoises, it was largely only external
characteristics that were used. Unless a tortoise was observed digging a nest or engaging in
cooperative copulation, it cannot be identified as female because it could potentially be an
immature male. In this years census, two individuals were initially classed as female. However, there
were no classifying female observations, and they had been noted as having slightly concave
carapaces indicating a potential male. Taking such factors into consideration, these two individuals
were re-classified accordingly for analysis. So, when considering only external characteristics (Table
5), this years census population of 114 individuals contains 2 juveniles, 31 potential males, 71 full
males, and 10 tortoises that were classified as unknown, as they havent displayed male
characteristics or female behaviour. Since 2015, three were classified as potential males but had
been identified as full males in 2016. As such, it is possible that these tortoises have recently
developed male sexual characteristics. An additional two tortoises appear to have changed from
unknown to potential male this year. Moreover, there are an additional nine individuals that appear
to have changed classification either from potential male to unknown or full male to potential male,
which would suggest variation in judgement of sexual determination characteristics between
researchers during one of the previous two censuses.

28



Table 5. Mean standard deviation of measurements taken for each age/sex class in the 2016
census. Note: there is no standard deviation for the 3rd dorsal measurements of juveniles (n=1,
measurement missing).

Juvenile Potential Male Full Male Unknown
(n=2) (n=31) (n=71) (n=10)
Mean SD Mean SD Mean SD Mean SD

3rd Dorsal (cm) 19.50 N/A 29.36 3.11 39.21 35.62 27.14 2.47
Width (cm) 67.75 1.77 103.34 12.12 126.20 8.41 94.01 9.93
OCCL (cm) 64.50 0.14 101.31 15.67 130.25 10.83 90.12 8.70
Toenail Length (cm) 1.80 0.00 3.70 0.72 4.52 0.57 3.01 0.71

Table 5 displays the average width, OCCL, 3rd dorsal width and toenail length for tortoises in each
age/sex class in 2016. When toenail length is considered as a percentage of the OCCL it is potential
males (3.65%) that have the longest, followed by full males (3.47%), then unknown (3.34%) and
juvenile (2.79%). This does differ somewhat to 2015, where it was the unknown category with the
longest toenails. A very strong relationship (R2=0.97; Pearsons Correlation Coefficient) was observed
between carapace length and width in all 114 tortoises (Figure 8).

180.0

160.0

140.0

120.0
Length (cm)

100.0 2
R = 0.97
80.0

60.0

40.0

20.0
125 108 037 112 078 024 049 023 080 040 009 032 088 013 084 034 113 051 010 061 101 015 100
Individual Tortoises

Figure 8. The relationship between OCCL (black line) and carapace width (blue line) measurements
over the 2016 giant tortoise census.

29



Average increases in 3rd dorsal size, carapace width, and OCCL measurements between 2015 and
2016 for tortoises in each of the age/sex classes are shown in Table 6. In 2015, juveniles had the
highest growth of all the classes; however, this years data was insufficient for analysis on this group,
as there were only two individuals and one having a missing data point. This year, growth in the 3rd
dorsal scute was highest in full males followed by unknown, then potential males. However, growth
in width over the previous year saw potential males exhibiting the highest growth rate, followed by
unknown, then full males. The unknown class exhibited the highest growth in OCCL, followed by
potential males and ultimately full males.

Table 6. Mean growth measurements for each of the age/sex classes between the 2015 and 2016
censuses.

Potential Males Full Males Unknown
(n=31) (n=69) (n=9)

Mean SD Mean SD Mean SD

3rd Dorsal
(cm) 0.29 0.70 4.59 36.12 0.31 0.51

Width (cm) 1.24 3.60 0.25 1.23 0.76 1.38

OCCL (cm) 0.94 1.31 0.15 0.99 1.13 1.87

There has been some tag loss during this years census. The metal Curieuse discs installed on
tortoises in 1997 have been lost by three individuals since the 2015 census. Of the 113 tortoises
given plastic discs in 2013, 111 of these are being analysed in this report (due to death or not being
found this year), and of these, 54 (48.64%) have lost their plastic disc. Nine tortoises have lost their
plastic disc since 2015 census. With regard to the metal Aldabra discs installed when individuals
were first relocated to Curieuse, none of the six individuals included in the first GVI census in 2013
have lost their tags to date.

Table 7. Mean growth for each of the age/sex classes between the 2013 and 2016 censuses.

Potential Male Full Male Unknown
(n=30) (n=71) (n=10)
Mean SD Mean SD Mean SD
OCCL (cm) 2.72 3.49 1.32 2.07 3.74 1.11
3rd Dorsal
(cm) 0.49 0.88 4.49 35.62 0.99 4.55

30



Table 7 shows the average growth in centimetres for three of the age/sex classes since the
commencement of the GVI census in 2013. The unknown category showed the greatest growth in
OCCL over this time, with 3.74cm average increase (growth rate of 0.10cm/month), followed by
potential male (2.72cm, growth rate of 0.08cm/month) and full male (1.3cm, growth rate of
0.04cm/month). The growth rate for the 3rd dorsal measurement was the highest in full males
(1.50cm/month), followed by unknown (0.33cm/month) and potential males (0.14cm/month). For
carapace width, the observed trend was again different, with unknown exhibiting the highest
(2.53cm/month) growth rate, followed by full males (1.07cm/month), and finally potential males
(0.98cm/month).

Prior to July 2016, there were 28 tortoises in the nursery, including 26 young hatchlings and
numbers 126 and 128 (approximately five years old). There were 15 hatchlings that had
measurements taken over 12 months (May 2015 to May 2016), which had an average monthly
growth rate of 0.28cm (3.39cm/year). However, 10 hatchlings had been in the nursery long enough
for measurements to be gathered over the space over 24 months (May 2014 to May 2016) which
had an average monthly growth rate of 0.33cm (3.95cm/year). Only one hatchling (RM11) had 18
months of data (November 2014 to May 2016), exhibiting a monthly growth rate of 0.27cm
(3.27cm/year). All of the 29 captive hatchlings were showing consistent positive growth (Figure 9).
Unfortunately, 25 of the hatchlings were stolen by poachers this year (Meriton-Jean & Bonnelame
2016), with RM1, 126, and 128 being left behind. Tortoise 126 had data for 12 months (May 2015 to
May 2016), and exhibited a monthly growth rate over six months of 0.53cm (6.38cm/year), whereas
number 128 had a growth rate of 0.33cm/month (4.01cm/year).

31



45

40

35

30
OCCL (cm)

25

20

15

10

0
Nov-14

Jan-15
Feb-15
Mar-15
Apr-15
May-15
Jun-15
Jul-15
Aug-15
Sep-15
Oct-15
Nov-15
Dec-15
Jan-16
Feb-16
Mar-16
Apr-16
May-16
Jun-14

Aug-14
May-14

Dec-14
Jul-14

Sep-14
Oct-14

Figure 9. The tracked growth measurements over 2014, 2015 and 2016 surveys displaying changes in
the OCCL for the 28 captive hatchlings in the nursery before July 2016 (note: this includes individuals
126 and 128).

Since August this year, GVI and SNPA personnel have been locating new hatchlings throughout the
island. At the time of writing there were 33 new hatchlings in the nursery as well as the remaining
three. With only one set of measurements (Table 8) for these tortoises, growth over time is not able
to be observed at this stage, however average 3rd dorsal is 3.53cm, average OCCL is 9.30cm, average
weight 0.06kg and average width 9.67cm.

Table 8. The average 3rd dorsal scute measurement, OCCL length, weight and carapace width of the
33 hatchling tortoises that were in the nursery as of the October 2016 and the standard deviation for
each measurement type.

Hatchlings Found in 2016

(n=33)
Mean Standard Deviation

3rd Dorsal (cm) 3.53 0.06
OCCL-mid (cm) 9.30 0.10
Weight (kg) 0.06 0.003
Width (cm) 9.67 0.10

32



Discussion
The 2016 annual giant tortoise census was successfully undertaken and resulted in a total of 114
free-ranging individuals being documented throughout the island. There were 10 tortoises not
accounted for in this year's census. Four of these individuals were also not found in 2014 and an
additional two have not been measured for the past three years. It is possible that these animals
have either died without being discovered or migrated to inaccessible areas of the island. As such,
there are a remaining four tortoises not found this year that were in the 2015 census. They were
previously located in areas outside of the Ranger Station (areas 1, 5, 11 and 12; Table 4), and it is
likely that they have simply evaded search teams. Tortoises have proven again this year that they are
capable of passing over difficult terrain. This year has also demonstrated that some undiscovered
tortoises may remain, as we are still finding tortoises new to the census.

In the 2016 census, tortoise number 108 was initially encountered alive, but was found dead behind
the Ranger Station several months later. When combined with individuals known to be dead from
previous years, a total of at least five have died during the course of annual censuses on Curieuse. It
is difficult to determine whether losing this number of adult tortoises over four years should be
considered normal for a giant tortoise population, especially with regard to Curieuse Island. Bourn &
Coe (1978) did not specify a mortality rate at Aldabra Atoll, but over a period of two years examined
547 deceased tortoises, stating that the major causes were becoming trapped or overturned, being
unable to move into adequate shade, malnutrition, disease, and age. It is not possible to make direct
comparisons of course, as Curieuse is a very different habitat with different environmental factors
affecting mortality i.e. more shade and tree cover, more consistent rainfall, etc. It may be that some
of the individuals unaccounted for in the past few years have died undetected, but this cannot be
known for certain. However, when considering that the GVI census has only recorded five tortoises
to have died within the past four years, it may be cautiously suggested that the mortality rate does
not appear as high on Curieuse as elsewhere, although it is only with more years of consistent data
collection that a better picture of mortality rates can be established.

The locations of where tortoises were initially encountered has been somewhat varied during
previous censuses (Table 7). This includes data from surveys in 1986 (Samour et al. 1987), and 1991
(Lewis et al.) as well as the 2013, 2014, 2015 and 2016 GVI Curieuse censuses. In each census the
majority of the giant tortoises have been located in areas three & four which make up the Ranger
Station and immediate surrounding area, with 92 being encountered there this year. There was
initially an attempt to keep all of the animals in this area, a moat which over time was breached,

33



meaning all are now free to roam (Samour et al, 1987). The fact that tortoises are not distributed
evenly across the island has been accounted for by difficult terrain and loss through poaching
(Hambler 1994, Samour et al. 1987). Despite the steep and often rocky terrain, certain individuals
have left the Ranger Station and have proven that giant tortoises are more than capable of
navigating over difficult ground (Hambler 1994, Samour et al. 1987, Stoddart et al. 1982). Of
particular note are tortoises 014 and 095, having previously made the journey over to the south side
of the island and remained there. These two individuals have had to cross very difficult, rocky
terrain, with at least number 014 making the journey twice after being returned to the other side of
the island by SNPA.

This year, 11 individuals have been encountered at Grand Anse, which represents an increase over
previous years. Grand Anse is one of the largest beaches on the island, with a large stretch of
swamp/forest behind it. It is possible that there is a feature of this habitat (e.g. food availability,
shelter, etc.) that results in tortoises remaining there. Three individuals were also found here for the
previous three years, indicating they permanently moved to this area. It will be interesting in future
years to observe whether the number of individuals in this area continues to rise. Also, tortoise 124,
which has been at Anse Badamier on the north coast of Curieuse for the past three years, appears to
have established residency in the area.

Curieuse Island is a vastly different habitat to Aldabra Atoll, and different environmental pressures
therefore affect the tortoise populations. It is not known at what age or size Curieuse tortoises begin
to display sexual characteristics; therefore, we are not able to confidently state how many males and
females are on the island. Reportedly on Aldabra, giant tortoises become sexually mature when they
reach a size of 70cm OOCL and a 3rd dorsal scute of more than 21cm (Lewis et al. 1991). In 2015, it
was considered, using these criteria that the 15 tortoises in the unknown category encountered that
year would have to be females as they had not displayed any male sexual characteristics. However,
of these, seven were then classified in 2016 as potential males. If we were to consider these same
criteria from Lewis et al. (1991), then the 10 unknown tortoises from this year's census would also
need to be considered as female, which would create a sex ratio of adult male to female of 7:1. All
tortoises displaying slight indications of being male (i.e. slightly concave plastron) have an OCCL of
greater than 77.5cm and a 3rd dorsal of more than 23.3cm. This implies that simply because a
tortoise is larger than the threshold given by Lewis et al. (1991) and has not yet displayed sexual
characteristics does not mean that they will not. It could be that tortoises on Curieuse are either
reaching sexual maturity at a later age, or growing at a faster rate. It has been previously

34



hypothesised that tortoises on Curieuse may grow at a faster rate than on Aldabra, based on data
from tortoises with a known age (Sanchez et al. 2015). This could be the case, though reports of the
age of tortoises are unreliable. Only when a tortoises exact age is known will we be able to
accurately compare growth rates on Curieuse and Aldabra, as well as determine the ages at which
tortoises are displaying sexual characteristics. It will surely take several years before we can begin to
make deductions on this matter, and the recent theft of a large number of hatchlings is an
unfortunate setback in this area.

In the past, a theory has been suggested that female giant tortoises have proportionally longer
toenails in relation to their size than males. In 2015, full males had the shortest toenails, and it was
the unknown class that was found to have the longest toenails in relation to their body size. Despite
differences in toenail length between the age/sex classes being small, it may have appeared to offer
some support to the theory, as any females in the population would have been in this class.
However, the same was not seen again this year, with potential males having the longest relative
toenail length. It is difficult to say exactly why this might have occurred, but it should be considered
that seven of the individuals classed as unknown last year have become potential males in the 2016
census, which could explain why it is now this class displaying the longest toenail length. This theory
still appears to be somewhat untested and is currently not well supported by our data. For this
theory to be examined thoroughly it will require much more investigation in future years.

In the past, GVI census size measurements have been compared with data collected in the 1997
census (Mortimer, 1998). It has however been decided that from this point on, it would be of more
use to be able to compare annual data consistently taken by GVI (2013, 2014, 2015, and 2016). The
width measurement has only been taken since 2014; therefore, a complete comparison could only
be done for OCCL and 3rd dorsal for the 111 (Table 8) individuals for which this data was available.
For this comparison, tortoises have been categorised into age/sex classes according to the 2016
data. There are 10 unknown tortoises able to be considered, one of which had appeared to decrease
in OCCL. Two individuals in the unknown class appear to exhibit zero growth in OCCL in the
intermittent years, one at 96.5cm and the other at 101.7cm (note that measurements in the
intermittent years were not consistent). These tortoises may have reached full size and could
therefore be viewed as female as they are still showing no signs of male sexuality. However, there
are seven of these tortoises that still appear to be growing although at varying OCCL growth rates
from 0.17cm/year to as much as 4.2cm/year. Since the unknown class would contain immature
males as well as females, it is expected that they have the highest average OCCL growth rate of

35



0.10cm/month. Full males grow at a mean rate of 0.04cm/month of growth, which suggests they are
growing more slowly and are closer to being fully grown.

The fact that several tortoises throughout the census had negative growth between 2013 and 2016
should be viewed with caution, and the possibility of measurement error considered. When
considering growth increase over the previous four years for all tortoises, it is worth noting that
there are obvious discrepancies in measurements, wherein recorded measurements are smaller
than previous years. This demonstrates the importance of a consistent methodology and measuring
technique. Though human error in biological surveys may be considered unavoidable, steps are
being taken to ensure that the process is reviewed so that training techniques will continue to result
in greater precision and accuracy.

All located individuals known to be tagged were successfully scanned, though some changes
between this year's census and previous ones in the location of some of these tags, e.g. a tag in the
left rear hip may this year be recorded as being in the tail. Despite PIT tags and the required reader
being expensive, and tags being known to migrate within the animals (Plummer and Ferner 2012),
the tagging of tortoises with PIT tags is still proving to be an effective method. Three tortoises were
newly tagged this year (numbers 125, 129, and 130).

The first reports of the Curieuse population naturally reproducing were in 1980, two years after they
were relocated to the island (Stoddart et al. 1982), and Hambler (1992) found the number of
emerging hatchlings to be slightly less than Aldabra due to differences in soil acidity. Mating is rare
in the dry season on Aldabra, but on Curieuse the weather is wetter and makes for better mating
conditions all year round (Hambler 1992; Hambler 1994; Lewis et al. 1991; Swingland 1977).
However, the presence of more fresh water is thought to allow populations of predators, primarily
feral rats (Rattus norvegicus) to become more abundant and pose a threat to eggs and hatchlings
(Hambler 1994). Due to the fact that GVI personnel and SNPA staff do not generally see nests being
laid or emerging, it is very difficult to estimate the current rate of reproduction. However, since 29
hatchlings were in the nursery before being stolen, 33 more were found between August and
December 2016, and there are likely to be more that have not been found yet, it can be suggested
that this tortoise population is reproducing well.

Considering this and the fact that 38 years have passed since the relocation, this population should
be increasing, with greater numbers of juveniles and sub-adults. Hambler (1994) estimated that

36



2,100-3,900 hatchlings had been produced on Curieuse by 1993 alone. However, juvenile tortoises
are difficult to locate as they tend to hide in the leaf litter (Grubb 1971; McFarland et al. 1974;
Swingland and Coe 1979), and it is thought they may head for higher ground to avoid predation
(Hambler 1994). However, it would be expected that if this were the reason for not finding them,
they would return to lower ground as sub-adults which does not appear to be occurring. This
suggests that when hatchlings remain free-roaming on the island, mortality is high in the first few
years (Gibson and Hamilton 1984). Curieuse does have a large number of feral rats, which have been
linked to the depletion of giant tortoise populations on other West Indian Ocean islands and in the
Galapagos (Hambler 1992; McFarland et al. 1974; Swingland and Coe 1984). It is currently thought
that Curieuse tortoises only lay one clutch a year (Lewis et al. 1991), and that rat predation could
certainly be having a negative effect (Rainbolt 1996) on hatchlings that are not found and taken to
the safety of the nursery.

The tortoise nursery was established in order to bring hatchlings discovered on the island into a safe
environment until they grow large enough to not be at risk from predation by rats. Additionally, the
nursery allows GVI to take consistent biannual growth data beginning shortly after hatching and
ideally continuing throughout life. The nursery was also intended to protect these valuable tortoises
from the risk of poaching. However, the July 2016 theft of 25 hatchlings was certainly a setback to
both the increase of the population and the growth studies that had been begun with these
individuals. For all hatchlings currently housed in the nursery, the ultimate plan is that with the
recently increased security they will remain guarded and safe, able to be monitored until they reach
the appropriate age when they will be tagged, given an ID number, released, and included in the
census of the free-ranging population.

Conclusion
This study aimed to continue the annual census of the population of Aldabra Giant Tortoises
relocated here to Curieuse. This years dataset included 114 tortoises in the free-ranging population,
the majority of which have remained close to the release site at the Ranger Station. Despite the fact
some individuals were not found this year, looking back at previous censuses it is not unusual for
several tortoises not to be found. Although we know that at least five of the census tortoises have
died, it may be that other missing individuals are alive in more inaccessible areas of the island. It is
advisable therefore that next year, any area that it is possible to access should continue to be
searched to ensure that the census remains thorough.

37



Whilst the population is not growing significantly, there also doesnt currently appear to be a large
population decrease. However, it seems that many of the tortoises originally from Aldabra are no
longer on Curieuse. With a large number of hatchlings being found, reproduction appears healthy,
and this is a promising sign for the future of the tortoise population on Curieuse. Additionally, with
such a large concentration of valuable animals in such close proximity to the larger islands, it is likely
that poaching has contributed greatly to the low population size. Despite setbacks, the protocol of
bringing hatchlings to the nursery improves hatchlings chances of survival.

GVIs annual census of the Aldabra Giant Tortoise population on Curieuse Island over the past four
years is beginning to provide the basis of a good quality long-term study. However, with only four
years of data and considering the life history of this species, it could be many more years before
significant trends in population structure, age at sexual maturity, and growth rates for the tortoises
on Curieuse become clear.

