Vegetation Type 2: Upland Rainforest
The survey on this principal vegetation type was restricted to forested areas at
elevations of 650-850m. Three forest types were observed and quantified using 41
plots in six transects, and were characterised as follows:
Ridge top forest type
A total of 24 plots along three transects were used to analyze this forest type. There
was an average of 30 trees (range 17-55) recorded in each plot, and an average of
eighteen species (range 12-29) per plot. The most common species were Syzygium
spp. (yasiyasi) and Litsea
sp. (lidi). The largest tree encountered was Agathis
(dakua makadre) with a dbh of 152cm. Other large trees included
sp. (150cm) and two other A. macrophylla
(140, 111cm). The average dbh
was 18cm (range 5-152cm) with the dominant species being A. macrophylla
relative dominance). All dominant trees recorded from each plot are important
timber tree species with dbh ranging from 31cm to 152cm and these include
Calophyllum vitiense, Dacrydium nidulum
spp., A. macrophylla
spp. (bau). Together these five species comprised approximately 75% of the
biomass of all trees in this forest type.
Slope forest type
Fifteen plots along two transects contained an average of 28 (range 10 to 40) trees per
) with an average number of 17 (range 8 to 25) species per plot. The most
common species were Gironniera celtidifolia
(sisisi) and Cyathea
spp. (balabala) whilst
other common species in some of the plots included C. vitiense
and Saurauia rubicunda
(mimila). The largest tree encountered was Garcinia myrtifolia
with a dbh of 89cm.
Other large trees recorded included Endospermum macrophyllum
(dbh of 71, 65cm), A.
(69cm) and C. vitiense
(68, 65cm). The average dbh was 19cm (range 5-152
cm). Overall there was no dominant species for this forest type
, but across the plots
the trees that together made up 80% of the total biomass were E. macrophyllum, A.
, C. vitiense, Semecarpus vitiensis
masiratu) and Buchanania attenuata
(maqo ni veikau).
Vegetation Type 3: Cloud Forest
On Emalu the cloud forest was restricted to mountain tops and ridges above 850m
and is almost always shrouded in clouds. Precipitation is high and temperature is
generally much lower with trees generally stunted and heavily covered with
bryophytes. A series of eleven plots placed along a fragmented transect over a
slope/ridge towards the summit of Mt. Vonolevu (1,111m) was used to quantitatively
assess this forest type.
An average of 42 trees per plot (range 31 to 70) with an average number of 15 species
per plot (range 11 to 22) was recorded for the area. The most common species were
spp. and Cyathea
spp. The largest tree, with a dbh of 181cm, was Syzygium
sp. Other large trees included Calophyllum vitiense
(60cm dbh) and Degeneria vitiensis
(43 cm dbh). The average dbh was 12cm (range 5 -181cm) and the average bole height
was 1.8m (range 1 to 5m). C. vitiense
sp. were dominant species from
some of the plots assessed but overall D. vitiensis
and C. vitiense
were the dominant
species with greater than 80% relative dominance.
Vegetation Type 4: Dry Forest
The native dry forest vegetation type on the leeward side of Viti Levu has been
almost completely destroyed by combined grazing, agriculture and fire. The survey
on this principal vegetation type was restricted to forested areas adjacent to the
grassland in elevations ranging from 250m to 650m within the Nasa catchment. In a
seasonal dry forest the mean annual rainfall is about 2,000mm.
Three habitat or forest types were quantified using 44 plots in six transects. The
forest types and their characteristics were as follows:
River flat forest type
The nineteen plots assessed along three transects held an average of fifteen (range 7
to 24) trees and an average of seven species (range 4 to 11) were present within a
plot. The most common tree species were Syzygium malaccense
(kavika) and Citrus
(moli kania). The largest trees were C. grandis
and Dysoxyllum richii
kei rakaka), having a dbh of 89cm and 86cm, respectively. The average dbh of trees
was 21cm (range 10-89cm). The dominant species from some of the plots assessed
was D. richii
with a relative dominace of 80% and S. malaccense
with 69%. Overall,
there were no dominant species across the haitat as the biomass was fairly evenly
distributed amongst the larger trees. Most of the trees discussed above are associated
with human habitation.