38



Mangroves
Introduction
Seven species of mangrove are present in the Seychelles, of which six were once present on Curieuse
(SNPA 2012) and five currently, along with a mangrove associate species. Mangrove ecosystems play
an important role in ensuring a high level of water quality and clarity, and are essential for adjacent
corals to thrive by trapping sedimentation and land run-off. Mangroves are vital nurseries for fish
and crustaceans, and provide an important habitat for birds, algae, and bryozoans. Mangroves
supply essential nutrients for marine creatures such as fish and shrimp. Additionally, they represent
a crucial buffer zone for protecting inland areas from high wave action and natural disasters such as
tsunamis (Lewis 2005; Yoshihiro et al. 2002).

The mangrove forest on Curieuse is of particular interest. In 1910, a seawall was built at Baie Laraie
in a failed endeavour to rear sea turtles. The wall had a lasting impact on the bay, as it reduced wave
intensity, providing a suitable environment for mangrove seedlings to settle and grow. In December
2004, a tsunami damaged the wall, allowing bigger waves to enter the bay more frequently, causing
an influx of sediment. This is altering the mangrove population structure by decreasing abundance
and species richness (SNPA 2012).

Aims
The primary aim of the surveys is to provide baseline data, which will help facilitate decision-making
regarding the placement of mangrove nurseries in the near future. Previous surveys were developed
in an effort to determine the mangrove distribution pattern in relation to hydrology and salinity, and
current ones aim to assess mangrove growth, mortality, and recruitment rates. Mangrove nurseries
are needed to rehabilitate the mangrove forest on Curieuse, as the forest is thought to be
decreasing in abundance and species richness, and the current surveys are also intended to help
inform rehabilitation efforts.

Methodology
In February 2013, 28 permanent transects were placed in the mangroves and monthly data
collection began in March 2013. The transects were placed 10m apart, spanning the Baie Laraie
mangrove forest in a north-westerly direction perpendicular to the coastline (Figure 10). Along each
transect PVC pipes mark waypoints every 50m. A total of 164 physical waypoints were set up on
permanent transects. Since then, ten waypoints have been uprooted due to weather damage and

39



tortoise trampling. Until now, waypoints have not been replaced once lost, but will be replaced in
2017.

Figure 10. 28 transects, placed 10m apart, span the mangrove forest in a north-westerly direction
perpendicular to the coastline. Numbers in blue show the locations of the five permanent 10m x
10m quadrats.

Annual GBH Survey
At each waypoint, the Girth at Breast Height (GBH; 130 cm from the ground) of one mangrove tree
located within 4m of the waypoint marker was marked and measured. Once a tree was chosen, a
nail was placed 120cm up the trunk and made more visible with a bowtie. Ten centimetres above
the nail, the girth of the tree trunk was measured.

Quadrat Survey
In the past, non-permanent 1m x 1m (2013) and 3m x 3m (2014) quadrats at each selected waypoint
within the mangroves were used; however, there were issues with inconsistent data collection due
to the varying positions of the quadrats. The previous quadrats were also found to be too small to
obtain the data required. As a result, larger permanent quadrats with fixed positions are now being
used. Five 10m x 10m permanent quadrats were set up in June 2015 in various locations throughout
the mangroves. The locations of these quadrats were chosen by SNPA and all lie within the seaward
half of the mangrove forest (Figure 10). The abundance and growth rate of individuals within these
quadrats was measured biannually. Within each 10m x 10m quadrat are four 1m x 1m quadrats
positioned at each corner.

40



The total number of mangrove trees (>1m high; >4cm GBH) and their species were recorded within
the 10m x 10m quadrat. Within each 1m x 1m quadrat, all species of mangrove seedlings (< 1m
high), saplings (>1m high; <4cm GBH, measured beneath the first stem), and trees were counted. All
mangrove trees within the 1m x 1m quadrats also had their GBH measured, which was set at 130cm
from the trunk base during the initial survey in June 2015 or beneath the first stem if the trunk was
less than 130cm. When no seedlings, saplings or trees were present inside of the 1m x 1m quadrat,
the species of roots present were recorded, or a lack of mangroves was noted.

Results
Annual GBH Survey
Of the 164 waypoints assessed during the 2016 GBH survey in April, May and June, 65 had no
species recorded, nine had no data recorded, 89 had data recorded, and one tree had fallen. Of the
waypoints with data recorded, 25 were disregarded due to negative growth and three had unreliable
measurements. In total, 61 trees were included in the analysis consisting of 22 Avicennia marina, 13
Bruguiera gymnorrhiza, nine Lumnitzera racemosa, 16 Rhizophora mucronata, and one Xylocarpus
moluccensis. To show average annual growth rate, trees that were measured in 2013, 2014, 2015
and 2016 have been included in the analysis, and differences in GBH from 2013 to 2016 have been
analysed (Table 9). Bruguiera gymnorrhiza exhibited the fastest mean growth rate per year (2.37cm
2.91 SD) and Avicennia marina exhibited the slowest growth rate per year (1.03cm 1.43 SD).

Table 9. Average annual growth rate (cm) and SD of the five species of mangroves at waypoints
along transects. No SD is provided for X. moluccensis as only one tree was recorded.

Average Annual
Species Number of Trees Growth Rate (cm) Standard Deviation
Avicennia marina 22 1.03 1.43
Bruguiera gymnorrhiza 13 2.37 2.91
Lumnitzera racemosa 9 2.07 2.03
Rhizophora mucronata 16 2.11 1.80
Xylocarpus moluccensis 1 1.37 -


Quadrat Survey
Results from the surveys completed in July 2015, December 2015 and June 2016 have been included
in this report. Data was not collected in December 2016 due to lack of manpower, and results from
January 2017 will be included in the following report, which will reflect an improvement in the
methodology that should make future quadrat data more accurate and comparable.

41



Results from the 10m x 10m quadrats indicate that R. mucronata is the dominant tree species in all
quadrats, except Quadrat 3 where A. marina is most abundant (Figure 11). R. mucronata has
exhibited an increase in abundance between July 2015 and December 2015 in all quadrats where it
is most abundant. However, only Quadrat 5 has shown a continued increase from July 2015 to June
2016. In Quadrat 1, there was a significant decrease in R. mucronata between December 2015 and
June 2016 with a loss of 30 individuals. In Quadrat 2, there was a slight decrease in abundance from
December 2015 to June 2016 yet an overall increase since July 2015. Quadrat 4 exhibited a decrease
in abundance from December 2015 to June 2016, where numbers were very similar to the survey in
July 2015.

120

100

80

60

40 July 2015
December 2015
20
June 2016

0
Avicennia marina

Bruguiera gymnorrhiza

Lumnitzera racemosa

Rhizophora mucronata

Avicennia marina

Bruguiera gymnorrhiza

Rhizophora mucronata

Avicennia marina

Bruguiera gymnorrhiza

Rhizophora mucronata

Bruguiera gymnorrhiza

Rhizophora mucronata

Bruguiera gymnorrhiza

Rhizophora mucronata

1 2 3 4 5
Mangrove species within each quadrat


Figure 11. Mangrove species distribution and abundance throughout the five 10m x 10m quadrats,
for the periods of July 2015, December 2015 and June 2016.

None of the quadrats contained Xylocarpus trees, and only Quadrat 1 contained L. racemosa;
however, L. racemosa had decreased from five to three trees between July and December 2015, and

42



then increased to four trees in June 2016 (Figure 11). The second most abundant tree species
throughout the quadrats in the December 2015 survey, after R. mucronata (n=308), was B.
gymnorrhiza (n=45), followed by A. marina (total n=37). In the most recent survey (June 2016), the
second most abundant species after R. mucronata (n=267) was B. gymnorrhiza (n=144), followed by
A. marina (n=110).

Seedling and sapling data from the 1m x 1m quadrats for the survey months of December 2015 and
June 2016 show a decrease in the total number of seedlings (42 and 37, respectively), and an
increase in the total number of saplings (14 and 17, respectively). The data also indicate that
Quadrat 2 has the most seedlings with an average of 32, which is considerably more than the other
quadrats. Quadrat 2 also contains the most saplings with an average of 15 (Table 10). Throughout all
the quadrats, Quadrat 2 accounts for 85.4% of all seedlings and saplings. Quadrat 3 has no seedlings
or saplings.

Table 10. Number of mangrove seedlings and saplings located within the 1m x 1m quadrats of the
five permanent quadrats, for the periods of December 2015 and June 2016. Quadrat 3 contained no
seedlings or saplings.

Life Dec- %
Quadrat Species Jun-16 Difference
Stage 15 Change
B.
1 gymnorrhiza Seedling 1 1 0 0
R. mucronata Seedling 1 1 0 0
R. mucronata Seedling 9 4 -5 -0.05
R. mucronata Sapling 11 10 -1 -0.01
2 B.
gymnorrhiza Seedling 25 26 1 0.01
B.
gymnorrhiza Sapling 3 6 3 0.03
R. mucronata Seedling 2 2 0 0
4 B.
gymnorrhiza Seedling 1 0 -1 -0.01
5 R. mucronata Seedling 3 3 0 0
R. mucronata Sapling 1 1 0.01
Total # of Seedlings 42 37
Total # of Saplings 14 17

43



Between the two surveys conducted, there has been a slight percentage increase in B. gymnorrhiza
seedlings and saplings (0.04% combined), and a slight percentage decrease in R. mucronata
seedlings and saplings (-0.06% combined) in Quadrat 2.

R. mucronata and B. gymnorrhiza were the only species to have seedlings and/or saplings present in
the 1m x 1m quadrats (Table 11). R. mucronata had the most saplings, with an annual average of 11,
compared to 4.5 for B. gymnorrhiza. However, B. gymnorrhiza had the most seedlings, with an
annual average of 27 compared to 12.5 for R. mucronata. Overall, B. gymnorrhiza had the most
seedlings and saplings, with a combined annual average of 31.5. For the 1m x 1m quadrats that
contained no seedlings, saplings, or trees, the roots for A. Marina, R. Mucronata, and B. gymnorrhiza
were observed. The substrate for quadrats that contained no roots, seedlings, saplings or trees was
recorded as either sand or pond. There is insufficient data at this early stage to be able to
compare the growth rate of the mangrove trees located within the 1m x 1m quadrats.

Table 11. Seedling and sapling abundance based on species, for the periods of December 2015 and
June 2016.

R. mucronata B. gymnorrhiza
Seedlings
December 15 15 27
June 16 10 27

Annual Average 12.5 27

Saplings
December 15 11 3
June 16 11 6
Annual Average 11 4.5
Combined Average 23.5 31.5

Discussion
For the GBH survey, a total of 61 trees were included in the analysis consisting of 22 Avicennia
marina, 13 Bruguiera gymnorrhiza, nine Lumnitzera racemosa, 16 Rhizophora mucronata, and one
Xylocarpus moluccensis. The two species which showed the highest average annual growth rate
between 2013 and 2016 were B. gymnorrhiza and R. mucronata (2.37cm and 2.11cm, respectively).
L. racemosa was the species with the third highest annual growth rate (2.07cm). However, the
sample size was relatively small. Due to the small sample size of trees, once data was separated into
species with reliable measurements, it became evident that more GBH measurements of all species

44



should be taken throughout the mangrove forest in order to gain a greater understanding of species
distribution and growth rates.

Between the 2013 and 2015 GBH survey, eight R. mucronata trees were recorded as dead. This was
to be expected, as this species dominates the coastal fringe of the forest and is therefore more
susceptible to damage from increased wave action and sedimentation. High mortality was also
recorded for this zone by Daig (2013). The 2016 GBH survey did not record any R. mucronata trees as
dead. This is a positive finding, as this species acts as a buffer for the rest of the forest.

The species abundance from the 10m x 10m quadrats were as expected based on the previous 1m x
1m and 3m x 3m surveys carried out in 2013 and 2014. R. mucronata is the most dominant species in
all quadrats, apart from Quadrat 3, where A. marina is the most dominant. Quadrat 3 is the most
landward quadrat, which is likely to be the reason for this difference in species distribution. The
most noticeable increase in the number of R. mucronata trees between December 2015 and June
2016 was recorded in Quadrat 5, which is the closest in proximity to the seaward edge of the forest.
There was also an increase in the abundance of this species between July 2015 and December 2015.
These results may indicate a positive rebound of this species following the increased amount of
wave action due to the partial destruction of the seawall.

The second-most dominant species in the quadrats was B. gymnorrhiza. There was no noticeable
increase in this species compared to R. Mucronata. However, during previous surveys, B.
gymnorrhiza had not been present in Quadrat 2. In the most recent survey (June 2016), one B.
gymnorrhiza has been recorded. This could suggest that this species may increase in the future,
though we can only gain a greater understanding of the distribution and abundance of the species
through future surveys.

With R. mucronata and B. gymnorrhiza being the most abundant species, this is likely to account for
why they were the only species to have seedlings and saplings present throughout the 1m x 1m
quadrats. Overall, there has been a decrease in the total number of seedlings and a slight increase in
the total number of saplings between December 2015 and June 2016. Looking at each species, there
has been a decrease of four R. mucronata seedlings, but the saplings of this species have remained
the same at 11. The seedlings of B. gymnorrhiza have also remained the same at 27; however, there
has been an increase of three saplings between December 2015 and June 2016. This results in a
slight percentage increase of B. gymnorrhiza and a slight percentage decrease of R. mucronata

45



seedlings and saplings in Quadrat 2, which accounted for 85.4% of all seedlings and saplings. These
were all located in the eastern 1m x 1m plot. It is unclear why exactly the eastern 1m x 1m plot
within Quadrat 2 was so productive. It is located on a slightly raised sandy bank surrounded by a
channel that is often inundated; the elevation of this area may offer a substrate for the seedlings to
establish themselves on with less tidal disturbance, and the channel may act as a funnel to direct
propagules towards this location. Furthermore, this concentration of seedlings may be self-
propagating, as more seeds become trapped within the stems of the existing bunched seedlings and
saplings. A high concentration of juvenile mangroves in one particular area increases the risk of high
juvenile mortality rates from threats such as giant tortoise grazing and tree fall. We are unable to
reveal whether there are any seasonal changes in seedling and sapling mortality rates, though this
should become more evident with time.

Moving forward, while the current positioning of the quadrats allows us to collect consistent data on
the mangroves in the seaward half of the forest, they exclude the middle and rear sections of the
forest. As a result of this, species such as Xylocarpus moluccensis, Xylocarpus granatum, L. racemosa
and A. marina are under-represented. The majority of the seaward edge of the forest is also
excluded, which is the area of highest concern as it is where we have anecdotally observed the
highest mortality rates. To undertake future assessments of mortality rates, and understand
whether or not this phenomenon has penetrated further into the forest, it is vital that more
permanent quadrats are set up along the seaward edge of the forest, which will begin in January
2017.

Since the partial destruction of the seawall in the December 2004 tsunami, there have been
concerns that the increased wave action and influx of sediment may be resulting in the degradation
of the forest. Therefore, establishing a mangrove nursery with the aim of rehabilitating the forest
has been a priority for SNPA.

If restorative planting of mangrove habitats is to go ahead it has been recommended that the
removal of stress should be considered prior to attempting restoration (Lewis 2005). There have
been ongoing discussions about whether or not to rebuild the seawall. When considering the
options, it is important to think of the implications that this may have on not only the mangroves,
but also on the multitude of species that inhabit this area, including the neonate sicklefin lemon
sharks that appear to use this area as a nursery ground. One of the options would be to not rebuild
the seawall and allow the mangrove forest to return to the state it was most likely in before the wall

46



was built in 1910. The concern surrounding this option is that it may lead to a decrease in the
currently high level of biodiversity in the area.

Another alternative to rebuilding the seawall would be to create natural buffer zones using R.
mucronata, enhanced sea grass beds, and coral reef restoration. Planting hypocotyls from R.
mucronata using the Rileys Encasement Method (REM) as outlined by SNPA (2012) could create a
natural seawall. REM was developed to facilitate planting where shorelines have high-energy waves
and in an effort to overcome the limitations of other mangrove planting schemes (Johnson and
Herren 2008). Restoring the buffer zone near the wall with R. mucronata, if successful, would restore
the hydrology of the mangrove system, which may allow the forest to naturally rebound. Increasing
the seagrass cover within the Turtle Pond may also assist in reducing the impacts of wave action and
sediment influxes on the mangroves. Studies in Florida have used combinations of eastern oysters
(Crassostrea virginica) and smooth cordgrass (Spartina altinaflora) to diffuse and absorb wave
energy, thus creating less erosion and sediment intake into coastal habitats (Manis 2013), and
perhaps a local seagrass species could be used at Baie Laraie. Conducting coral reef enhancement
directly beyond the seawalls eastern side could also assist in alleviating wave action on the
mangrove forest. With a coral nursery project currently underway on Curieuse Island, there is
potential that the project could include enhancing the reef beyond the seawall in the future.

Conclusion
The current mangrove monitoring activities are part of a long-term project aimed at maintaining and
improving the ecological function of the mangrove habitat. Last year saw the completion of the
salinity, temperature and inundation surveys, which has provided three years of data on which to
base sound decisions for future rehabilitation plans. With the establishment of the five permanent
10m x 10m quadrats within the seaward half of the mangrove forest, together with the future plans
of establishing more permanent quadrats to cover a higher and more varied area of the forest, these
quadrats should provide us with a better insight into species distribution and abundance as well as
seedling and sapling distribution and abundance. The data that has been collected has indicated that
it is important to also establish permanent quadrats throughout the forest in order to represent all
mangrove species present and provide sufficient insight into the changes occurring throughout the
forest. However, focus will be placed on increasing the number of quadrats along the seaward fringe
of the forest to document the rate of change in this most critical area. The annual GBH survey has
shown a continued pattern of dying R. mucronata trees at the seaward edge of the forest until this
year, when a positive change was observed with no dead trees recorded. This could indicate a

47



rebound of this species based on positive annual GBH growth rates and an increase in R. mucronata
trees within the permanent quadrats over the last year. Continued monitoring is required in order to
assess whether the seaward edge of the forest will continue to degrade or whether a natural state of
equilibrium has been reached.

The mangrove forest of Curieuse Island is an integral landscape for multiple faunal communities as
well as neighbouring ecosystems such as the adjacent seagrass beds and coral reefs. Additionally,
the area is heavily visited by tourists and school groups, with most island visitors walking along a
boardwalk through the mangroves where there are educational signs. Many tour guides also stop
their groups in this area to point out flora and fauna of interest. Moving forward, this high
biodiversity area may benefit from the development of natural buffer zones, such as the planting of
R. mucronata to act as a natural seawall, increasing seagrass cover and carrying out coral reef
enhancement to help mitigate the impact of increased wave action and sediment movement since
the partial destruction of the seawall in 2004. Considering the value that mangrove forests provide
in terms of ecosystem services and the potential to improve its state, it is vital that mangrove
monitoring continues in order to better understand, protect, and rehabilitate the area.

48



Sea Turtles
Introduction
Seychelles hosts globally important populations of sea turtles, including one of the five largest
nesting populations of the critically endangered hawksbill turtle (Eretmochelys imbricata) remaining
in the world (Mortimer and Donnelly 2008). Green turtles (Chelonia mydas) also nest in Seychelles,
mostly on Aldabra Atoll and a few in the inner granitic islands. Other sea turtle species that can be
found in Seychelles waters include the leatherback (Dermochelys coriacea), loggerhead (Caretta
caretta) and olive ridley (Lepidochelys olivacea).

The largest hawksbill populations remaining in the Western Indian Ocean occur in Seychelles, where
an estimated average of 1,500 females nested annually in the early 1980s (Mortimer 1984). Since
then, populations have suffered declines due to the nearly complete harvest of nesting females from
the 1960s to the 1990s (Mortimer 1998), following which a total ban on turtle harvesting was
implemented in 1994. An exception to this downward trend was noted at Cousin Island, which has
been well protected since 1970. The Cousin population has seen an eight-fold increase in annual
nesting numbers over the past 20 years (Allen et al. 2010). The exploitation of hawksbill turtles in
Seychelles became particularly intense after the mid-1960s with the advent of the mask and snorkel,
spear guns, underwater lights, outboard engines, and the high prices paid for raw shell (Mortimer
1984). Mortimer (1984) estimated that 4771% of the total estimated annual nesting population in
the granitic Seychelles Islands was killed during the 198082 nesting seasons. Although it is now
illegal to harvest any species of turtle in Seychelles, a small degree of poaching does still occur. In
addition, the destruction of breeding and foraging habitat, especially in the granitic islands, is an
increasingly serious problem (Mortimer 1998).

There are also small numbers of nesting females of the endangered green turtle on Curieuse
(Seminoff 2004, Burt et al. 2015). Green turtles have been heavily exploited for their meat since the
17th century and are now very rare in the Inner Islands (Mortimer 1984), although there is some
evidence to suggest they may have started to recover following the protection of all turtles in
Seychelles in 1994 (see discussion).

The waters surrounding Curieuse are home to both green and hawksbill turtles, as the surrounding
reefs and seagrass beds provide ample food resources. Beaches also provide a nesting habitat for
both species, with Curieuse hosting one of the most important nesting hawksbill populations in the

49



Inner Granitic islands (Burt et al. 2015). This alone is enough to highlight the importance of the
Curieuse Marine National Park for sea turtles. There is also evidence to suggest that the number of
hawksbills nesting on Curieuse has increased by as much as 100% since 1984. It should be noted
however, that this increase is substantially lower than on several other islands that have benefitted
from a much higher level of protection than Curieuse, such as the special reserves of Aride and
Cousin (Burt et al. 2015).

Hawksbill turtles in Seychelles, and along the East African coast, nest primarily during daylight hours
in contrast to hawksbill turtle populations elsewhere, which tend to nest either strictly or primarily
at night (Mortimer and Bresson 1999). Green turtles, on the other hand, nest primarily at night
(Mortimer 1984). Historical data gathered in Seychelles shows that both hawksbill and green turtles
can nest during any month of the year. However, hawksbill turtles show a distinct peak in nesting
from October to February (Mortimer 1998).

Aims
Curieuse is an important sea turtle nesting rookery in the inner granitic islands of Seychelles. Sea
turtle patrols are carried out in an effort to identify the annual nesting female population. There
were few estimates for the annual number of nesting sea turtles on Curieuse before GVI Seychelles
began beach patrols. Another objective of the sea turtle surveys is to measure hatching success rate
on each of the nesting beaches through nest excavations. GVI Seychelles aims to continue to
monitor nesting beaches and expand on current methodology.

Methodology
In the Inner Granitic islands of the Seychelles, the hawksbill sea turtle nesting season is at its height
from October to February, with small numbers of females nesting in the months either side of this
period. Small numbers of green turtles nest all year round with a peak in June to August. Therefore,
beach patrols of the main nesting beaches are conducted four to five days a week from October to
February, with a minimum of weekly checks on all other nesting beaches. Outside of hawksbill
season, all beaches are checked at least once a week so that green turtle nesting is sufficiently
monitored.

Turtle patrols involve walking along the high tide line and recording any sea turtle activities. For all
nesting activity, the date, time, recorder, beach and turtle species were recorded. Track width was

50



measured perpendicular to the direction of the track at its widest point. Estimated time of
emergence was recorded as 0, 1, or 2 where 0 identifies the activity as estimated to have occurred
less than 12 hours prior to encounter, 1 indicates 12-24 hours prior, and 2 indicates over 24 hours
prior. The time of an emergence can be estimated by a) knowing when the last patrol occurred, b)
looking at the clarity of the track in the sand, and c) how much of the track has been washed away
by the tide. Each track is further classified as one of nine emergence types (Table 12). If attempts at
nesting have occurred, the number of attempts is recorded. For all activities, a GPS waypoint is taken
using the code TUN for a nest, and TUA for all other types of activities. For nests where eggs are
located, the location is triangulated and marked with flagging tape, with the distance from each
mark (L, C, and R) recorded. This facilitates nest excavations at a later date, following emergence
of hatchlings.

Table 12. Nine categories of sea turtle emergence types.

A. Wandering (but no digging) below high tide line
Half Moon B. Wandering (but no digging) above high tide line
ESBO. Emergence stopped by obstacle
C. Considerable disturbance, evidence of digging (body pit & egg
chamber) but no covering.
Did Not Lay
D. Evidence of body pitting, but no digging of egg chamber or
covering.
Laid E. Considerable disturbance, evidence of digging and covering.
F. Prob DNL. Probably Did Not Lay
Variations G. Prob Laid. Probably Laid
?. Cannot tell if laid or not

When a nesting turtle was encountered on a beach patrol, expedition members followed
appropriate behaviour in line with training in order to not disturb the turtle. The turtle was observed
until she began laying, at which point she goes into a trance-like state and could be slowly
approached from behind. Depending on number of staff and volunteers, one person may be
assigned to count the number of eggs she laid, using a manual click counter. Once the turtle was at
least halfway through laying, measurements could be taken including two over-the-curve carapace
lengths: mid to tip (M-T) and tip to tip (T-T), and width of the carapace at the widest point, usually
across the third vertebral scute (Figure 12).

51



Figure 12. Measurements taken for each turtle encountered.

Each measurement was taken three times to check for accuracy. Photographs of each side of the
head were taken (without a flash) as well as photos of any distinguishing features. Tag numbers (if
she was tagged), tag scars, evidence of disease/injuries or other distinguishing features were also
recorded. If the turtle was untagged, the field team would wait until she had almost finished
covering her eggs before administering two tags, one in the fleshy part of each of the front flippers,
just before the first scute. Tags administered during both the 2014-2015 and 2015-1016 seasons are
SCA series, while tags administered during the 2016-2017 season have been largely E series. The
location of the eggs was triangulated as above; this could be done while she was laying if enough
people were on hand to assist, or after all other data had been taken. Once an activity had been
recorded, all evidence of tracks, body pits and egg chambers were erased so that it could not be
mistaken for a new activity at a later date.

Hatching Success
Hatching success can be difficult to measure since hatchlings usually emerge at night. However,
success rates can be determined by excavating recently hatched nests. When hatchlings emerge
they leave behind a depression in the sand, the result of the sand sinking down to fill the space the
hatchlings had occupied. Teams monitored each nest around its due date and looked for this
depression.

When teams excavated a nest, they recorded the number of hatched eggs, any pipped (half in, half
out of the egg) hatchlings, live or dead hatchlings in the nest, as well as the number of unhatched
eggs. Unhatched eggs were broken open and recorded as either undeveloped, stage one, stage two
or stage three. Definitions of each excavation category can be found in Table 13. Nest depth was
measured before the contents were replaced and re-buried. Hatching success rate was calculated by

52



dividing total number of hatched eggs by total number of eggs laid. This indicates how many turtles
successfully hatched from their eggs. Additionally, emergence success was calculated by subtracting
the number of hatchlings found in the nest, either dead or alive, and dividing this by the total
number of eggs, indicating how many hatchlings successfully emerged from the nest. Often, a small
number of live hatchlings were found in the nest; these were released onto the beach to enter the
sea otherwise unaided.

Table 13. Nest excavation categories and definitions.

Hatched Empty eggshells.
Live pipped Hatchling has broken through eggshell but not entirely emerged.
Dead pipped As above, though hatchling is no longer living.
Undeveloped No discernible embryo.
Stage one Discernible embryo; eyes, spine, blood development but mostly yolk.
Stage two Partially developed embryo. Yolk sac is larger than the turtle fetus.
Stage three Mostly developed embryo. Turtle embryo is larger than yolk sac.
Predated Egg obviously consumed by crabs or dogs.
Predated Maggot and/ bacteria predation beyond stage recognition.
beyond *When a small amount of maggots, bacteria or fungus was within an egg and the
recognition stage was still recognizable, the number of eggs with predation were accounted for
in [ ].
Example: Stage one: 5 [2]
*5 is the total number of eggs within the stage one category
*2 of those eggs contained maggots, fungus and/or bacteria

Results
This report contains a summary of the 2015 2016 nesting season, since the last annual report was
written in January 2016 before the completion of the season. It also covers the results from the 2016
2017 nesting season until 31st December 2016. The next annual report will contain a summary of
the complete 2016 2017 season.

Nesting Adults Hawksbills
2015 2016 nesting season:
The peak of the nesting season was during December, with 36% of all activities recorded during this
month (Figure 13). The total number of activities for this season was 596, of which 368 were nests,
giving a population estimate of between 92-123 individuals (Table 14). The highest percentage of
laid nests was on Grand Anse with 80%. The lowest percentage of laid nests was on Anse Laraie and
Anse St. Jose with only 1% (Figure 14).

53



Table 14. The total number of activities and nests for hawksbill and green turtles recorded for the
past six nesting seasons, 2010-2016. Data from the 2016-2017 season is complete until 31st Dec
2016.*Population estimate is calculated by dividing total number of nests by a bracketed mean of 3-
4 clutches per female per season for hawksbills, and 3-5 clutches per female per season for greens,
in line with Burt et al. (2015). Numbers in red indicate years where monitoring was not consistent.


2011- 2012- 2013- 2014- 2015- 2016-2017
Nesting Season 2010-2011
2012 2013 2014 2015 2016 (incomplete)

Activities 312 367 522 323 428 596 352
Hawksbill

Total nests 151 186 282 128 225 368 136

* Population
38-50 47-62 71-94 32-43 56-75 92-123 -
estimation
Activities 8 14 9 6 53 47 6
Green

Total nests 0 8 2 4 22 27 3
*Population
1-2 1-2 1-2 1-2 5-7 5-9 -
estimation



250 200

200
150
# Acfvifes

# Acfvifes

150
100 Nest
100
Non-nest

50
50 All Acvies

0 0




Figure 13. The number of activities by month (nest, non-nests and all activities) for the 2015-2016
(left) and 2016-2017 (right) hawksbill seasons. The 2016-2017 season shows activities until 31st
December 2016.

54



90%

80%
Percentage of successfully laid nests

2010-2011

70% 2011-2012

2012-2013
60%
2013-2014

50% 2014-2015

2015-2016
40%
2016-2017
30%

20%

10%

0%

Figure 14. The percentage of total nests laid on each of the seven nesting beaches on Curieuse over
the past seven hawksbill seasons. The 2016-2017 season shows total nests until 31st December 2016.

Beach suitability is measured by assessing nesting success on each beach. The higher the proportion
of successful nesting attempts versus aborted nesting attempts (i.e. non-nests), the higher the beach
suitability. The most suitable beach for nesting was Anse St. Jose, with a nesting success rate of
100% (Figure 15). The least suitable beach was Anse Badamier, with a nesting success of 23%. It
should be noted however, that small sample numbers from Anse St. Jose (n=4), all of which were
nests, are likely skewing the results for this beach. Grand Anse, which showed the highest
percentage of laid nests, had a beach suitability success rate of 70% with 293 nests and 124 non-
nests. This suggests that, with a larger sample size, Grand Anse is a very successful beach for nesting.

55



100 100

Nesfng Success (%)

Nesfng Success (%)

80 80

60 60

40 40

20 20

0 0

Figure 15. Beach suitability for the 2015-2016 (left) and 2016-2017 (right) seasons expressed in
nesting success (number of nests divided by total number of activities) for each beach. The 2016-
2017 season shows nesting success until 31st December 2016.

2016 2017 nesting season:
The total number of activities for hawksbills from September until 31st December 2016 was 352, and
of these, 136 were recorded as nests (Table 14). This nesting season started off slowly compared to
the previous season. In line with the previous season, December exhibited the highest number of
activities (n=174) (Figure 13). Our encounter rate for turtles peaked in December with 36, which was
higher than the previous December, when 22 turtles were encountered. Grand Anse remained the
most utilised nesting beach, with 75% (n=102) of all hawksbill nests to that point in time. The least
utilised beach for nesting was Anse Laraie, with only 1% of nests for the season so far (Figure 14).

The most suitable beach for nesting during the current season until 31st December 2016 was Grand
Anse, showing the highest success rate so far at 47% (Figure 15). This differs to the previous season
when Anse St. Jose and Anse Laraie had a 100% success rate. During the current season, the success
rate at Anse Laraie was 40% and at Anse St. Jose 30%, which were second and fourth highest,
respectively. However, low sample numbers (n=5 and n=10, respectively) are likely skewing the data
again. Once the season is complete, further analysis will help us gain a better understanding of
beach suitability.

The least suitable beach for the current season until 31st December 2016 was Anse Cimitiere, with a
success rate of 14% (Figure 15). There have been a total of 22 activities on Anse Cimitiere so far, with
only three of these activities recorded as nests. The majority (50%) of activities on Anse Cimitiere
were abandoned egg chambers, followed by body pits (23%).

56



Nesting Adults Greens
2015 2016 nesting season:
Green sea turtles lay nests throughout the year in the inner granitic islands, though the low number
of encountered tracks gives a poor indication of the nesting population (Table 14). Green sea turtles
lay at night and infrequently throughout the entire year, making tagging and photo identification on
Curieuse impractical. For the 2015 2016 nesting season, there were a total of 47 activities, of which
27 were nests. Of these 27 nests, 26 were laid on Grande Anse, and one on Anse Papaie. The
population estimate for green turtles was five to nine individuals (Table 14).

2016 2017 nesting season:
The current season has seen a much lower number of green turtles than the previous season. The
first evidence of a green turtle occurred in November, and a total of six activities have been
recorded until 31st December 2016, of which three were recorded as nests (Table 14). Two of these
nests were laid on Grand Anse and one on Badamier. Once the season is completed, further analyses
will be performed.

Hatching Success
2015 2016 nesting season:
A total of 152 hawksbill nests and eight green nests were excavated in the 2015 2016 season
(Table 15). Hatching success was higher for hawksbills (84.7%) than for greens (73.9%). Overall,
hawksbill hatching success (%) ranged from the 70s to the low 90s (where enough excavations were
done to obtain a reliable result): Grand Anse (92.4%, n=114), Anse Papaie (73.6%, n=13), Anse
Mandarin (93.7%, n=13), and Anse Badamier (87.4%, n=4). Three excavations were done on Anse
Caiman (73.0%), two on Anse Laraie (79.9%), and only one on Anse St. Jose (90.8%).

Table 15. Hawksbill and green turtle excavation parameters collected for the 2015-2016 season.

Hawksbill Green
Number of Excavations 152 8
Hatching Success (%) 84.7 73.9
Emerging Success (%) 82.8 72.5
Average Clutch Size 142 97
Average Nest Depth (cm) 50.3 73.6
Total Hatched Eggs 18356 591

57



2016 2017 nesting season:
Since the 2016-2017 hatching season (typically November April) only recently began at the time of
writing, only five successful excavations have been carried out for hawksbill nests (Table 16). This
small sample so far has provided an average hatching success rate of 90.9%. Excavations have been
carried out on Grand Anse (n=2), Anse Papaie (n=1), Anse Badamier (n=1), and Anse Cimitiere (n=1).

Table 16. Hawksbill and green turtle excavation parameters collected for the 2016-2017 season until
31st December 2016.

Hawksbill Green
Number of Excavations 5 -
Hatching Success (%) 90.9 -
Emerging Success (%) 90.0 -
Average Clutch Size 142 -
Average Nest Depth (cm) 53.3 -
Total Hatched Eggs 644 -

The high level of green activities seen on Curieuse during the past few seasons has allowed for
hatching success rates to be calculated. However, this year we have observed a much lower number
of nesting greens (n=3) which has resulted in no marked nests thus far. Therefore, we have been
unable to carry out any excavations of green nests. As the season continues, and with green turtles
nesting throughout the year, we hope to successfully mark and excavate a number of green nests to
report hatching success rates.

Nesting Hawksbill Identification
2015 2016 nesting season:
There were a total of 65 encounters in the previous season. Of these, 32 turtles had tags
administered by us and were recorded as newly tagged turtles. We also encountered 20 turtles that
were already tagged that we recorded as old tags. The remaining 13 encounters included turtles
returning to the sea; therefore it was not possible to check for tags. One turtle was encountered
with old tags that were replaced with new ones. For those turtles where tag numbers were
recorded, the majority were only encountered once, however seven were encountered twice.

2016 2017 nesting season:
So far in the current season, we have encountered 61 hawksbills. Of these, 23 turtles had tags
administered by us and recorded as newly tagged turtles. We also encountered 30 turtles that

58



already had tags; these were recorded as old tags. One turtle we encountered only had one old tag
so we also applied a new tag. The remaining seven included turtles returning to the sea, therefore it
was not possible to check for tags. For those turtles where tag numbers were recorded, the majority
were only encountered once, however four were encountered twice and two were encountered
three times.

No turtles encountered last season have been seen this year (as expected, in line with biennial
reproduction of sea turtles).

Discussion
A slow start in nesting was experienced this season, and we currently have slightly fewer hawksbill
activities and nests compared to the previous season. However, this is expected due to the nature of
biennial nesting in populations of hawksbill turtles. With this in mind, the encounter rate for
hawksbills until 31st December has been much higher in the current season than the previous (n=60
and n=42, respectively). This higher encounter rate will enable us to more closely investigate the
inter-nesting interval of hawksbills (discussion follows).

It was estimated that 71-94 individual hawksbills nested on Curieuse during the 2012-2013 season,
32-43 during the 2013-2014 season, 56-75 during the 2014-2015 season, and the 2015-2016
seasons estimations show the highest so far with 92-123 individuals (Table 14). It is important to
consider the amount of effort spent on recording this data by means of turtle patrols. Over the last
few seasons, more time has been spent on turtle patrols and therefore more data collection has
been possible, which could suggest why our estimations of individuals are higher. Estimations will be
made for the 2016-2017 season once the season is complete. This will aid in our understanding of
population trends.

This is the fourth consecutive season that metal tags have been applied to turtles by staff members.
Tagged individuals encountered this season have once again included tags placed on turtles on other
islands in the Seychelles, indicating the possibility of movement of females between islands within a
nesting season. The above-mentioned population size estimates are based on the assumption that
hawksbills lay an average of 3-4 clutches per season (Burt et al. 2015), and that all of these clutches
were laid on Curieuse. Therefore, if there is indeed movement of nesting females between islands
within a season, then this is an underestimation of the number of females nesting on Curieuse
annually. The continuation of metal flipper tagging and recording of tag numbers will hopefully allow

59



for a greater understanding of hawksbill inter-nesting intervals and changes in nesting sites. This
may also lead to more accurate estimates of number of nesting females, though this requires
collaboration and sharing of data between islands. The photo ID system will continue in order to
compare newly tagged females with previously identified individuals. Once all photo ID individuals
are given metal tags, photo ID will supplement flipper tag numbers as a backup system. It may also
allow for the identification of turtles that are encountered but not tagged (such as those already
leaving the nesting site).

Records indicate that for the 2016-2017 season we have encountered two individual turtles on three
separate occasions. During the 2015-2016 season, the highest number of repeat encounters was
two. With a higher proportion of turtles encountered more frequently, we can gain a better
understanding of their inter-nesting intervals. For the two turtles encountered three times this
season, both were encountered a second time after 43 days followed by a third time after 13 days.
However, we must consider that they may have nested between these days without being
encountered. The average number of days between nesting was 14, which supports the known
inter-nesting behaviour of these sea turtles.

Green turtle activities for the 2014-2015 and 2015-2016 seasons have been remarkably higher than
previously seen on Curieuse. However, for this current season there has been a considerable drop in
green turtle activities (Table 14). Annual fluctuations of over 70 turtles have been recorded on
various islands (Mortimer 2004). However, few green turtles are estimated to nest in the Inner
Granitic Islands. A study from the 2001-2002 and 2002-2003 nesting population on Curieuse
estimated that one to two greens nest on Curieuse annually (Mortimer 2004). Data from 2012-2014
indicates a similar number annually (Burt et al. 2015). However, data from the 2014-2015 and 2015-
2016 seasons suggests a significantly higher number (five to seven and five to nine, respectively) of
nesting green turtles on Curieuse. The current seasons population estimate will be included in the
next annual report when it has been completed. So far however, only six green turtle activities have
been recorded on Curieuse, of which three were nests. With only a few years of year-round, regular
beach surveying, and unknown re-migration intervals (time between nesting periods) for greens in
the Seychelles, it is impossible to draw any conclusions from this with regards to changes in
population size. However, if green turtles in the Inner Granitic Islands are indeed recovering, it is
imperative that nesting females are protected and nesting is monitored consistently.