Ridge top forest type
Within the ten plots used to analyze this habitat there was an average of 35 trees
(range 21 to 51) and sixteen species (range 10 to 25) per plot. The most common
species was Litsea
sp., followed by Garcinia myrtifolia
and Citrus grandis.
tree observed was Ficus obliqua
(baka ni viti) with a dbh of 132cm. Other large trees
included several Bischofia javanica
(86, 77 and 76cm dbh) and Dysoxyllum
dbh). The average tree dbh was 23cm (range 10-132cm). The dominant trees from
some of the plots assessed were F. obliqua
sp. with a relative
dominance of 57% and 53%, respectively. Overall there was no dominant tree species
in this habitat despite B. javanica
having a relative dominance overall of 48%. Like the
river flat above, the presence of tree species like C. grandis, S. malaccense
is indicative of past human habitation and activity in this area.
Slope forest type
Fifteen plots along two transects were used to analyze this habitat. An average of 23
trees (range 8 to 37) with an average number of ten species (range 6 to 14) per plot
was recorded. The most common species was Litsea
sp. with the largest tree
encountered being F. obliqua
with a dbh of 148 cm and other large trees that included
and Neonauclea fosteri
(vacea) with dbh of 120 and 113cm,
respectively. The average dbh was 22cm (range 6 to 148). The dominant species
assessed were D. quercifolium, F. obliqua
and N. fosteri
all with relative dominance of
greater than 74%. Overall, the dominant species for this forest type were the
spp. (4 taxa) and Ficus obliqua
which together comprised 75% of the total
biomass in this habitat.
Overall, the forest or habitat types found in this principal vegetation type are best
described as an anthropogenic primary forest as most of the more dominant and
common tree species are associated with human activity. Other species not found in
the plots that testify to this include Codiaeum variegatum
(sacasaca), Cordyline fruticosa
(qai), Schizostachyum glaucifolium
(bitu dina) and Veitchia joannis
Vegetation Type 5: Talasiga Vegetation
Grassland/talasiga habitat type
The grassland is restricted to the slopes and ridge tops and is mostly made up of the
grass Pennisetum polystachyon
(mission grass), Sporobolus
spp. (wire grass),
spp., (qato or bracken ferns), Pteridium esculentum, Miscanthus floridulus
(gasau or reed) and many other smaller weedy plants. The general lack of tree cover
is characteristic of such a landscape. The grassland is regularly set on fire to allow for
new re-growth of grass for use as fodder for cattles and horses. Areas closer to the
edge of the gully forest are used for subsistence farming.
Woody shrubland habitat type
This vegetation was observed growing between the grassland and the forest edge
and is also referred to as savannah grassland. The area was dominated by secondary
pioneer plant species like Commersonia bartramia
(sama), Parasponia andersonii
(vakaceredavui), Trema orientalis
, Dillenia biflora
(nuqanuqa) and larger patches of Schizostachyyum glaucifolium
. This habitat is where active agricultural activities are occurring both at the
subsistence level and on a semi-commercial scale. Gardens or plantations of yaqona,
banana and taro are common and so are patches of abandoned (fallow) gardens.
Such activity expands the grassland habitat types into forested areas and as noticed
from the survey will continue to do so especially with increasing pressure from
subsistence farming and a growing population.
River bank/riparian habitat type
The vegetation along the creek and river system that is found adjacent to the
grassland was mostly dominated by important introduced and native fruit trees.
Also found here were important trees species that have cultural uses
, such as
(ivi, chestnut), Pometia pinnata
(dawa), several species of Citrus
(uto, breadfruit), Cocos nucifera
(niu), Spondus dulcis
(kavika) and Terminalia catappa
(tavola). Other culturally important trees
include Aleurites moluccana
, Bischofia javanica
, Cananga odorata
(qai) and Euodia hortensis
Conclusions and recommendations
The discovery of the focal species detailed above, in particular, the priority
conservation species on the IUCN Red List, as well as rare orchids, parasitic plants
and palms is an indication of the micro-sensitivity and function of the upper
catchment areas that have yet to be fully explored, and which needs protecting.
Based on current knowledge of these taxa, any level of development (logging or
agricultural) could seriously affect their existence, thus more effort needs to be
invested in their protection. Overall the presence of a large number of these high
value conservation species within the Nasa
, Mavuvu and Waikarakarawa Creek
catchment highlights the biodiversity importance of Emalu. For Viti Levu (and for
Fiji as a whole) it is an area with the highest concentration of important plants of
In terms of the vegetation, the level of human impact decreases as you move further
inland or away from current human habitation, and also towards the higher altitudes
of Mt. Vonolevu. Demarcation of these habitat types is quite obvious in the grassland
vegetation but almost near impossible to detect under heavy canopy in the forested
lowland and upland vegetation including the riparian system running across these
vegetation types. It is recommended that more extensive future surveys be carried
out in these areas.