60



A study of hawksbill hatching success was conducted on Curieuse for the 2001-2002 and 2002-2003
nesting seasons for a selection of nests (n=65). Overall, the hatching success (number of hatched
eggs) was approximately 60% (Mortimer 2004). This differs somewhat from the current
approximation, although excavation categories also differ slightly. The overall hawksbill hatching
success rate of 84.7% for the 2015-2016 nesting season (n=152) seems high when compared to
previous data and other islands. Hatching success for green turtle nests was lower (73.9%). However,
the small sample size (n=8) should be taken in to consideration.

While in the past turtle nests were more evenly distributed across Curieuses beaches, they are now
mostly concentrated on 240m of beach at Grand Anse and Anse Papaie, resulting in an annual
nesting density of 34 clutches per 100m (Burt et al. 2015). In the 2015-2016 season, Grand Anse
continued to be the most utilised nesting beach for hawksbills, followed by Anse Papaie, with 88% of
nests laid on these two beaches. The other beaches could be less utilised for a variety of reasons,
including erosion (Anse Mandarin, Anse Badamier, Anse Cimitiere), high levels of disturbance from
tourists/residents (Anse Laraie, Anse St. Jose, Anse Caiman), and a limited area of plateau behind the
beach (Anse Badamier).

It is imperative that Grand Anse and Anse Papaie remain safe, undisturbed areas for nesting
hawksbills and greens in the Inner Islands. Burt et al. (2015) state a number of recommendations in
line with this, including preventing access by tourists to Anse Papaie and Grand Anse, and installing
educational boards to inform tourists of the appropriate code of conduct if encountering turtles on
those beaches populated by tourists during the day. Also, as recommended, GVI Seychelles staff and
volunteers, with the help of SNPA rangers, will be attempting to clear areas of Grand Anse that are
currently inaccessible to turtles due to fallen trees in the hope that this reduces the number of
unsuccessful nesting attempts due to obstacles. This may also potentially reduce the number of
nests laid below the high tide line and increase hatching success for greens in particular, since
distance from the high-tide line can be correlated with hatching success if the beach is prone to
inundation by storm swells (Mortimer 1990).

Conclusion
Protection at the nesting beaches may be the most critical component of any sea turtle conservation
program (Mortimer 2004). The knowledge that Curieuse Island may be used by up to 123 nesting
hawksbills, and up to nine green turtles annually shows that it is essential to monitor these nesting
populations and maintain high standards of conservation. It is possible that large-scale annual

61



fluctuations occur in the number of females arriving at nest sites (Limpus and Nicholls 1988) and
therefore long-term monitoring is essential to document true population change (Meylan and
Donnelly 1999). Therefore, the existing monitoring schedule of four times a week during peak
hawksbill nesting season, and at least once a week outside of hawksbill season, should be continued
to ensure reliable monitoring of green turtle nesting. The fact that the Curieuse population of
nesting hawksbills has not experienced the degree of recovery witnessed at other more protected
islands stresses how imperative it is that Curieuse Island turtle nesting beaches are not affected by
any future development. Instead, a higher level of protection should be implemented to ensure the
future of Curieuse as a vital hawksbill rookery. With further data collection we will continue to
highlight the value and importance of Curieuse Island for nesting turtles.

62



Lemon Sharks
Introduction
Global shark populations face a number of threats including overfishing and habitat loss. In recent
years there has been an unsustainable decline of most shark species, with annual estimates of
fishing mortality of up to 7.9% of the total global shark population a rate which most species are
unable to sustain (Worm et al. 2013). Significant reductions in shark populations have been shown to
disrupt trophic balance in many ecosystems and cause cascading effects, leading to population
declines in other elements of the food web (Jackson et al. 2001, Myers et al. 2007). Once depleted,
shark populations do not recover easily due to their relatively slow growth rates, late sexual
maturity, low fecundity, and long development periods (Compagno 1990, Cortes 2000). The coastal
habitat preference of many shark species also leads to further challenges related to fishing pressure
and habitat degradation (Holland et al. 1999). In order to effectively conserve, manage, or rebuild
populations of sharks, it is essential to possess a reasonable knowledge of the status of a species and
its reproductive and life history patterns. Therefore, any new information resulting from the
scientific study of sharks can aid in their management and conservation.

The sicklefin lemon shark (Negaprion acutidens; Ruppell 1835) is one of two extant species of lemon
shark. Known regionally as simply the lemon shark, it is a large shark of the family Carcharhinidae
(requiem sharks) and typically grows to a length of approximately 3.0m (Carpenter and Niem 1998).
It is distinguished by the almost equal size of its two dorsal fins, and by the pale yellow colouration
which gives rise to the name lemon shark.

Distribution and Habitat
The sicklefin lemon shark is known to inhabit coastal waters throughout the Indian and southwest
Pacific Oceans (Bester 2014; Figure 16). The range of this species has also recently been found to
extend further to the northeast in the Pacific, with the discovery of several individuals at Palmyra
Atoll in the central Pacific (Papastamatiou 2014), and was historically found in the Persian Gulf
(Moore et al. 2010). The range of this species spans most of the islands throughout the Indian
Ocean, including many islands in Seychelles.

63



Figure 16. Distribution of the sicklefin lemon shark, N. acutidens (Image from IUCN).

N. acutidens is known to inhabit coastal waters up to 92m depth, including coral reefs, shallow
sandy-bottom lagoons, and mangrove swamps (Bester 2014). The high biomass of prey items on
coral reefs and in mangroves provides a food-rich environment, and in the closely related N.
brevirostris, juvenile preference for very shallow waters has been suggested as a predator avoidance
strategy (Morrissey and Gruber 1993b), as large sharks are known to prey on smaller sharks (Harvey
1989).

Conservation Status
The sicklefin lemon shark is one of 58 shark species known to inhabit the waters of Seychelles (Seret
2002). Categorised as vulnerable (IUCN 2014), in part due to its coastal preference and consequent
proximity to human activity, it faces many threats to its continued survival. The last assessment date
of the global conservation status of the species was 2003, and it is currently due for review. The
species is fished throughout its range (Compagno 1990), and its small habitat range and limited
movement patterns make it susceptible to local depletion (Stevens 1984, Stevens et al. 2000).
Schultz et al. (2008) also demonstrated that the seascape of the Indo-Pacific results in particularly
limited dispersal across the range of the species. High fishing pressure in Southeast Asia and
Indonesia has led to locally extinct populations in India and Thailand and to very low numbers in
Indonesia. Other threats to the conservation of N. acutidens consist of habitat loss (coral reefs and
mangroves) a threat of wider global concern (IUCN 2014) and the international trade in shark fin,
whereby captured sharks have their fins removed with the rest of the shark being discarded.

64



Recent efforts have been undertaken to highlight the economic importance of shark species in areas
other than removal of sharks from the population (e.g. ecotourism). Gallagher and Hammerschlag
(2011) comprehensively examined the worldwide distribution, frequency and economic value of
shark-based ecotourism, including four operations in the Seychelles. They discovered that when
sharks were used as the subjects of ecotourism (e.g. shark feeding or viewed by divers), the value of
a live shark to a countrys economy is many times its value as a fishery resource. The sicklefin lemon
shark supports a successful shark feeding ecotourism activity in Moorea Island, French Polynesia
(Clua et al. 2011), where the contribution of each individual shark to the local economy during its
lifespan was calculated to be up to USD 2.64 million.

In 2007, the government of Seychelles produced a National Plan of Action for the Conservation and
Management of Sharks (NPOA; Seychelles Fishing Authority 2007). The plan was updated in 2016
and recognises the nations commitment to, and sets out national strategies for, the conservation of
all shark species in Seychelles waters. The key aim is that shark stocks in the Seychelles EEZ are
effectively conserved and managed so as to enable their optimal long-term sustainable use, and
one of the main mechanisms to achieve that aim is to collect more information on these shark
species. The assessment of the NPOA confirmed that shark stocks in Seychelles have followed a
similar pattern of decline over the past few decades as seen in the majority of shark populations
worldwide.

Current Knowledge of the Sicklefin Lemon Shark
A comprehensive review of the scientific literature on lemon sharks has revealed a significant
difference in the level of knowledge of the two extant species of lemon shark. There are
approximately 200 publications on the closely related Atlantic and eastern Pacific lemon shark (N.
brevirostris) over a wide variety of research areas. In contrast, there are only approximately 35
publications relating to N. acutidens from a very limited number of populations in specific
geographical locations (Moorea Island in French Polynesia, Viti Levu in Fiji, St. Joseph and Aldabra
Atolls in Seychelles, Heron and Lizard Islands, Darwin, Shark Bay, and Ningaloo Reef in Australia), on
a much narrower range of topics. Studies relating to N. acutidens are restricted to a single study on a
single population and single topic in most cases, and research findings related to N. brevirostris
cannot necessarily be assumed to apply to N. acutidens.

Certain behavioural characteristics displayed by the sicklefin lemon shark, such as home range and
site fidelity, make it more susceptible in areas with high human pressure. The home range of N.

65



acutidens has been broadly assessed by Stevens (1984) at Aldabra Atoll, Seychelles, where 91% of
recaptured sharks were found within 2km of their initial tagging location, suggesting a very limited
range of activity. Morrissey and Gruber (1993) measured the home range of juvenile N. brevirostris
at Bimini, Bahamas, and found a similarly small range covering an average area of 0.68km2.

Similar to several published studies on N. brevirostris (e.g. Chapman et al. 2009, DiBattista et al.
2008, Edren and Gruber 2005, Feldheim et al. 2014, Freitas et al. 2006, Morrissey and Gruber 1993,
Wetherbee et al. 2007), N. acutidens has also been shown to display a high degree of site fidelity,
from studies of populations in Fiji (Brunnschweiler et al. 2014), St. Joseph Atoll (Filmalter et al. 2013)
and Aldabra Atoll (Stevens 1984) in Seychelles, French Polynesia (Mourier et al. 2013), Australia
(Speed et al. 2011), and range-wide (Schultz et al. 2008).

Studies on the activity patterns of N. acutidens have been published, and they have been found to
display movement patterns on a diel and tidal basis. Filmalter et al. (2013) suggested that the N.
acutidens of St. Josephs Atoll, Seychelles refuge in cooler waters during the day, contrary to
DiGirolamo et al. (2012), who suggested N. brevirostris in Bimini, Bahamas, seek out areas of higher
temperature during the day. It was also suggested (but not tested) that the same population follows
a similar crepuscular and nocturnal foraging behaviour as N. brevirostris, as shown by DeMarignac
(2000) and Gruber et al. (1988), and that tidal cycles have a strong influence on their movements.

Only a very small number of studies have examined the habitat preference of N. acutidens (Speed et
al. 2011, Stevens 1984, White and Potter 2004), and only one which has specifically assessed the role
of mangroves as habitat (White and Potter 2004). In contrast, there are a multitude of studies which
have confirmed the importance of mangrove habitat as nursery areas for N. brevirostris (e.g.
DeAngelis et al. 2008, Gruber et al. 2001, Morrissey and Gruber 1993, Murchie et al. 2010, Yeiser et
al. 2008), and it is highly likely that mangroves are also an important nursery habitat for N.
acutidens.

The Sicklefin Lemon Sharks of Curieuse Island
Through observations by staff and volunteers from SNPA and GVI Seychelles, it has been known for
many years that juvenile lemon sharks are present in the Curieuse mangrove habitat and Turtle
Pond. There appears to be a clear annual cycle of parturition beginning in September and lasting for
three or four months (similar to observed parturition times on other Indian Ocean islands; Stevens

66



1984), with an influx of many small lemon sharks. Population numbers appear to decline throughout
the year with very few individuals observed between January and August each year.

The current study on the Curieuse N. acutidens population began in October 2014 and is currently in
its third season. This presents an opportunity to improve our overall knowledge of the species in
order to aid in conservation efforts whilst working towards project objectives relating to increased
understanding of the mangrove habitat on Curieuse. With the initial funding provided by Seychelles
British High Commission, GVI Seychelles is working alongside SNPA, our primary partner, to achieve
this objective.

In addition, a part of the funding was directed towards education initiatives. Throughout Seychelles,
public knowledge of the importance of keystone species such as sharks to overall ecosystem health
is often limited. In 2015, fixed information boards were installed alongside the board walk through
the Curieuse mangroves, detailing parameters and aims of the project. Additionally we hosted Green
Islands Foundation, assisting with a shark education fun day for local school children. Finally we have
hosted Seychelles Broadcasting Corporation (SBC) who filmed our survey sessions to produce a
national news article on the project. Ongoing public education about sharks is being conducted at
public events throughout Seychelles by GVI staff and volunteers.

Aims
The project aims to gather life history data for the Curieuse lemon shark population, which allows
for the estimation of population size and establishment of size and growth parameters. In doing so,
this project contributes to the key aim of the Seychelles National Plan of Action for sharks by
increasing knowledge on local shark populations. This knowledge will be used to educate local
people on the importance of sharks and the value of Curieuse Marine National Park.

Methodology
Study Site
The study is being conducted within the Turtle Pond and fringing mangrove forest (Figure 17). The
Turtle Pond is a 40-acre shallow lagoon resulting from the construction of a wall across Baie Laraie in
1910, which was originally intended for the farming of hawksbill turtles (Eretmochelys imbricata);
however, this was unsuccessful and quickly abandoned. The wall, now partially destroyed by the
2004 Indian Ocean Tsunami, has created a unique environment that has allowed the lagoons

67



fringing mangrove forest to flourish into one of the largest and most diverse remaining in the
Seychelles Inner Islands.

Figure 17. N. acutidens study area within the Turtle Pond at Curieuse Marine National Park,
Seychelles.

The lagoon was chosen based on previous studies of nursery areas and site fidelity in N. acutidens
and N. brevirostris, as it was believed that the shallow waters and mangroves would provide a
suitable nursery area for neonates. It is also easily accessible for the transportation of large nets and
other sampling equipment. The seaward edge of the mangrove forest is predominantly comprised of
Rhizophora mucronata, which is inundated up to 1.24m during spring tides (unpublished data: GVI).
Sand flats compose the landward edge of the lagoon, and they are exposed at low tides, while the
seagrass beds located in the central area of the lagoon are only partially exposed at the lowest tides.
There are several deeper sections abutting the wall, with sandy substrate and sporadic patches of
coral. At tides below 0.7m, the southernmost section of the lagoon is isolated, forming a pool
approximately 25x50m, which is referred to as Pats Pool (Figure 17).

68



Capture Methods
Surveys were conducted around dawn (approximately 05:00-08:00) and dusk (approximately 17:00-
19:00). Due to sampling limitations resulting from the heterogeneous nature of the study area,
several methods of capture were used:
1. Seine nets 90 x 0 .75m and 10 x 1.5m (the latter being decommissioned in July 2016), with
a stretched mesh of 10mm. The 90m seine was designed to be deployed in the open waters
of the Turtle Pond and used either as a purse or beach seine, or placed at the mouth of a
drainage channel for the mangroves at low tide (coined the Lemon Shark Highway). The
10m seine was designed for blocking narrow channels and openings.
2. Gill nets 25 x 1.5, 18 x 1.5m, and 10 x 1.5m, with a stretched mesh of 60mm (the 10m net
was decommissioned in August 2016 and replaced with the 25m net). Gill nets were used
under constant observation, either static or dragged slowly in the shallows.
3. Hook and line size 14 barbless circle hooks, with fish used as bait. These were used in the
first season but discontinued in April 2015 due to concerns over the handling stress of
hooked individuals.
4. Cast net 3m in diameter, similar mesh to the gill nets. This method proves most useful in
very shallow water in the Turtle Pond or in restricted areas within the mangroves.
5. Dip nets 60cm diameter, similar mesh to the seine nets. Used either independently or to
safely remove and/or transport sharks from the aforementioned nets to the work-up
station.

Tagging and Data Collection
Upon capture each individual was transported to the work-up station by hand, dip net, or in a water
filled transfer crate, then placed in large water filled holding crate. During the work-up process,
sharks were transferred to a water filled PVC trough with an integrated measuring tape (100mm
diameter October 2014 September 2015; 150mm September 2015 onwards). This method greatly
reduces stress by allowing the sharks to respire in the water whilst tagging and data collection is
occurring. Each new capture was first tagged with both an internal Passive Integrated Transponder
(PIT) tag injected into the musculature beneath the first dorsal fin on the left side of the shark. For
most of the first season, a uniquely numbered external spaghetti tag (Floy Long T-Bar Anchor) with
a phone number for reporting capture by any external parties was applied to the right side of the
shark below the first dorsal fin. Double tagging with both PIT and Floy tags was chosen, as a PIT tag
shedding rate of 12.1 % has been previously recorded in Negaprion sharks (Feldheim et al 2002).

69



Furthermore, it was hoped that the external tags would provide additional information. Floy tagging
was discontinued in March 2015, following concerns over its impact on the sharks.

After tagging, the following measurements were taken to the nearest millimetre: pre-caudal length
(PCL, from the tip of the snout to the caudal pit in front of the caudal fin), fork length (FL, from the
tip of the snout to where the tail begins to fork) and total length (TL). A DNA sample was taken
either using a fin punch from a pectoral fin (October 2014 - September 2015) or later a fin snip was
taken from the trailing edge of the anal fin (September 2015 onwards), as this proved faster and
offered a more permanent indication of prior sampling should PIT tag shedding occur. All samples
were immediately fixed in 100% ethanol. These DNA samples will allow for extensive genetic analysis
of the overall population. Weight was also taken using a sling and hanging scale (accurate to 50g)
before returning the shark to the holding crate. The shark was then overturned to expose the ventral
region to ascertain sex and state of umbilical scar closure (recorded as either: open, open, open,
open, closed (fresh) or closed). Any injuries were recorded and photographed along with the
genital region and umbilical scar state. Care was taken to ensure the mouth and gills were
submerged whenever possible. The shark was then released, and each individual followed for as
long as required to monitor recovery. For recaptured individuals, length, weight, sex, umbilical scar
closure, and injury data was collected using the same protocol as new captures. Additionally, a GPS
position was taken for each individual capture location and recorded alongside the capture method.
Water temperature and salinity were also measured once during each sampling session.

Analysis

Population Estimates:
Estimation of population size was calculated using the POPAN (Schwarz & Arnason 1996) module of
the MARK 8.0 mark-recapture software. This model calculates a super-population N using a Jolly-
Seber calculation (Jolly 1965, Seber 1965) from an input matrix consisting of capture histories of all
individuals marked during the sampling period. This is an open population method of calculating
abundance whereby individuals may enter or leave the study area from the super-population by
emigration, immigration, births, or mortality. A number of conditions common to all Jolly-Seber
models must be met: 1. Every animal present in the population has the same probability of capture.
2. Every marked animal has the same probability of survival until the following sampling time. 3. The
method of marking is permanent and cannot be overlooked. 4. All samples are instantaneous and
each release is made immediately after the sample (Pollock et al. 1990).

70




The input matrix consists of capture histories for each individual in binary form, e.g. 0011010
denoting the individual was first captured on sampling occasion three, and then again on occasions
four and six, after which it was not encountered again. It will generally not be known whether the
individual was not encountered due to not being present at the time of sampling, permanent
emigration from the study area, through mortality, or simply by evading capture. In most Jolly-Seber
formulations sampling mortality can be a major confounding factor, however in contrast, the POPAN
formulation accounts for sampling losses which may otherwise violate assumptions of the model,
denoted by "-1" following the individual's capture history.

Condition Factor:
Condition factor (k) [formula: k= (weight (g) / length (cm)3) x 100)] (Ricker 1975) is an insightful
method of converting length and weight into a single value that can be used to track trends in body
condition. Fish exhibiting a relatively high condition factor value may be indicative of favourable
environmental conditions (e.g. population density, prey availability, habitat quality, etc.), and
changes in condition may be related to changing environmental conditions over time (Blackwell
2000). This formula was used to analyse the condition of year-0 sharks over time.

Results
The results that follow are divided into the 2015/2016 and 2016/2017 research seasons for neonate
N. acutidens. In order to maintain inter-annual consistency, captures of individuals from the
2015/2016 cohort that took place during the 2016/2017 research season were analysed for
recapture growth rates only.

2015 2016 Research Season:
Research Effort
The first captured neonate N. acutidens of the 2015/2016 cohort was observed on 24/09/2015. This
capture marked the starting point for the 2015/2016 research season, which comprised a total of 67
surveys over 364 days, ending with the first captured neonate of the 2016/2017 cohort on
09/29/2016. The total number of surveys in this season was 67.5% higher than the 2014/2015
research season. Sampling effort varied throughout the research season (Figure 18), with an average
of 5.0 (2.2 SD) surveys per month. The highest sampling effort occurred in October 2015 (n=10
surveys), with relatively high levels (6 surveys per month) of research effort between October 2015
and February 2016, with the exception of January (n=3 surveys). Survey effort remained relatively

71



low (4 surveys per month) from March to August. The lowest levels of sampling effort were in
January, May, and August (n=3 surveys/month). Research effort increased to seven surveys per
month in September 2016, in anticipation of the start of the 2016/2017 pupping season.

12

10
# Surveys / Mean capture rate

Figure 18. Total monthly sampling effort in number of surveys (blue bars) and mean N. acutidens
capture rate per survey (red line) over the 2015/2016 research season (24/09/2015 to 22/09/2016).

Mark-Recapture Overview
A total of 142 captures were made over the season, comprised of 95 new captures and 47
recaptures (of 27 individuals). No previously marked individuals were recaptured from the
2014/2015 research season.

Compared to the 2014/2015 research season, the number of total captures in this season was 10.9%
higher, with 1.0% fewer new captures and 46.9% more recaptures. The total number of captures
varied among sampling areas (Figure 19), with the highest number of captures occurring at North
Turtle Pond (n=62), and lowest in the South Turtle Pond (n=14). Over the entire research season, the
mean capture rate was 2.12 (2.67 SD) captures per survey. However, capture rates varied by month
(Figure 18). The mean capture rate was relatively high (4 captures per survey) during a peak
between October and November, with the highest mean capture rate occurring in October (4.9
3.51 SD captures per survey). Mean capture rate declined to relatively low levels (<4 captures per

72



survey) in December, a trend which continued until the end of the research season. The lowest
mean capture rate was experienced in July (0.50 0.58 SD captures per survey).

70
60
Total captures

50
40
30
20
10
0
M NTP PP STP

Figure 19. Total number of N. acutidens captured by sampling location over the 2015/2016 research
season (24/09/2015 to 22/09/2016). M= Mangroves, NTP = North Turtle Pond, PP= Pats Pool, STP =
South Turtle Pond.

The most successful capture method involved the use of gill nets, which accounted for 48.6% of
captures, followed by the seine net, which accounted for 47.2% of captures. The cast net, dip net,
and hand capture methods accounted for 2.8%, 0.7%, and 0.7% of captures, respectively.

The mean water temperature at sampling locations throughout the research season was 29.1C
(1.80 SD), range: 25.5 34.5. Mean salinity was 33.1 (2.8 SD), range: 24.5 35.8. No strong
temporal trends were evident in either temperature or salinity at the time of capture.

Population Estimate
The size of the 2015/2016 N. acutidens cohort within the study site was estimated at 246 (range: 171
398; 95% CI). This population estimate was 21.4% lower than that of the 2014/2015 season (313).

Sex Ratio
A total of 43 males and 52 females were captured, resulting in a sex ratio of 83 males: 100 females.
This differs from the 2014/2015 cohort, which experienced a more male-dominated sex ratio of 126
males: 100 females.

Observed Pupping Season
The first observed neonate of the 2015/2016 cohort was captured on 09/24/2015, marking the
beginning of the observed pupping season. This was determined by the presence of an open

73



umbilical scar on the captured individual. The final individual to be observed with an open umbilical
scar was captured on 08/12/2015, thereby marking the end of the observed pupping season.
Therefore, the 2015/2016 pupping season was an estimated 75 days long.

Size and Weight
All captured individuals were neonates; therefore, all available capture and recapture data was used
to produce the following summary statistics for the 2015/2016 cohort over the entire research
season. With regards to size, mean PCL was 51.5 cm (4.3 SD), range: 45.3 70.9 cm. Mean FL was
44.7 cm (4.8 SD), range: 44.7 77.3 cm. Mean TL was 66.1 cm (5.5 SD), range: 59.0 90.3 cm. On
average, captured females in this population were significantly larger (p<0.001) than males in TL
(females: 66.1 cm 5.5 SD, range: 59 90.3 cm; males: mean 56.6 cm 4.3 SD, range: 51.5 77.3
cm). No significant difference in TL (= 0.05) was detected between the 2014/2015 and 2015/2016
research seasons.

Upon comparing mean TL by month (Figure 20), it appears to steadily increase from September 2015
until January 2016, followed by a fluctuating trend between March and April. Then, a large increase
in the mean TL of captures is evident between May and June, which is followed by a large decrease
in July. From July to September 2016, the mean TL of captures increased drastically until the end of
the research season. However, mean monthly values between May and August were represented by
relatively few individuals (n3). Moreover, variation in TL was very low between September and
December 2015, and increased steadily from January 2016 to the end of the research season.

4.5
98
4
93
3.5
88
3
Weight (kg)

83
TL (cm)

2.5
2 78

1.5 73

1 68
0.5 63
0 58

Figure 20. Mean weight ( SE; blue line) and total length (TL) ( SE; red line) by month for N.
acutidens captures over the 2015/2016 research season (24/09/2015 to 22/09/2016).

74




The mean weight of neonates over this season was 1.65 kg (0.58 SD), range: 1.0 5.35 kg. In
contrast to mean TL values, the mean weight of captures (Figure 20) steadily decreased between
September and December 2015. Mean weight of captures then fluctuated between January and May
2016, which was followed by an increase in June, then a decrease in July. However, mean monthly
values between May and August are represented by relatively few individuals (n3). From July to
September 2016, the mean weight of captures increased until the end of the research season. No
significant difference (= 0.05; t-test) in weight between the two sexes was identified. No significant
difference (= 0.05; t-test) in weight was detected between the 2014/2015 and 2015/2016 research
seasons.

Condition Factor
The mean condition factor of this population over the 2015/2016 research season was 0.89 (0.11
SD), range: 0.63 1.57. Upon comparing mean values by month (Figure 21), a significant (p<0.001; t-
test) decrease in the condition factor of captured individuals was evident between late September
2015 and February 2016. A slight increasing trend was then evident between March and April,
followed by a steady decrease until July. Mean condition factor appeared to increase from July to
September. No significant difference (= 0.05; t-test) in condition factor was detected between the
2014/2015 and 2015/2016 research seasons.

1.2

1
Mean condifon factor (k)

0.8

0.6

0.4

0.2

Figure 21. Mean condition factor (k) by month ( SE) for N. acutidens captures over the 2015/2016
research season.

75



Recaptures and Growth
A total of 47 recaptures were made of 27 individuals, which accounted for 33.1% of total captures
during the season. Recaptured individuals were captured an average of 1.8 times (1.3 SD), range: 1
6. Recaptured individuals were studied for an average of 115.4 days (110.5 SD) at large, range: 6.0
347.0. In order to calculate accurate growth rates, any recaptures that resulted in negative TL
growth (n=11) were removed to avoid including any potential measurement error.

Mean growth in TL per day was calculated at 0.042 cm-1 day-1 (0.035 SD), range: 0 0.114. Mean
annual growth in TL was calculated at 15.35 cm-1 year-1 (12.76 SD), range: 0 41.71. Mean growth
in weight per day was calculated at 0.67 g -1 day-1 (6.56 SD), range: -13.3 +12.09. Mean annual
growth in weight was calculated at 243.2 g-1 year-1 (2394.9 SD), range: -4867 +4415.

One individual (ID# 153) was recaptured a total of seven times, which allowed for the analysis of
individual growth trends over time. Another individual (ID# 190) was recaptured during the
2016/2017 season only, and is included in this section.

Individual ID # 153:
This individual was initially captured on 16/11/2015 and studied for a total of 347 days over six
recaptures. The most recent capture took place during the 2016/2017 research season
(28/10/2016). Throughout the study period, TL increased by 28.9% and weight increased by 137.0%.
One recapture suggested negative TL growth and was subsequently removed to avoid including
human error. TL increased steadily between November 2015 and September 2016 (Figure 22), with a
more drastic increase between September and November 2016. Daily and annual TL growth rates
were 0.051 cm-1 day-1 and 18.72 cm-1 year-1, respectively.

76



4.500 98.0
4.000 93.0
3.500 88.0
3.000
Weight (kg)

83.0

TL (cm)
2.500
78.0
2.000
73.0
1.500
1.000 68.0

0.500 63.0
0.000 58.0

Figure 22. Weight (blue line) and total length (TL; red line) over 347 days at large and 5 captures for
an individual N. acutidens (ID# 153).
Overall, this individual decreased in weight between November 2015 and February 2016 (Figure 22).
A steady increase in weight was observed between February and August 2016, followed by a larger
increase in weight between August and October. Daily and annual weight growth rates were 5.3 g-1
day-1 and 1945.5 g-1 year-1, respectively. Condition factor decreased steadily between November
2015 and February 2016 (Figure 23), and then gradually increased between February and October
2016.

1.2

1
Condifon factor (k)

0.8

0.6

0.4

0.2

Figure 23. Condition factor (k) over 347 days at large and 5 captures for an individual N. acutidens
(ID# 153).

Individual ID # 190:

77



This individual was initially captured on 11/08/2016 and studied for a total of 97 days over two
recaptures. Throughout the study period, TL increased by 10.9% and weight increased by 34.3%.
Daily and annual TL growth rates were 0.09 cm-1 day-1 and 34.24 cm-1 year-1, respectively. Daily and
annual weight growth rates were 11.9 g-1 day-1 and 4327.3 g-1 year-1, respectively. Condition factor
decreased by 4.7% over the study period.

Sampling Mortality and Observed Injuries
No sampling mortality was experienced over the 2015-2016 research season. Injuries were observed
in 23.2% of captured sharks. A total of 22 individuals were identified with injuries, of which 90.9%
(n=20) appeared to be the result of other animals (bites and tears) and 9.1% (n=2) represented other
injuries.

2016 2017 Research Season:
Research Effort
The first N. acutidens neonate capture of the 2016/2017 cohort was on 29/09/2015. This marked the
starting point for captures of the 2016/2017 research season. At the time of writing, this seasons
research activities comprised a total of 20 surveys over 93 days (until 31/12/2016). The total number
of surveys in this season was 23.1% lower (n= 20) than by the same point in the 2015/2016 research
season. Sampling effort varied throughout the research season (Figure 24), with an average of five
(3.56 SD) surveys per month. The highest sampling effort by number of surveys was in October and
November 2016 (n=8 surveys/month). Survey effort remained relatively low for the month of
December due to lack of available manpower.

10
9
# of surveys/Mean capture rate

8
7
6
5
4
3
2
1
0
Sep-16 Oct-16 Nov-16 Dec-16

78



Figure 24. Total monthly sampling effort in number of surveys (blue bars) and mean N. acutidens
capture rate per survey (red line) over the 2016/2017 research season thus far (29/09/2016 to
31/12/16).

Mark-Recapture Overview
A total of 144 captures have been made over this season until 31/12/16, comprised of 113 new
captures and 31 recaptures (of 24 individuals). All newly captured individuals were tagged,
measured, and tissue samples were taken. Two previously marked individuals were recaptured from
the 2015/2016 research season.

Compared to the same point in the 2015/2016 research season, the total number of captures in this
season was 42.6% higher, with 43.0% more new captures and 40.9% more recaptures. The total
number of captures varied among sampling sites (Figure 25), with the highest number of captures
occurring at the Mangroves (n=101), and lowest in the South Turtle Pond (n=11). Over the entire
research season, the mean capture rate was 7.2 (4.17 SD) captures per survey. However, capture
rates varied by month (Figure 24). The mean capture rate was relatively high (5 captures per
survey) during a peak between October and November, with the highest mean capture rate
occurring in November (8.75 4.06 SD captures per survey). Mean capture rate declined to relatively
low levels (<5 captures per survey) in December (4.33 1.53 SD captures per survey).

120

100

80
Total Captures

60

40

20

0
M NTP PP STP

Figure 25. Total number of N. acutidens captured, by sampling location, over the 2016/2017
research season thus far (29/09/2016 to 31/12/16). M= Mangroves, NTP = North Turtle Pond, PP=
Pats Pool, STP = South Turtle Pond.

79



The most successful capture method involved the use of gill nets, which accounted for 83% of
captures, followed by the seine net, which accounted for 9% of captures. The dip net and cast net
methods accounted for 8% and 0% of captures, respectively.

The mean water temperature during sampling sessions throughout the research season was 28.4C
(1.56 SD), range: 25.4 30.5. Mean salinity was 34.4 (.1.0 SD), range: 33.6 36.2. No strong
temporal trends were evident in either temperature or salinity.

Population Estimate
The size of the juvenile N. acutidens population for this season was estimated at 368 (92.12 SE,
range: 242 620; 95% CI) within the study site. This population estimate represents an increase of
49.6% from the 2015/2016 research season.
Sex Ratio
A total of 57 males and 56 females were captured, resulting in a sex ratio of 102 males: 100 females.
This differs from the 2015/2016 cohort, which had a higher proportion of females (83 males: 100
females). However, the 2014/2015 season was also male-biased (126 males: 100 females).

Observed Pupping Season
The first observed neonate of the 2016/2017 cohort was captured on 29/09/2016, marking the
beginning of the observed pupping season. This was determined by the presence of an open
umbilical scar on the captured individual. The final individual to be observed with an open umbilical
scar was incidentally captured on 12/12/16 (not recorded), thereby marking the end of the observed
pupping season. Therefore, the 2016/2017 pupping season was an estimated 74 days long one
day shorter than the estimate for the 2015/2016 pupping season.

Size Trends
All newly captured individuals were neonates, with the exception of two individuals from the
2015/2016 season which were removed from the 2016/2017 analysis. The remaining capture and
recapture data was used to produce the following summary statistics for the 2016/2017 cohort over
the research season thus far. With regards to size, mean PCL was 50.7 cm (2.0 SD), range: 42.3
55.4 cm. Mean FL was 55.9 cm (2.1 SD), range: 51.3 60.8 cm. Mean TL was 64.9 cm (2.5 SD),
range: 58.4 71.4 cm. Upon comparing mean TL by month (Figure 26), the mean TL of captures
appears to steadily increase between September and December 2016, with a low level of variation

80



among individuals. No significant difference (= 0.05; t-test) in TL was detected between sexes or
research seasons.

4.5 98

4 93
3.5
88
3
83
Weight (kg)

TL (cm)
2.5
78
2
73
1.5
68
1

0.5 63

0 58
Sep-16 Oct-16 Nov-16 Dec-16

Figure 26. Mean weight ( SE; blue line) and total length (TL) ( SE; red line) by month for N.
acutidens captures over the 2016/2017 research season thus far (29/09/2016 to 31/12/16).

The mean weight of neonates over this season was 1.53 kg (0.19 SD), range: 1.0 2.0 kg. The mean
weight of captures (Figure 26) increased slightly between September and November 2016, but
decreased slightly in December. No significant difference (= 0.05; t-test) in weight was detected
between the two sexes or research seasons.

Condition Factor
The mean condition factor of this population over this season was 0.87 (0.09 SD), range: 0.56
1.11. Upon comparing mean values by month (Figure 27), condition factor appears to decrease over
time. A significant (p<0.001; t-test) decrease in the condition factor of captured individuals was
detected between October and December 2016. No significant difference (= 0.05; t-test) in
condition factor was detected between the sexes or research seasons.

81



1.2

1
Condifon factor (k)

0.8

0.6

0.4

0.2

0
Sep-16 Oct-16 Nov-16 Dec-16

Figure 27. Mean condition factor (k) by month ( SE) for N. acutidens captures over the 2016/2017
research season thus far (29/09/2016 to 31/12/16).

Recaptures and Growth
A total of 31 recaptures were made of 24 individuals, which accounted for 21.5% of total captures in
this season. Of these recaptures, two individuals from the 2015/2016 research season were
recaptured, and have been analysed with the 2015/2016 data. In order to calculate accurate growth
rates, any recaptures that resulted in negative TL growth (n=10) were removed to avoid including
potential measurement error.

Over the current research season, recaptured individuals were captured an average of 1.19 times
(0.49 SD), range: 1 3. Recaptured individuals were at large for an average of 18.3 days (12.3 SD),
range: 4 52. Mean growth in TL per day was calculated at 0.04 cm-1 day-1 (0.04 SD), range: 0
0.13. Mean annual growth in TL was calculated at 16.17 cm-1 year-1 (15.4 SD), range: 0 46.33.
Mean growth in weight per day was calculated at 0.22 g -1 day-1 (18.22 SD), range: -50.0 +21.42.
Mean annual growth in weight was calculated at 83 g-1 year-1 (6653.6 SD), range: -18250.0
+7821.4.

Sampling Mortality and Observed Injuries
Thus far, sampling mortality for the current season was 0.69% (n=1). Sampling mortality ranged
between 0 3.9% over the two previous seasons. The individual in question (ID# 299) was captured
on 1/12/2016 in the mangroves at 06:19 using a gill net. It was processed and released using the
standard methodology, and exhibited signs of stress (colour blotching) in the final ~30 seconds of

82



the work up process. The shark was then promptly released and monitored for signs of normal
recovery behaviour. It remained on the substrate auto-ventilating for over 30 minutes before it
became deceased. Injuries were observed in 10.6% of captured sharks. A total of 12 individuals were
identified with injuries, of which 91.7% (n=11) appeared to be the result of other animals (bites and
tears) and 8.3% (n=1) represented an intestinal prolapse.

Natural Mortality
A total of 18 juvenile sharks were found stranded and dead in the Baie Laraie area between
19/12/2016 and 5/1/2017. Of these individuals, eight were available to be scanned for tags. The
majority (75%) of those scanned did not contain tags. Those that were tagged (n=2) had been
studied for between 27 and 339 days and grew at a normal rate, and upon examination did not
display any signs of abnormal physical condition.

Discussion
Now in its third research season, the sicklefin lemon shark monitoring programme at Curieuse has
provided a wealth of information pertaining to the life history of this species, including baseline data
regarding pupping season, body size, growth and condition, juvenile population size, and sex ratio.
The continuation of this monitoring programme will be critical in detecting any changes that may
threaten this population of top predators that remain vital to the health and function of Curieuse
Marine National Parks marine ecosystems.

Fishing Effort and Capture Rates
Fluctuations in capture rates have been observed between and within research seasons. The highest
capture rates tended to occur within the first two months of a research season
(October/November), and become reduced to low levels from December onward. This trend is
consistent with juvenile populations of other shark species, which experience a population boom
followed by a marked decrease; it is suggested that the large influx of neonates into an ecosystem at
the start of a pupping season can be supported for a period of time, though natural selection
through predation, competition for resources, and starvation results in large reductions in
population size (Gruber et al. 2001; Lowe 2002; Duncan and Holland 2006; Heupel et al. 2007).
Moreover, as year-0 sharks grow and mature they may also be utilising the Turtle Pond study area
less. Stevens (1984) noted that N. acutidens in Aldabra Atoll moved an average distance of 1.3km
from their initial tagging site, with a maximum of 5km. This may indicate that the lower capture
rates could also be the result of changes in habitat use. In previous research seasons, year-0 N.

83



acutidens captured in the Turtle Pond were tagged externally using plastic spaghetti (Floy Long T-
Bar Anchor) tags. This resulted in several individuals being opportunistically spotted by GVI staff in
shallow coastal areas outside of the Turtle Pond, and possibly off the coast of Praslin by a local
fisherman. A project involving active acoustic tracking of year-0 sharks in Curieuse Marine National
Park is slated to begin mid-February 2017, which should further elucidate the habitat use and
movement patterns of individuals both within and beyond the current study area.

The 2016/2017 research season has already exceeded the two previous seasons in terms of
captures. However, this could be the result of a combination of fluctuations in cohort population size
and variation in fishing effort (i.e. fishing duration and the size, number, and type of fishing gear
used). For example, gill nets were used more often during the 2016/2017 season compared to
previous seasons, while less surveys have been conducted.