Nunia Thomas and Isaac Rounds
This report documents the first record of herpetofauna biodiversity within the Emalu
study area. Emalu, like many other parts of Viti Levu, contains habitats ideal for
herpetofauna. Despite weather and time constraints this survey produced results
similar to surveys carried out in other areas of Viti Levu
, encountering six species of
herpetofauna: three endemic, two native and one invasive. Further surveys will very
likely reveal the existence of additional herpetofaunal species.
To date, there has been no documented information on the herpetofauna of the
Emalu area. This report is therefore the first documented study of these organisms.
The objectives of this baseline herpetaofauna survey were to:
Document the herpetofauna diversity in the study area.
Identify ideal herpetofauna habitat.
Trial herpetofauna survey methods with recommendations for long-term
monitoring in the study site.
During the survey periods the weather was generally fine every day with occasional
and sometimes heavy afternoon showers. At the Waikarakarawa survey site heavy
rain on one day resulted in flashflooding. Weather conditions dictated the number of
days, type of traps and survey methods conducted, and these are summarised in
Appendix 5. Average air temperatures recorded for the nocturnal surveys were
20.6°C and 26°C for the Nasa and Waikarakarawa catchments respectively.
The objective of the expedition was to record all herpetofauna species captured
and/or observed within the study site; and develop appropriate long term
monitoring methods. For this reason, all potential habitats within good forest cover
and outside of the forest were surveyed. The study area generally had ideal
herpetofauna habitats: riparian vegetation, ridge forest, forest floor cover of leaf litter
and rotting wood, and trees with dense epiphyte cover. Systematically, the survey
targeted a ridge habitat, riparian forest habitat and lowland forest habitat. A total of
ten sites were intensively surveyed (Map 5).
Diurnal and Nocturnal Herpetofauna Surveys
There are several accepted methods for herpetofauna surveys that generally fall
under two categories: opportunistic
diurnal and nocturnal searches and trapping,
nocturnal and diurnal searches and trapping. A summary of the
methods used in this survey is given in Appendix 5
Herpetofauna surveys in Fiji have generally been opportunistic, but their methods
standardised to allow for comparison between sites. Other long term herpetofauna
monitoring plots on Viti Levu: the Sovi Basin Conservation Area and the Wabu
Forest Reserve are limited to nocturnal frog searches. Because of the cryptic and
heliophilic nature of Fiji’s reptiles; and Fiji’s climate, survey and trap methods are
wide ranging, albeit limited by weather conditions.
The herpetofauna surveys in the Emalu study site consisted of three techniques but
were constrained by the rainy weather. These are described below:
Standardised sticky trap transects
whereby sticky mouse traps (Masterline®) are
laid out at intervals along a transect. Each station is designated a station number (1-
10) with a cluster of three traps per station for three placements to represent local
habitat structure at each location (tree, log and ground). Transects are laid out along
identified ideal habitats e.g. ridge tops and along river banks/ riparian vegetation.
Leaf litter cover, canopy cover and undergrowth are all recorded. Left overnight,
traps are checked regularly for captured specimens. These traps target both
terrestrial and arboreal species.
Frogs and geckoes are active and more visible at night. Standardised (time
constrained) nocturnal visual encounter surveys
(2 hours) in ideal frog habitat are
used. This method gives an encounter rate for comparison with other surveys within
Fiji. Search efforts with a minimum of two observers at any one time targeted
streams and adjacent banks/ flood plains.
Opportunistic Visual Encounter Surveys
outside of the standardised searches allow
for a record of presence/absence of herpetofauna. Skinks are more likely to be seen
during the day, particularly during hot and sunny conditions. Opportunistic diurnal
surveys were conducted along trails enroute to the camp site, vegetation plots, along
stream edges, and in forest habitats surveyed by other survey teams in the
expedition. Search efforts targeted potential skink habitat and frog and burrowing
snake diurnal retreat sites. The diurnal surveys began at 09:00 and ended at 15:00 on
each of the survey days. The team had a minimum of two searchers at any one time.