Although somewhat problematic as an accurate measure of abundance (Harley et al. 2001), it is
suggested that catch per unit effort (CPUE) should be calculated in subsequent research seasons to
assess capture rates over time, as this would be a more rigorous and standardised method for
avoiding bias due to variation in fishing effort. In order to calculate CPUE, more precise logging of
gear type and size as well as net soak times will be necessary moving forward. Monthly and annual
capture rates based on CPUE would then be directly comparable in order to reliably track temporal
changes in capture rate in this population.

Population Estimate
Estimated population size appears to fluctuate from year to year, with the highest estimated
population being observed in the current cohort (n=368), and the lowest during the 2015/2016
season (n=246). These fluctuations could result from a combination of the reproductive periodicity
of N. acutidens, inter-annual changes in prey availability or habitat quality, and the number of
reproductive females giving birth in the area. Similar inter-annual fluctuations in juvenile population
size commonly occur in other wild shark populations (Bush 2003).

Sex Ratio
Between pupping seasons, sex ratio has remained relatively stable at approximately 1:1, ranging
between 83 and 126 males per 100 females between all research seasons. This is consistent with
other populations of N. acutidens such as the population at Aldabra Atoll, Seychelles, in which 59%
of captured individuals were female (Stevens 1984).

84



Observed Pupping Season
The actual duration of a pupping season is difficult to determine with a high degree of accuracy, as
the duration is currently estimated based on the time between the first and final observations of
open umbilical scars on captured individuals over a given research season. Although a higher level of
research effort is exerted in the weeks leading up to the previously estimated start dates, it is
possible that parturition in this population can begin days or even weeks prior the first individual
bearing an open umbilical scar being captured. Moreover, retention of the open umbilical scar state
varies greatly among individuals (GVI Seychelles, unpublished data). Therefore, the duration of
observed pupping seasons provided here serve as a rough estimate of an actual pupping season.
However, long-term research over future seasons should aid in determining the extent of the
pupping season for N. acutidens in this population with increased precision. To date, the earliest
estimated start date is the 24th of September (2015), and the latest estimated end date is the 12th of
December (2016), with a duration of approximately 75 days.

Size, Weight and Condition Trends
Based on available data from the previous and current research seasons, the consistent mean
length, weight, and condition factor values among years suggests stability in these variables.
However, long-term research must continue to be conducted in order to support this notion with
any degree of certainty. Completion of the 2016/2017 research season will result in data collection
over two complete pupping seasons that can be directly compared for the first time in this
population (as the 2014/2015 research season began shortly after the observed pupping season
range).

The mean length (TL) of female sharks was significantly larger in the 2015/2016 cohort, a pattern
that is consistent with the N. acutidens population at Aldabra Atoll, Seychelles (Stevens 1984).
Overall, TL of analysed individuals increased steadily throughout all research seasons, with relatively
little variation among individuals between September and February. From February onward,
variation in length among individuals was relatively high. In contrast, the mean weight of the
population generally decreased or remained stable between September and December, though also
exhibited growth (with increased variation) throughout the remainder of research season. This
increase in variation over the course of the research season was likely due to a combination of the
lower number of captures (monthly sample size) and increased variation in the length of individuals
as they mature, which has also been observed in N. brevirostris (Barker et al. 2005). For example,
this trend is particularly evident in capture data from the months of June and July 2016, which were

85



characterised by low capture numbers (n3) and large variation in size and weight among the
individuals captured (Figure 21). Increased fishing effort (e.g. larger nets and more sessions) may
result in a greater monthly sample size, which could provide more accurate data regarding growth in
size and weight over time.

Mean condition factor over all research seasons remained stable. However, it is evident that the
mean condition factor of neonate sharks decreased significantly over the first three to five months
following parturition (September February). This is the result of mean TL increasing over time with
a reduction in mean weight. This reduction in condition is likely due to intense competition over this
period. Mean body condition fluctuated between March and July, then appeared to increase until
September.

The trend of decreasing weight and body condition over the first 3 5 months of the research
season is consistent with the hypothesis of intense competition among congeners within a nursery
habitat causing deaths due to starvation, e.g. as suggested by Lowe (2002) for neonate scalloped
hammerhead sharks (Sphyrna lewini). Moreover, a neonate lemon shark population studied in
Bimini, Bahamas was reported to have an estimated 35 to 62% neonate mortality in the first year,
though largely due to predation (Gruber et al. 2001). It has been suggested that aside from
predation, such mortality rates could also imply that nursery areas may not always provide sufficient
resources (Heupel et al. 2007). However, predation should not be ruled out as a cause for reduced
capture rates later in the season, as between 10.6 23.2% of captured individuals showed injuries
that may indicate predation.

Growth Trends
The TL growth rates reported here (15.4 16.2 cm-1 year-1) are slightly higher than those of other
studies of N. acutidens populations (12.5 to 15.5 cm-1 year-1) (Stevens 1984). However, the growth
rates from Stevens (1984) were based on data from young of the year and older sharks, and are not
directly comparable. Thus, the elevated growth rates provided here are strictly representative of
young of the year individuals. Although negative growth has previously been recorded in some shark
studies (Pratt and Casey 1983; Meyer et al. 2015), individuals exhibiting negative growth (reduced
TL) were removed from analysis pending further investigation into the likely cause of apparent
negative growth.

86




Sampling Mortality and Injuries
With improved capture and handling techniques, sampling mortality has been reduced to levels
much lower than other studies of lemon sharks. For example, Gruber (2001) experienced 0 11.1%
mortality in a study of lemon sharks (N. brevirostris) in Bimini, Bahamas. We reported 0% and 0.69%
mortality over the 2015/2016 and 2016/2017 research seasons, respectively. This low level of
sampling mortality is likely the product of the continual review and optimisation of handling and
research procedures, which are conducted in order to keep sampling mortality as low as possible.

Natural Mortality
Without further specific research into potential factors leading up to the observed natural mortality
that resulted in 18 juvenile N. acutidens being found dead between 19/12/2016 and 5/1/2017, it is
difficult to determine the exact cause. However, the elevated capture rates and population estimate
in the 2016/2017 research season indicate that a relatively high number of neonate lemon sharks
were present in the ecosystem. Also, the majority of dead, stranded young of the year sharks found
by SNPA rangers and GVI staff anecdotally appeared relatively thin for their body length. Although it
is difficult to determine the cause of this mortality without a more specific study, these trends
appear to be consistent with the hypothesis of intense competition among congeners within a
nursery habitat causing deaths due to starvation. It is possible that the observed seasonality in
juvenile shark mortality at Curieuse could potentially be linked to the natural reduction of available
prey present in the form of baitfish, which appear seasonally at the end of the Southeast monsoon
and are preyed upon heavily by many piscivorous species throughout subsequent months - leading
to a reduction in available prey for juvenile lemon sharks. However, this theory cannot be tested
without further research, and other factors should not be excluded in contributing to this mortality
event. Therefore, this population must and will continue to be monitored intensively to determine
population trends. A protocol for stranded shark encounters should be developed and enacted in
order to facilitate data collection with regard to future mortalities. This should include increased
communication between SNPA, tour guides, and GVI so that deceased sharks can be scanned and
measured as soon as possible following discovery. Necropsies and the collection of tissue samples
should be performed, and samples should be sent to relevant experts to test for illness and disease.

87




Conclusion
The lemon shark monitoring programme, now in its third research season, has provided a robust set
of standardised and comparable baseline data regarding population parameters such as pupping
season, body size, growth and condition, juvenile population size, and sex ratio. Thus far, the
juvenile N. acutidens population monitored within the Turtle Pond area of Curieuse Marine National
Park appears to be stable in both size and condition, while exhibiting a higher growth rate than the
population at Aldabra Atoll. This baseline data will be used to compare against future trends in order
to continually inform park management actions regarding this important species. As such, the
following recommendations are made based on the available data:

1) Increase daytime and night patrols aimed at preventing illegal fishing within the marine protected
area in Curieuse Marine National Park during and following the observed sicklefin lemon shark
pupping season (September to December):

The ability to enact this recommendation would result in direct increased protection for breeding
adult sicklefin lemon sharks and their offspring, with additional benefits to other species targeted by
illegal fishing within the protected area. A project involving the active acoustic tracking of sharks is
slated to begin in mid-February 2017, and will collect data on shark movements and critical habitat
throughout the protected area. This additional data should also help inform park management on
critical habitats in need of special attention and monitoring, and result in healthier marine habitats
that would improve both local fisheries and tourism potential.

2) Establish a protocol to monitor stranding and mortality events for sicklefin lemon sharks. The
mortality event experienced during the course of this research season highlighted the need for a
standardised protocol to collect information to inform park management of the specific factors
leading to natural mortality:

An agreed upon protocol among SNPA, GVI and local tour guides should be developed and enacted
prior to the next research season, which would involve calling GVI personnel immediately upon
encountering a stranded shark so that they can be scanned and measured. Necropsies and the
collection of tissue samples should also be performed, and samples should be sent to relevant
experts to test for illness and disease. Enacting these recommendations would facilitate a greater

88



understanding of future mortality events and thereby improve management of this species within
the park.

89




Beach Profiling
Introduction
Curieuse Island is located on the Seychelles Bank, where coastal plateaus are comprised of
calcareous sand accumulated from fringing reefs surrounding the granitic islands (Nentwig et al.
2014). Throughout the year, Curieuse is subjected to changes in wind and wave direction. During the
Southeast Monsoon, wave energy comes from the southeast between May and September, after
which it then switches to the northwest between November and March when the Northwest
Monsoon sets in (Payet & Agricole, 2006). Between the monsoon periods, there are several weeks
where wind direction fluctuates and the sea tends to be calmer.

Since GVI became established on Curieuse in 2009, substantial seasonal morphological changes to
the beaches have been observed. However, until 2015 there was no continuous quantifiable data
collection on this. In the past, Seychelles has been impacted by significant storm events such as the
2004 Tsunami (Ramalanjaona, 2011) and Tropical Cyclone Felleng in 2013 (Leister, 2013). The
collection of baseline data would therefore be vital in our ability to measure the impact that any
future storm events or changes in sea level may have on Curieuses beaches.

The beaches of Curieuse Island also provide important nesting habitats for the critically endangered
hawksbill turtle and the endangered green sea turtle (IUCN 2017). Having an understanding of the
changing morphology of these nesting beaches, particularly during peak hawksbill nesting season
from October to February (Mortimer, 1998) will guide SNPA in future decision-making regarding
habitat protection measures.

Aims
The primary aim of this survey was to monitor the changes of six beaches on Curieuse Island by
recording their beach profiles over time. Monitoring these changes allows us to assess the rate of
erosion and accretion along the beaches and investigate how this corresponds with changes in wind
and wave direction between the two monsoon seasons. In turn, this should provide a greater
understanding of any net loss or gain of beach sediment annually, as well as on a long-term time
scale in years to come.

90



Methodology
A total of 18 transects were surveyed on six of the beaches on Curieuse: Anse Caiman (2), Anse
Cimitiere (1), Anse St. Jose (6), Anse Laraie (4), Anse Papaie (2), and Grand Anse (3). Note that there
were originally seven transects on Anse St. Jose, but one has been removed for practical safety
reasons. The number of transects installed on each beach was dependent on the beach length, with
longer beaches having more transects. The positions of transects were chosen by SNPA, and
currently only beaches located along the eastern, southern, and western coastline were being
surveyed due to time and resource constraints (Figure 28).

Figure 28. Map of Curieuse Island and the approximate positions of the beach profiling transects
along Anse Caiman, Anse Cimitiere, Anse St. Jose, Anse Laraie, Anse Papaie and Grand Anse in 2016
(note: AJ07 is no longer being monitored within this project).

Individual transects were surveyed once every two months, two hours either side of the lowest low
tide of the month, as surveying at low tide usually permitted the access required to the offshore
step. The beaches were separated into two groups with Anse Caiman, Anse Cimitiere, and Anse St.
Jose all being surveyed in the same month, and Anse Laraie, Anse Papaie, and Grand Anse all
surveyed the following month. Each transect had a fixed reference mark painted on a tree, rock, or
building from which the profile was started, and each reference mark had its location recorded by
GPS. It is important that the trajectory of transects from the reference mark towards the sea is in the

91



same direction each time it is done, which was known from a fixed compass bearing to minimise
variations between profiles each month.

Each transect was surveyed by following a set method each time it was completed. Firstly, the height
of the reference mark was measured from the ground to the top of the spray-painted line of the
mark (note that all measurements were recorded in metres and measured to the nearest
centimetre). Using a compass and the fixed bearing given for that profile, the transect trajectory was
then established. The transect was then surveyed in segments using an Abney Level and two ranging
poles from the reference mark down to the sea. One pole was initially placed by the reference mark
with the second pole placed where the terrain changes in slope angle. Then, one person (this should
be the same person for an entire transect to ensure consistency) used the Abney level to measure
the angle of the slope in degrees and minutes. The Abney level should be held at a comfortable
height level at one of several pre-prepared marks on the ranging pole and read by using the
corresponding mark on the other pole. The length of the segment was also measured, ensuring this
was done once again between two corresponding points on the poles to ensure accuracy, and no
segment measured was greater than 10m in length. After this the first pole was then moved past the
second to be placed at the next point of slope change for the next segment. Once the segment
including offshore step had been recorded one further segment after this was measured to
complete the profile. Once each transect was complete, a photograph was taken of the entire profile
(perpendicular to the beach). For each segment surveyed, the angle, horizontal distance, and any
obstacles/substrates of interest were recorded (e.g. rocks, logs), and a sketch of the beach profile
noting the rough positions of the ranging poles was also drawn. If a segment was very short and the
Abney Level was unable to be used, then an inclinometer may have been used instead, and the angle
converted to degrees and minutes.

All data was entered into the Beach Profile Analysis (Version 3.2) software, which was used to
produce profile graphs and provide beach width (m) and area (m) measurements. All metadata (e.g.
dates, times, survey teams, and comments) was recorded in a separate Excel spreadsheet.

Results
For the majority of the beach profiling transects in this coastal monitoring project, six months of
data has been gathered over the course of 2016 (i.e. each beach surveyed every other month). This
year, the decision was made to take transect AJ07 out of the survey due to it being physically
inaccessible at certain times of the year, meaning very limited data could be gathered for this point.

92



Due to recorder error and transects being unable to be repeated, two transects on Anse St. Jose
(AJ01 & AJ02) and the one on Anse Cimitiere (ACIM01) have data missing from July, but have been
included for the remaining months. All remaining results allow for an overall view of what represents
the first full year of beach monitoring of the beaches of Curieuse. The two transects along Anse
Caiman and one transect along Anse Cimitiere were excluded from the 2015 analysis due to them
being set up in October and therefore having only one month of data. In this report, when all data
gathered over both years is being considered, data for AJ01, AJ02 and ACIM01 include October 2015
to November 2016 while all others include between June/July 2015 to November/December 2016. It
should also be noted that the software Beach Profile Analysis (version 3.2) has extrapolated the data
for some profiles, which could lead to some inaccuracies.

120
January

100 March
May
80 July
Width (m)

September
60 November

40

20

0
AJ01 AJ02 AJ03 AJ04 AJ05 AJ06
Survey Transect

Figure 29. The beach width (m) of Anse St. Jose in 2016 for each transect in the months of January,
March, May, July, September, and November (note: data for July 2016 for AJ01 and AJ02 is not
available). Transect AJ01 is at the north-western end of the beach and AJ06 at the south-eastern
end.

93



400 January
350
March
300
May
250
Area (m)

July
200

150

100

50

0
AJ01 AJ02 AJ03 AJ04 AJ05 AJ06
Survey Transect

Figure 30. The beach area (m) of Anse St. Jose in 2016 for each transect in the months of January,
March, May, July, September, and November (note: data for July 2016 for AJ01 and AJ02 is not
available). Transect AJ01 is at the north-western end of the beach and AJ06 at the south-eastern
end.

In general, beach width and area (Figures 29 and 30) for Anse St. Jose was initially reduced, and then
increased throughout the year; however there were fluctuations and exceptions to this. The general
trend appears to be that transects AJ01, AJ02, AJ03, AJ04, AJ05 lost width and area from January to
May, with a more steady increase throughout the remainder of the year. The further south-east the
beach spans, the wider and larger in area it gets (Figures 31, 32, and 33). The greatest losses and
increases were observed in transect AJ06, which displayed slightly differing results from the rest of
the beach, with an increase in width between January and July at an average rate of increase of
4.60m/month and an average rate of loss in width between July and November of 14.98m/month.
With regards to area, transect AJ06 fluctuated somewhat, displaying an increase in area between
January and March (average accretion rate of 28m/month), after which erosion took place until
May (at a rate of 34.41m/month), and from May to July it began accreting sediment again at
52.42m/month. After July there was a rather high rate of erosion seen at AJ06 up until November of
57.46m/month.

94



Figure 31. The beach profile for transect AJ04 on Anse St. Jose, Curieuse Island, for January, March,
May, July, September, and November 2016. The profile was generated using the program Beach
Profile Analyses (version 3.2). (The horizontal axis depicts the distance (m) and the vertical axis
shows depth (m)).

Figure 32. The beach profile for transect AJ05 on Anse St. Jose, Curieuse Island, for January, March,
May, July, September, and November 2016. The profile was generated using the program Beach
Profile Analyses (version 3.2). (The horizontal axis depicts the distance (m) and the vertical axis
shows depth (m)).

95



Figure 33. The beach profile for transect AJ06 on Anse St. Jose, Curieuse Island, for January, March,
May, July, September, and November 2016. The profile was generated using the program Beach
Profile Analyses (version 3.2). (The horizontal axis depicts the distance (m) and the vertical axis
shows depth (m)).

Table 17. The mean width for each beach transect calculated from all data over 2015 and 2016.
Data includes average width (m), standard deviation, minimum and maximum width measurements
and the difference in these measurements.

Transect Name Average SD Minimum Maximum Min-Max Difference
AC01 41.90 49.02 15.26 29.24 13.99
AC02 20.07 4.96 11.99 25.70 13.71
ACIM01 22.48 3.58 16.86 25.17 8.31
AJ01 25.92 5.89 17.36 33.75 16.39
AJ02 18.25 3.92 12.70 24.22 11.51
AJ03 27.33 2.91 24.10 32.81 8.71
AJ04 27.27 7.96 19.33 63.38 44.04
AJ05 44.20 18.31 21.01 71.26 50.25
AJ06 93.33 22.03 53.53 129.41 75.89
AL01 40.76 7.84 25.78 50.76 24.98
AL02 41.38 10.87 30.53 67.26 36.73
AL03 36.68 3.04 32.62 40.35 7.74
AL04 32.80 2.79 28.98 36.68 7.69
AP01 24.85 10.42 16.12 52.77 36.65
AP02 34.92 12.57 23.08 67.99 44.91
GA01 27.98 5.11 21.55 36.15 14.60
GA02 34.46 3.67 29.64 40.95 11.31
GA03 47.60 17.50 23.42 73.76 50.34

96



Table 18. The average statistics regarding area for each beach transect calculated using all data
gathered in 2015 and 2016. Data includes average area (m2), standard deviation, minimum and
maximum area measurements and the difference in these measurements.

Transect Name Average SD Minimum Maximum Min-Max Difference
AC01 25.92 9.57 13.80 38.40 24.60
AC02 21.59 9.51 11.04 31.69 20.66
ACIM01 23.12 6.13 15.43 32.33 16.89
AJ01 31.43 12.17 18.48 55.66 37.18
AJ02 20.82 6.52 12.30 29.23 16.93
AJ03 27.33 8.03 30.53 58.38 27.86
AJ04 37.81 15.40 18.35 41.79 23.44
AJ05 75.16 37.82 29.49 129.33 99.84
AJ06 205.91 86.38 107.27 337.99 230.72
AL01 59.42 13.43 23.44 68.86 45.42
AL02 43.50 4.49 37.22 49.60 12.38
AL03 52.25 7.41 36.92 60.28 23.36
AL04 44.31 3.88 38.50 51.59 13.09
AP01 25.96 10.42 8.67 40.08 31.42
AP02 50.55 5.42 41.69 56.16 14.47
GA01 46.40 14.50 29.67 73.25 43.58
GA02 57.92 4.66 51.22 64.89 13.67
GA03 80.20 29.77 27.75 120.24 92.49

Averages have been calculated for all 18 beach transects in width (Table 17) and area (Table 18)
using all the available data from 2015 and 2016. This data demonstrates averages and the extremes
of all measurements recorded for the beaches being monitored. It can clearly be seen that AJ02 has
the lowest average width and area of any transect and the lowest average beach width over this
period. Interestingly, transect AJ06 exhibits the highest averages in area and width, with the greatest
increase at 75.89m between its shortest ever recorded width to its widest, and a difference of
230.72m from its lowest recorded area to its largest.

97



35
January
30
March
25 May
Width (m)

20 July
15 September

10 November

0
AC01 AC02 ACIM01
Survey Transects

Figure 34. The beach width (m) of Anse Caiman and Anse Cimitiere in 2016 for each transect in the
months of January, March, May, July, September, and November (note: data for Anse Cimitiere in
July 2016 is not available). Transect AC01 is at the north-western end of the beach, with ACIM01
being the furthest to the south-east.

45
January
40
March
35
30 May
Area (m)

25 July
20 September
15 November
10
5
0
AC01 AC02 ACIM01
Survey Transects

Figure 35. The beach area (m) of Anse Caiman and Anse Cimitiere in 2016 for each transect in the
months of January, March, May, July, September, and November (note: data for Anse Cimitiere in
July 2016 is not available). Transect AC01 is at the north-western end of the beach, with ACIM01
being the furthest to the south-east.

Anse Caiman and Anse Cimitiere have been represented together in Figures 34 and 35, as possible
trends can be better considered this way with the two beaches being in such close proximity to each
other. All three transects showed a reduction in width between January and March. However, the
north-westernmost transect (AC01) exhibited loss into May in both width (an average rate of

98



decrease of 1.93m/month) and area, with an average erosion rate of 3.35m/month. From July to
September, this same transect increased 20.75m in area and 13.63m in width. The next transect to
the south-east (AC02), demonstrated losses between May and July (19.33m area, 10.60m width)
and a noticeable gain between September and November (20.66m in area, 11.54m in width). In
September and November, all three of these transects generally exhibited an increase in area, and
two transects increase in width (AC02 & ACIM01). In January, May, and November transects across
Anse Cimitiere exhibited slightly longer transects (all approximately 25m) and a higher area of
sediment (23.36m, 27.87 m and 32.33m), with March and September showing slightly shorter
transects with less accumulation. It is worth noting that there are many trees growing on the beach
between Anse Caiman and Anse Cimitiere which may trap sand or affect sediment accumulation to
some extent.

80
February
70
April
60
June
50
Width (m)

August
40
October
30
20 December

10
0
GA01 GA02 GA03
Survey Transects

Figure 36. The beach width (m) of Grand Anse in 2016 for each transect in the months of February,
April, June, August, October, and December. Transect GA01 is at the south-western side of the
beach and GA03 at the north-eastern side.

99



140 February
120
April
100
June
Area (m)

80
August
60
October
40

20

0
GA01 GA02 GA03
Survey Transects

Figure 37. The beach area (m) of Grand Anse in 2016 for each transect in the months of February,
April, June, August, October, and December. Transect GA01 is at the south-western side of the
beach and GA03 at the north-eastern side.


The results for Grand Anse (Figures 36 & 37) reveal that between February and June there appeared
to be some movement of sediment away from transect GA01 on the south-easternmost side of the
beach. There was both a decrease in width (average rate of loss 3.85m/month) and area (average
rate of erosion of 10.90m/month). In August and October, the area and width increased once more
and fell again in December. In general, a similar though less distinct pattern was observed for
transect GA02, with only small fluctuations throughout the year. The main areas of note regarding
this beach involve transect GA03, which showed a slight loss in width between February and April
before increasing again in August (average rate of increase 5.7m/month). There was a 21.77m
increase in width between April and August before a dramatic loss by October of 41.86m in width of
82.53m in area during this time. With regards to beach area between February and August, a steady
increase in area is observed at an average accumulation rate of 5.4m/month. It is important to note
that there are some practical issues with transect GA03, as it has a large amount of rock in
approximately two thirds of the profile. Therefore, it is very difficult to obtain accurate readings
from the level of the sediment. As such, any results from this transect should be treated with
caution.

100



60
February

50 April

40 June
Width (m)

30 August

20 October

10

0
AL01 AL02 AL03 AL04
Survey Transects


Figure 38. The beach width (m) of Anse Laraie in 2016 for each transect in the months of February,
April, June, August, October, and December. Transect AL01 is at the south-western edge of the
beach and AL04 at the north-eastern side.

80
February
70
April
60
June
50
Area (m)

August
40
October
30
December
20
10
0
AL01 AL02 AL03 AL04
Survey Transects

Figure 39. The beach area (m) of Anse Laraie in 2016 for each transect in the months of February,
April, June, August, October, and December. Transect AL01 is at the south-western edge of the
beach and AL04 at the north-eastern side.

Anse Laraie exhibited some fluctuations (Figures 38 & 39) in width and area, exhibiting small
amounts of accretion and loss, though remaining in a fairly stable state. The greatest changes in
2016 occurred on the two most westerly transects. Transect AL01 observed a loss of 34.66m of area
and 12.01m in width between April and June, with another sudden increase in the following two
months, with an accretion of 44.21m and increase in width of 21.44m. Transect AL02 also saw

101



fluctuations in area and width, though they were not as large. In general, transects AL03 and AL04
increased in width and area until April, and then began to decrease until October when they
increased once more (the exception being AL04, which began to increase in width very slightly in
August). In December, both transects began to lose width and area again. It is to be noted that
transects AL01 & AL02 are separated from AL03 & AL04 by a rocky headland.

60
February
50
April
40 June
Width (m)

30 August

October
20
December
10

0
AP01 AP02
Survey Transects


Figure 40. The beach width (m) of Anse Papaie in 2016 for each transect in the months of February,
April, June, August, October, and December. Transect AP01 is at the south-western edge of the
beach and AP04 at the north-eastern side.

60
February
50
April
40
Area (m)

June
30
August
20

10

0
AP01 AP02
Survey Transects


Figure 41. The beach area (m) of Anse Papaie in 2016 for each transect in the months of February,
April, June, August, October, and December. Transect AP01 is at the south-western edge of the
beach and AP04 at the north-eastern side.

102



Both transects on Anse Papaie are shown in Figures 40 and 41. Transect AP01 exhibited an increase
in both width (average increase rate 18.33m/month) and area (average accretion rate of
15.71m/month) between February and April, which was reduced again by June (average decrease
of 36.07m/month and erosion rate of 8.81m/month). Following this, both width and area in AP01
increased until October, when area began to fall slightly. Transect AP02 had area measurements that
generally remained stable with the exception of two periods of erosion in June and October, with a
similar pattern for width.

Calculations regarding annual net losses and gains have not been performed due to insufficient data.
It is important to note that net losses and gains to beach sediment do not denote sediment budget
changes (please refer to the discussion section for further explanation).

Discussion
This beach profiling project on Curieuse Island is currently only in its second year, with 2016 having
been the first full year of data collection. Due to data still being rather limited, the analysis was not
extensive. However, we have been able to analyse data which show some trends that may prove to
become annual patterns in the future.

In the 2015 annual report, it was suggested that data for Anse St. Jose between June and October
may indicate a sediment shift from the south-eastern end of the beach to the north-west. When
analysing the results from this same beach in 2016, there also appears to be a similar indication of
this trend. In general, it appears that from July to November sediment is moving away from AJ06 at
the south-eastern end of the beach and increases at the other transects further north-west, which
coincides with the Southeast monsoon taking place predominantly from May to September. During
this time, the generally stronger winds account for longshore drift moving in the direction of the
north-west, shifting sediment towards AJ05 & AJ04 and the remaining transects. However, during
the Northwest monsoon season (November to March) the trade winds are blowing from this
direction towards the south-east; therefore we would be expecting to see opposite trends. During
this time of the year, beach width and area does seem to generally decrease between January and
March, with the exception of AJ06, which increases. This suggests sediment is being moved down
the beach towards the south-east accumulating at the end where a rocky outcrop at the end of the
beach would naturally trap sediment. Interestingly, AJ06 is not only the longest transect on this
beach, but is observed to have changed the most in width and area between the two extreme
measurements in all the surveys to date. This perhaps suggests that this transect undergoes the

103



most dramatic changes over the course of the seasons. This notion appears to be supported by
results from Anse Caiman, where there also appears to be a south-easterly movement of sediment
between the two transects during the Northwest monsoon, and a more north-westerly one in the
Southeast monsoon. Less clear are the results from Anse Cimitiere, and with the added factor of
missing data from July, it is difficult at this point to say whether the same patterns are seen for this
transect. However, this project is in a relatively early stage, and the results we are seeing here do
appear to fit expected seasonal changes affecting this side of the island. This is a positive sign, as
these important sea turtle nesting beaches appear to be losing and accreting sediment in what is
expected to become clear as an annual cycle.

In 2015, it was difficult to draw conclusions regarding the coast of the island on which Anse Laraie,
Anse Papaie, and Grand Anse are located due to limited data. Now that a full year of measurements
have been recorded, we may consider the effect, if any, that the monsoon seasons may be having on
Curieuse. The location and orientation of these three beaches (Figure 28) means that during the
Southeast monsoon the wind would be approaching the beach at an approximate right angle, and
overall it is quite difficult to predict which direction the longshore drift would be in.

When considering the observed results, it may be suggested that the Northwest monsoon may not
be having a significant effect on these beaches, and that the shape of the island shelters these
beaches from much of the season's effects the main exception being the first transect on Anse
Papaie, where April saw a noticeable increase in both width and area. This suggests that perhaps
milder winds and calmer seas that typically follow a monsoon season allowed for sediment
accumulation to occur simply via tide and wave action.

All three beaches do appear to experience movement in sediment to the north-east between June
and August, a trend which is expected as the winds are typically stronger during this period. The
wind and waves appear to be hitting this stretch of coast with longshore drift moving towards the
north-east. It is between August and October that a significant drop in beach width and area at GA03
has been detected, which is generally the peak of the Southeast monsoon season and suggests it is
only at this time a large amount of sediment moves from the north-eastern end of the beach, some
of which may be seen to accrete on the two transects to the south-west. It should be considered
that there may be an additional factor influencing the pattern of observed substrate movement. A
raised rocky ledge at the low tide mark extending some of the way across the middle towards the
north-eastern end of Grand Anse affecting only transect GA03 may be naturally trapping sediment

104



on this beach section. However, it would take detailed investigation into local longshore drift
currents, the effects of surrounding substrates, and sediment movements at the north-eastern end
of the beach to fully understand what is occurring to explain this accumulation pattern.

It would be wise to consider the suitability of the location and distribution of transects along the
study beaches. There appears to be some skew in where transects are located; for example, Anse St.
Jose originally had seven transects (AJ06 no longer used), whereas there were only three installed on
Grand Anse and only one on Anse Caiman. It would appear that the main area of Grand Anse is
underrepresented at the north-eastern end of the beach. It is advisable that a fourth transect at that
end should be added, and due to the questionable reliability of GA03, this transect could possibly be
moved to an area of the beach with a lower quantity of rock. Moreover, an additional transect on
Anse Cimitiere would allow for greater confidence in observed changes over time. However, if this
beach is not deemed long enough for another transect, it may be that continuing analysing it along
with Anse Caiman is a suitable alternative. Also, as it has previously been suggested, repositioning
the second transect on Anse Laraie (AL02) further towards the centre of the beach in front of the
Ranger Station may better represent this section of the beach when profiles are produced.

It should also be noted that reference marks at the beginning of each transect are not set back at a
consistent position from the sand, making it difficult to make direct comparisons between transects
on the same beach regarding the widths and areas of the beaches. This will continue to be an issue
unless the markers are altered and there are additional reasons for doing this. For example, many
of the reference marks are painted onto trees, which of course often continue to grow. As such, it
may be that some height changes attributed to sediment accretion could actually be due to the
movement of reference marks. It is therefore prudent to install new, more permanent reference
marks, in the form of solid and well installed posts, with marks at comparable heights. This has been
communicated with SNPA and plans are currently underway for this to occur in 2017. However, if
and when this occurs, changes in height and position of reference marks may in some cases result in
the inability to make direct comparisons to past data, if it is not possible to measure and calculate
adjustments from the existing reference marks to any new ones installed in different positions.
Because of this, it would be advisable at the same time to consider the possible relocation of
transects as discussed above.

Sediment budget, defined as the balance between the sediment gains and losses within a specified
control area over a given time, requires the identification of sediment sources (sediment inputs,

105



such as from land erosion/river outputs/coral reef erosion, etc.) and sinks (a point/area where beach
sediment is irretrievably lost from the system such as estuaries, sand dunes or deep seabed
channels) of a defined system (Rosati 2005). The rate of sediment movement must also be
estimated, and unpredictable variables can make it difficult to obtain accurate sediment budget
estimates (Rosati 2005). Ideally, we would be able to calculate the sediment budget for Curieuse
Island in order to better understand the effect of the coastal processes on the island, and whether
they are cycling annually as would be expected or whether each year more sediment is being lost
than gained. If such an imbalance were to be discovered, then future management plans to preserve
these valuable sea turtle nesting beaches would need to be explored and implemented to preserve
the viability of the island as a sea turtle rookery. To properly investigate sediment budgets on
Curieuse would require a much more detailed analysis of the island and its littoral sediment
movements, which GVI is not in a position to conduct at this time.

Several issues have been noted with the current computer program being used (Beach Profile
Analysis (Version 3.2)), one of these being that the program extrapolates the data from the end of
the profile inputted. However, depending on the data entered this is only done for some transects
and not for others. This appears to be having some effect on results, and at this stage data should
still be treated with caution. Further training could rectify these problems and allow for more
reliable analysis, and it is suggested that this is explored further in the coming months. However,
some initial research has begun into possible alternative computer programs, and it is hoped that in
the next year of this project other programs may be trialled. The current program provides limited
ways in which profiles can be viewed, a problem which could be addressed if a new program was
available to GVI that is perhaps compatible with GIS software. Being able to view the changes in our
beaches with GIS images would allow us to get a much clearer view of what is happening on the
island not just within the yearly cycle but also over the coming years.

A major contributing factor in determining the shape of a beach profile is sediment characteristics,
including grain size, sorting, and distribution (Hanson, 2016). Last year it was suggested that
collecting data on these factors would provide greater insight into the reason behind some of the
changes in beach profiles being seen. Other major influencing factors affecting beach profiles are
wave climate (wave length/period/height) and wave-generated currents (e.g. longshore drift), tides,
and the strength of swash and backwash (Hanson, 2016). However, collecting this type of data is
currently beyond the capabilities of our project and staff. Though collecting such additional
information may still be worth considering in time, as it would widen the extent to which the results

106



can be analysed and increase the usefulness of the conclusions being drawn. This type of initiative
would require collaboration with SNPA and possibly other organisations with further training,
equipment, and staff.

At present, the beach profiling conducted by GVI is strictly scheduled and takes place every month,
usually over the span of two days, and timed where possible to the lowest tides of the month.
Significant storm events can be the cause of considerable beach erosion (Morton, 2002), and it was
previously suggested that it would be very useful to be able to capture the effects of these events to
get a better idea of what is causing erosion and when exactly it is occurring. However storm events
are often unpredictable and adding extra beach profiling surveys is generally not logistically possible
at this time.

Conclusion
It is evident that there are still several issues that need to be addressed, including reference points,
suitable computer software, reference point construction, and the location of transects that are
being used. Over 2017, it will be imperative that these issues are resolved in order to produce
reliable beach profile data to support future management decisions. Other recommendations for the
future that have been discussed here include consideration of studies on sediment grain
size/sorting/distribution, attempting to capture additional data recording the effects of significant
storm events.

Completion of the first full year of beach profiling data has allowed us to gain some indication of
what could become clear trends for the island of Curieuse over the coming years. However, at this
stage, it remains too early to be able to undertake the analysis required to be able to come to
stronger conclusions. It is hoped that improvements in the methodology and analysis will also
support stronger conclusions regarding the observed trends in beach profile island-wide. The trends
observed thus far demonstrate that a project such as this is worth running at Curieuse Marine
National Park in order to guide future management of its vitally important sea turtle nesting
beaches.

107



References
Allen AC, Shah NJ, Grant A, Derand GD, Bell D. 2010. Hawksbill turtle monitoring in Cousin Island
Special Reserve, Seychelles: an eight-fold increase in annual nesting numbers. Endangered Species
Research 11:195-200

Barker MJ, Gruber SH, Newman SP, Schluessel V. 2005. Spatial and ontogenetic variation in growth
of nursery-bound juvenile lemon sharks, Negaprion brevirostris: a comparison of two age-assigning
techniques. Environmental Biology of Fishes 72: 343355

Bester C. 2014.Sicklefin lemon shark. Florida Museum of Natural History.
http://www.flmnh.ufl.edu/natsci/ichthyology/gallery/Descript/SicklefinLemonShark/sfLemonShark.
html

Blackmore S, Chin S, Chong Seng L, Christie F, Inches F, Utami P W, Watherston N and Wortley A H.
2011. Observations on the morphology, pollination and cultivation of Coco de Mer (Lodoicea
maldivica (J F Gmel.) Pers., Palmae) Journal of Botany

Blackwell BG, Brown ML and Willis DW. 2000. Relative weight (Wr) status and current use in fisheries
assessment and management. Reviews in Fisheries Science 8: 1-44

Bourne D and Coe M. 1978.The size, structure and distribution of the Giant Tortoise Population of
Aldabra. Phil. Trans. R. Soc. Lond. 282: 139-174.

Brunnschweiler JM, Abrantes KG, Barnett A. 2014. Long-term changes in species composition and
relative abundances of sharks at a provisioning site. PLoS ONE 9:e86682

Burt AJ, Dunn N, Mason-Parker C, Antha S & Mortimer JA. 2015. Curieuse National Park, Seychelles:
Critical Management Needs for an Important Nesting Habitat. Marine Turtle Newsletter 147: 6-11.

Bush A. 2003.Diet and diel feeding periodicity of juvenile scalloped hammerhead sharks, Sphyrna
lewini, in Kaneohe Bay, Oahu, Hawaii. Environ. Biol. Fish 67:111

Carpenter KE and Niem VH. 1998. FAO species identification guide for fishery purposes: The living
marine resources of the Western Central Pacific. Volume 2. Cephalopods, Crustaceans, Holothurians
and Sharks. Food and Agriculture Organization of the United Nations, Rome, 1998.

Chapman DD, Babcock EA, Gruber SH, Dibattista JD, Franks BR, Kessel SA, Guttridge T, Pikitch EK,
Feldheim KA. 2009. Long-term natal site-fidelity by immature lemon sharks (Negaprion brevirostris)
at a subtropical island. Molecular Ecology 18:3500-3507

Clua E, Buray N, Legendre P, Mourier J, Planes S. 2011. Business partner or simple catch? The
economic value of the sicklefin lemon shark in French Polynesia. Marine and Freshwater Research
62:764-770
Compagno LJV. 1990. Shark exploitation and conservation. Pratt HL Jr., Gruber, SH, and Taniuchi T
(Eds) Technical Report NMFS 90, 391-414. 1990. Honolulu, Hawaii, National Oceanic and
Atmospheric Administration. NOAA.

Cortes E. 2000. Life history patterns and correlations in sharks. Reviews in Fisheries Science 8: 299-
344

108



Daig. 2013. A baseline study of the distribution and abundance of mangrove species in the proximal
zone (outermost fringe) of Curieuse wetland. University of London, Dissertation.