Environmental variables such as air temperature, water temperature, weather
conditions (rain/fine) and cloud cover (%) were taken at the beginning and end of
each nocturnal survey. Habitat characteristics and other basic ecological and
biological information of herpetofauna found were recorded. Observations on
possible threats to herpetofauna species and populations were also noted.
Geographic coordinates of survey sites were captured using the Thales Mobile
Mapper Pro Navigator and Garmin GPSmap 60CSx
Based on the current knowledge of herpetaofauna on Viti Levu there are a total of 26
species that could potentially occur in the study area (Appendix 6). Prior to the
survey a target list of 12 of these species was drawn up
, based on their endemism
and conservation status.
In total six species were encountered over the course of the survey, including three of
the 12 target species. These were the green tree skink (Emoia concolor
), the bronze-
headed skink (E. parkeri
) shown in Fig. 13 and the Fiji tree frog (Platymantis vitiensis
shown in Fig. 14.
A total of six species were captured during the survey of the Nasa catchment. Three
of these were endemics (Platymantis vitiensis, Emoia parkeri
and E. concolor);
native (Nactus pelagicus,
and Gehyra vorax,
and one was an invasive
species (Bufo marinus,
Fig. 64). These findings were the result of over 14 man-hours of
diurnal survey, 49 hours of sticky trapping and six man-hours of nocturnal surveys.
Two species were reported to occur by local villagers: the endemic banded iguana
) and the Pacific boa (Candoia bibroni
), but were not encountered
during the expedition.
Herpetofauna were observed at all the three habitat types targeted
; but at only two of
the survey sites. The majority of the species were encountered during opportunistic
surveys (4 species); with lower encounter rates for the sticky traps (2 species), and
standard diurnal (2 species) and nocturnal surveys (1 species).
Interestingly, the sticky traps did not yield any rats or invasive ants – which have
been encountered in other survey sites on Viti Levu.
Waikarakarawa and Mavuvu Catchments
For these two catchments the same six herpetofauna species that were encountered in
the Nasa catchment were also found here. The survey of the Waikarakarawa and
Mavuvu catchments consisted of 8 man-hours of diurnal survey, 14 hours of sticky
trapping and 3.3 man-hours of nocturnal surveys.
One of the main target species known to occur from historical records to occur in the
area of Waikarakarawa catchment, the Fiji burrowing snake (Ogmodon vitianus
not encountered during this survey.
This report documents the first record of herpetofaunal diversity in the Nasa,
Waikarakarawa and Mavuvu catchments on land belonging to the mataqali Emalu.
Fiji’s terrestrial herpetofauna are significantly impacted by introduced mammalian
predators. This is particularly true for Viti Levu which has experienced the
extirpation of two large terrestrial skinks (Emoia trossular
and E. nigra
) in the presence
of the mongoose, feral cats, feral pigs and rats.
The presence of the Fiji Tree Frog, Platymantis vitiensis
(Fig. 14) in the study area is of
exceptional interest – this is a new record for the area and is possibly the western-
most record of the occurrence of the species (in relation to the wet parts of Viti Levu)
in Southern Viti Levu to date.
The apparent absence of the common ground skinks such as E. cyanura
the study area and in the agricultural land is interesting and warrants more intensive
searches both within and outside the forested areas, taking into consideration that
weather impacts the observer’s ability to find these species.
The low encounter rates and low diversity of herpetofauna in the study sites do not
necessarily mean an absence of the species. Low encounter rates of heliophlic species
is not uncommon in Viti Levu’s forests; and is typical globally in rainforest habitats
(Ribeiro-Junior et al
., 2006, Ribeiro-Junior et al
., 2008). Consequently, there are efforts
to develop better quantitative survey methods of forest dwelling herpetofauna – and
these will be considered in the development of an appropriate long term monitoring
method for the Emalu study area. However, sites to target for the establishment of
long-term monitoring plots should ideally be adjacent to the vegetation sample plots,
as done so in this study because of the dependence of native herpetofauna on the
health of the forest. Herpetofauna survey sites will also be extended to the non-
forested parts of the study area to assess the presence/absence of the more common
native ground skinks in the area.
Selecting which herpetofauna species could act as indicators of high conservation
value forest was problematic for several reasons. Firstly, the tree frog (P. vitiensis
was not a suitable indicator species as it was found all over the study area from
disturbed areas right up to the cloud forest. Furthermore, skinks or geckos are not
ideal indicator species as they are cryptic. The invasive cane toad (B. marinus
found everywhere except the cloud forest.