DeAngelis BM, McCandless CT, Kohler NE, Recksiek CW, Skomal GB. 2008. First characterization of
shark nursery habitat in the United States Virgin Islands: Evidence of habitat partitioning by two
shark species. Marine Ecology Progress Series 358:257-271

deMarignac JRC. 2000. Home range and diel movement pattern of sub-adult lemon sharks in Bimini,
Bahamas. Master of Science The Faculty of Moss Landing Marine Laboratories

DiBattista JD, Feldheim KA, Thibert-Plante X, Gruber SH, Hendry AP. 2008.A genetic assessment of
polyandry and breeding-site fidelity in lemon sharks. Molecular Ecology 17:3337-3351

DiGirolamo AL, Gruber SH, Pomoroy C, Bennett WA. 2012. Diel temperature patterns of juvenile
lemon sharks Negaprion brevirostris, in a shallow-water nursery. Journal of Fish Biology 80:1436-
1448

Duncan KM and Holland KN. 2006. Habitat use, growth rates and dispersal patterns of Juvenile
Scalloped hammerhead sharks Sphyrna lewini in a nursery habitat. Marine Ecology Progress Series
312: 211-221

Dunn N, Mason-Parker C, Davies D, Quinton K, Portal B, Ward C, Lang R, Woods P, Jones Z, Parker-
Jurd F, Lovell P. 2014. GVI Seychelles- Curieuse Terrestrial Conservation Expedition Annual Report.

Edren SMC, Gruber SH. 2005. Homing ability of young lemon sharks, Negaprion brevirostris.
Environmental Biology of Fishes 72:267-281

Edwards PJ, Kollmann J, Fleischmann K. 2003.Life history evolution in Lodoicea maldivica
(Arecaceae). Nord. J. Bot. 22:227-237. Copenhagen

Feldheim KA, Gruber SH, Ashley MV. 2001. The breeding biology of lemon sharks at a tropical
nursery. Proc. R. Soc. Lond. B 269:1655-1661

Feldheim KA, Gruber SH, Dibattista JD, Babcock EA, Kessel ST, Hendry AP, Pikich EK, Ashley MV,
Chapman DD. 2014. Two decades of genetic profiling yields first evidence of natal philopatry and
long-term fidelity to parturition sites in sharks. Molecular Ecology 23:110-117

Filmalter JD, Daghorn L, Cowley PD. 2013. Spatial behaviour and site fidelity of the sicklefin lemon
shark Negaprion acutidens in a remote Indian Ocean Atoll. Marine Biology 160:2425-2436

Fleischer-Dogley F, Huber MJ, Ismail S. 2011a. Lodoicea maldivica. In: IUCN 2012. IUCN Red List of
Threatened Species. Version 2012.2. www.iucnredlist.org. Downloaded on 29 May 2013

Fleischer-Dogley F, Kettle CJ, Edwards PJ, Ghazoul J, Mtnen K, Kaiser-Bunbury CN. 2011b.
Morphological and genetic differentiation in populations of the dispersal-limited coco de mer
(Lodoicea maldivica): implications for management and conservation. Diversity and Distrib. 17:235-
243

Fleischer-Dogley F. 2006. Towards sustainable management of Lodoicea maldivica (Gmelin) Persoon.
PhD Thesis. University of Reading UK

109



Fleischmann K, Edwards P J, Ramseier D, Kollman J. 2005. Stand structure, species diversity and
regeneration of an endemic palm forest on the Seychelles. African Journal of Ecology.43:291-301.

Freitas RHA, Rosa RS, Gruber SH, Wetherbee BM. 2006. Early growth and juvenile population
structure of lemon sharks Negaprion brevirostris in the Atol das Rocas Biological Reserve, off North-
East Brazil. Journal of Fish Biology 68:1319-1332

Gallagher AJ, Hammerschlag N. 2011. Global shark currency: The distribution, frequency, and
economic value of shark ecotourism. Current Issues in Tourism 14:797-812

Gaymer R. 1968.The Indian Ocean giant tortoise Testudo gigantea on Aldabra. J. Zool. Lond 154:341-
363

Gerlach J, Rocamora G, Gane J, Jolliffe K, Vanherck L. 2013. Giant tortoise distribution and
abundance in the Seychelles islands: past, present, and future. Chelonian Conservation and Biology
12(1):70-83

Gibson CD, Hamilton J. 1984. Population processes in a large herbivorous reptile: the giant tortoise
of Aldabra atoll. Oecologia 61:230-240

Grubb P. 1971.The growth, ecology and population structure of giant tortoises on Aldabra. Phil Trans
R Soc Lond 260:327-372.

Gruber SH, deMarignac JRC, Hoenig JM. 2001. Survival of juvenile lemon sharks at Bimini, Bahamas,
estimated by mark-depletion experiments. Transactions of the American Fisheries Society 130:376-
384

Gruber SH, Nelson DR, Morrissey JF. 1988. Patterns of activity and space utilization of lemon sharks,
Negaprion brevirostris, in a shallow Bahamian lagoon. Bulletin of Marine Science 43:61-76

Hambler C. 1992.Giant Tortoise Reproduction and recruitment of Curieuse Island. Inform: Reports
from the Conservation and National Parks Service of Seychelles. Victoria: 1-18.

Hambler C. 1994. Giant Tortoise Geochelone gigantean Translocation to Curieuse Island (Seychelles):
Success or Failure Biological Conservation 69(3):293-299.

Hansen DM, Donlan CJ et al. 2010. Ecological history and latent conservation potential: large and
giant tortoises as a model for taxon substitutions. Ecography 33(2):272-284.

Hanson, L. 2016. Beach Profile: Influencing factors. Retrieved 15 January 2016 from
http://w3.salemstate.edu

Harley SJ, Myers RA, Dunn A. 2001. Is catch-per-unit-effort proportional to abundance? Canadian
Journal of Fisheries and Aquatic Sciences 58(9): 1760-1772

Harvey, JT. 1989. Food habits, seasonal abundance, size, and sex of the blue shark, Prionace glauca,
in Monterey Bay, California. California Fish and Game 75: 3344.

Heupel MR, Simpfendorfer CA, Bennett MB. 1998. Analysis of tissue responses to fin tagging in
Australian Carcharhinids. Journal of Fish Biology 52:610-620

110



Heupel MR, Carlson JK, Simpfendorfer C. 2007. Sharks Nursery Areas: Concepts, Definition,
Characterization and Assumptions. Marine Ecology Progress Series 337:287-297

Hill MJ. 2002. Biodiversity surveys and conservation potential of inner Seychelles islands. Atoll
research bulletin No 495. Issued by National Museum of Natural History Smithsonian Institution.
Washington D.C.

Hitchins PM, Bourquin O, Hitchins S. 2004. Nesting success of hawksbill turtles (Eretmochelys
imbricata) on Cousine Island, Seychelles. J. Zool., Lond. 264:383389

Hodgkiss R, Mason-Parker C, Grant A, McClelland J, Rademakers B, Sanchez C, Davies D. 2016. GVI
Seychelles - Curieuse Island Conservation Expedition Annual Report.

IUCN. 2014. Negaprion acutidens (sharptooth lemon shark). International Union for the
Conservation of Nature Red List of Threatened Species.

IUCN. 2017. International Union for the Conservation of Nature Red List of Threatened Species.
Retrieved 04th January 2017 from http://www.iucnredlist.org/

Johnson LK, Herren LW. 2008. Management of Fringing Mangrove Habitat in the Indian River Lagoon.
Florida Department of Environmental Protection.

Jolly GM. 1965. Explicit estimates from capture-recapture data with both death and immigration -
stochastic model. Biometrika 52: 225-247.

Koch B, Kaiser-Bunbury C. 2010. Long term survey of CDM growth in Fond Pepper: materials and
methods.

Leighton PA, Horrocks JA, Kramer DL. 2009. How depth alters detection and capture of buried prey:
exploitation of sea turtle eggs by mongooses. Behavioural Ecology 20(6):1299-1306

Leister, E. 2013. Tropical Cyclone Felleng Hits Madagascar and Nearby Islands. Retrieved 29 January
2017, from http://www.accuweather.com/en/weather-news/tropical-cyclone-felleng-hits/5329753

Lewis M, Adams D, Hill P, Heaney V. 1991.Findings of the Oxford University Expedition Study of the
Giant Tortoise Population on Curieuse Island.

Lewis R. 2005. Ecological engineering for successful management and restoration of mangrove
forests. Ecological Engineering 24:403-418

Limpus CJ, Nicholls N. 1988. The Southern Oscillation regulates the annual numbers of green turtles
(Chelonia mydas) breeding around northern Australia. Aust. J. Wildl. Res. 15:157-161

Lowe CG. 2002. Bioenergetics of free-ranging juvenile scalloped hammerhead sharks (Sphyrna
lewini) in Kaneohe Bay, Oahu, HI. J. Exp. Mar. Biol. Ecol. 278:141156

Manis, JE.2013. Assessing the effectiveness of living shoreline restoration and quantifying wave
attenuation in mosquito lagoon, Florida. Masters Thesis.

111



Manson FJ, Loneragan NR, Skilleter GA, Phinn SR. 2005. An evaluation of the evidence for linkages
between mangroves and fisheries: a synthesis of the literature and identification of research
directions. Oceanography and Marine Biology: an Annual Review 43:485-515
McFarland CG, Villa J et al. 1974. The Galapagos Giant Tortoises (Geochelone elephantopus) Part 1:
Status of the Surviving Populations Biological Conservation 6(2):118-133

Meyer CG, O'Malley JM, Papastamatiou YP, Dale JJ, Hutchinson MR, Anderson JM. 2014. Growth and
Maximum Size of Tiger Sharks (Galeocerdo cuvier) in Hawaii. PLoS ONE 9(1): e84799

Meylan AB, Donnelly M. 1999. Status Justification for Listing the Hawksbill Turtle (Eretmochelys
imbricata) as Critically Endangered on the 1996 IUCN Red List of Threatened Animals. Chelonian
Conservation and Biology. 3(2):200224

Moore ABM, White WT, Peirc, R. 2010. Additions to the shark fauna of the Persian (Arabian) Gulf
(Carcharhiniformes: Hemigaleidae and Carcharhinidae). Zoology in the Middle East 50, 8388

Morrissey JF, Gruber SH. 1993. Home range of juvenile lemon sharks, Negaprion brevirostris. Copeia
1993:425-434

Morrissey JF, Gruber SH . 1993b. Habitat selection by juvenile lemon sharks, Negaprion brevirostris.
Environ Biol Fish 38: 311319

Mortimer JA, Bresson R. 1994.The hawksbill nesting population at Cousin Island, Republic of
Seychelles: 1971-72 to 1991-92. Pp. 115-118, in Schroeder, B.A. & Witherington, B.E.
(compilers).Proceedings of the 13th Annual Symposium on Sea Turtle Biology and Conservation.
NOAA Technical Memo. NMFS-SEFSC-341

Mortimer JA. 1998. Turtle & Tortoise Conservation. Project J1: Environmental Management Plan of
the Seychelles. Final report to the Ministry of Environment Republic of Seychelles and the Global
Environment Facility (GEF). Volume 1 (82pp) and Volume 2 (Appendices 1-50)

Mortimer JA, Bresson R. 1999. Temporal distribution and periodicity in hawksbill
turtles (Eretmochelys imbricata) nesting at Cousin Island, Republic of Seychelles, 1971-
1997.Chelonian Conservation and Biology 3(2): 292-298

Mortimer JA, Donnelly M. 2008.Eretmochelysimbricata. In: IUCN 2012. IUCN Red List of Threatened
Species. IUCN SSC Marine Turtle Specialist Group. Version 2012.2.www.iucnredlist.org.

Mortimer JA. 1984. Marine Turtles in the Republic of Seychelles: Status and Management.
Publication of the IUCN Conservation Library: Gland, Switzerland.

Mortimer JA. 1990. The influence of beach sand characteristics on the nesting behaviour and clutch
survival of Green turtles (Chelonia mydas). Copeia. 3:802-817

Mortimer JA. 1998. Turtle & Tortoise Conservation. Project J1: Environmental Management Plan of
the Seychelles. Final report to the Ministry of Environment Republic of Seychelles and the Global
Environment Facility (GEF). Volume 1 (82pp) and Volume 2 (Appendices 1-50)

Mortimer JA. 2004. Seychelles Marine Ecosystem Management Project (SEYMEMP): Turtle
Component. GEF Final Report, Vol. 1: 243 pp. Vol. 2: Appendix 1-11, 158 pp

112




Morton, RA. 2002. Factors Controlling Storm Impacts on Coastal Barriers and Beaches: A Preliminary
Basis for near Real-Time Forecasting, Journal of Coastal Research. 18(3): 486-501.
Mourier J, Buray N, Schultz JK, Clua E, Planes S. 2013. Genetic network and breeding patterns of a
sicklefin lemon shark (Negaprion acutidens) population in the Society Islands, French Polynesia. PLoS
ONE 8:e73899

Murchie KJ, Schwager E, Cooke SJ, Danylchuk AJ, Danylchuk SE, Goldberg TL, Suski CD, Philipp DP.
2010. Spatial ecology of juvenile lemon sharks (Negaprion brevirostris) in tidal creeks and coastal
waters of Eleuthra, The Bahamas. Environmental Biology of Fishes 89:95-104

Nentwig, V., Bahlurg, H., & Monthy, D. 2014.Sedimentology of Coastal Deposits in the Seychelles
Islands - Evidence of the Indian Ocean Tsunami 2004, Pure and Applied Geophysics.172(2015):641-
656

Nicholls H.2012. Conservationists report the exciting recovery of the Pinzn tortoise in the
Galapagos. The Guardian.<http://www.theguardian.com/science/animal-magic/2015/jan/16/rat-
eradication-galapagos-tortoise-pinzon>. Accessed 24 January 2015.

Obura D, Abdulla A. 2005. Assessment of tsunami impacts on the marine environment of the
Seychelles. Seychelles Ministry of Environment

Papastamatiou YP, Wood CL, Bradley D, McCauley DJ, Pollock AL, Caselle JE. 2014. First records of
the sicklefin lemon shark, Negaprion acutidens, at Palmyra Atoll, Central Pacific: A Recent
Colonization Event? Marine Biodiversity Records 7:1-3

Payet R and Agricole W. 2006. Climate Change in the Seychelles: Implications for Water and Coral
Reefs. Ambio 35(4): 182-189

Pemberton JW, Gilchrist JS. 2009. Behavior and Diet Preferences of a Released Population of Giant
Tortoises in the Seychelles. Chelonian Conservation and Biology 8(1):57-65

Pillans, R. 2003. Negaprion acutidens. The IUCN Red List of Threatened Species 2003.

Plummer MV, Ferner JW. 2012. Marking Reptiles In Foster M and McDiarmid R (Eds) Reptile
Biodiversity (pp. 143-150). Berkeley and Los Angeles, CA: University of California Press.

Pollock KH, Nichols JD, Brownie C, Hines JE. 1990. Statistical inference for capture-recapture
experiments. Wildlife Monographs 107: 3-97.

Pratt HL and Casey JG. 1983. Age and growth of the shortfin mako, Isurus oxyrinchus using four
methods. Canadian Journal of Fisheries and Aquatic Sciences 40:1944-1957

Ramalanjaona G. 2011. Impact of 2004 Tsunami in the Islands of Indian Ocean: Lessons Learned.
Emergency Medicine International, Volume 2011, pg 1-3

Rainbolt, RE 1996. Geochelone gigantean (Aldabra Giant Tortoise) Predation. Herpetological Review
27(4): 1999.

Ricker, WE. 1975. Computation and interpretation of biological statistics of fish populations.

113



Bulletin of the Fisheries Research Board of Canada 191:1-382

Rist L, Kaiser-Bunbury CN, Fleischer-Dogley F, Edwards P, Bunbury N, Ghazoul J. 2010. Sustainable
harvesting of coco de mer, Lodoicea maldivica, in the Vallee de Mai, Seychelles. Forest Ecology and
Management 260:2224-2231

Rosati, JD. 2005. Concepts in sediment budgets, Journal of Coastal Research. 21(2): 307-322
Samour HJ, Spratt DMJ, Hart MG, Savage B, Hawkey CM. 1987. A survey of the Aldabra Giant
Tortoise Population Introduced on Curieuse Island, Seychelles Biological Conservation 41:147-148

Sanchez C, Davies D, Grant A, Hodgkiss R, McClelland J, Kovacev S, Howard J, Dunn N. 2015. GVI
Seychelles - Curieuse Island Conservation Expedition Annual Report.

Savage JP and Ashton PS. 1984. The Population Structure of the Double Coconut and Some Other
Seychelles Palms. Biotropica 19: 282-283

Schultz JK, Feldheim KA, Gruber SH, Ashley MV, McGovern TM, Bowen BW. 2008. Global
phylogeography and seascape genetics of the lemon shark (Genus Negaprion). Molecular Ecology
17:5336-5348

Schwarz CJ, Arnason AN. 1996. A general methodology for the analysis of capture-recapture
experiments in open populations. Biometrics 52: 860-873.

Seber GAF. 1965. A note on the multiple recapture census. Biometrika 52: 249-259.

Seminoff J (Southwest Fisheries Science Center US). 2004. Chelonia mydas In: IUCN 2012. IUCN Red
List of Threatened Species. Version 2012.2. www.iucnredlist.org.

Seret B. 2002) Shark fisheries and conservation in the Seychelles. Consultancy Report for the
Seychelles Fishing Authority. December 2002.

Seychelles Fishing Authority. 2007. Seychelles national plan of action for the conservation and
management of sharks. Government of the Seychelles.

Seychelles National Parks Authority (SNPA). 2012. Curieuse mangroves a manual for field study.
Seychelles. Print.

SIF unpub. 14 January 2015. Facebook post at 8pm.<https://www.facebook.com/pages/Seychelles-
Islands-Foundation-SIF/1414466072110654>.

Silverton J. 1987. Possible sexual dimorphism in the Double Coconut: Reinterpretation of the data of
Savage and Ashton. Biotropical 19: 282-283

Speed CD, Meekan MG, Field IC, McMahon CR, Stevens JD, McGregor F, Huveneers C, Berger Y,
Bradshaw CJA. 2011. Spatial and temporal movement patterns of a multi-species coastal reef shark
aggregation. Marine Ecology Progress Series 429:261-275

Stevens JD (1984) Life-history and ecology of sharks at Aldabra Atoll, Indian Ocean. Proceedings of
the Royal Society of London B: Biological Sciences 222:79-106

114



Stevens JD, Bonfil R, Dulvy NK, Walker PA (2000) The effects of fishing of sharks, rays, and chimeras
(Chondrichthyans), and the implications for marine ecosystems. ICES Journal of Marine Science
57:476-494

Stoddart DR, Cowx D, Peet C, Wilson JR. 1982. Tortoises and tourists in the Western Indian Ocean:
The Curieuse experiment. Biological Conservation.24:6780

Stoddart DR, Peake JF, Gorden C, Burleigh R. 1979. Historical Records of Indian Ocean Giant Tortoise
Populations. Phil. Trans. R. Soc. Lond. B286: 147-161

Swingland IR, Coe M. 1979. The natural regulation of giant tortoise populations on Aldabra Atoll.
Reproduction J. Zool. Lond. 186:285-309.
Swingland IR. 1977. Reproductive effort and life history strategy of the Aldabra Giant tortoise.
Nature, Lond. 269: 402-404

Tortoise & Freshwater Turtle Specialist Group. 1996. Geochelone gigantea. The IUCN Red List of
Threatened Species. Version 2014.3.www.iucnredlist.org.Downloaded on 25 January 2015.

Wetherbee BM, Gruber SH, Rosa RS. 2007. Movement patterns of juvenile lemon sharks Negaprion
brevirostris within Atol das Rocas, Brazil: A nursery characterized by tidal extremes. Marine Ecology
Progress Series 343:283-293

White WT, Potter IC. 2004. Habitat partitioning among four Elasmobranch species in nearshore,
shallow waters of a subtropical embayment in Western Australia. Marine Biology 145:1023-1032

Worm B, Davis B, Kettemer L, Ward-Paige CA, Chapman DD, Heithaus MR, Kessel ST, Gruber SH.
2013. Global catches, exploitation rates, and rebuilding options for sharks. Marine Policy 40:194-204

Yeiser BG, Heupel MR, Simpfendorfer CA. 2008. Occurrence, home range and movement patterns of
juvenile bull (Carcharhinus leucas) and lemon (Negaprion brevirostris) sharks within a Florida
estuary. Marine and Freshwater Research 59:489-501

Yoshihiro M, Michimasa M, Hitonori N, Motohiko K, Toyohiko M, Nobuyuki K and Daijiro K.
2002.Coastal erosion due to long-term human impact on mangrove forests. Wetlands Ecology and
Management 10:1-9

115