6 Freshwater Macroinvertebrates Bindiya Rashni
6.1 Introduction The Fijian freshwater macroinvertebrate fauna is represented by 45 families, namely;
25 families of insects, eight families of molluscs, four families of crustaceans, three
families of segmented worms, two families of nematodes, two families of sponges,
and one family of flatworms (Haynes, 1988; Haynes, 1999; Haynes, 2001; Jeng et al.,
2003; Haynes, 2009). Many of these are yet to be fully described to genus and species
level and many aquatic insect larvae need to be matched to their described flying
Prior to this study, there have been no surveys conducted on the composition of
freshwater macroinvertebrate communities within the waterways of the study sites
detailed in this report or their tributaries. There is, however, some documentation of
previous macroinvertebrate surveys in other waterways of Vanua Levu focusing on
the freshwater gastropods (Haynes, 1988; Haase et al., 2006) and Atyid shrimps
(Choy, 1991) only. These studies were conducted to document the aquatic
gastropods and shrimps present in easily accessible streams in Vanua Levu.
Therefore the present study represents the first detailed and comprehensive study of
freshwater macroinvertebrates and the aquatic habitats within the Mt. Delaikoro,
Sorolevu and Savusa catchments.
The key objectives of the study were to provide a comprehensive list of taxa,
describe community structure and identify taxa that are unique, rare and
endangered in Fiji. This report also provides information relating to water
physicochemistry that supports macroinvertebrate communities at waterways
surveyed in the two main provinces (Macuata and Cakaudrove) of Vanua Levu.
6.2 Methodology Survey Stations During the Vanua Levu freshwater survey (September-October 2013), eight main
stations (VL1-VL8) were sampled within the Macuata province and four major
stations (VL9-VL11 and VL13) in Cakaudrove province. The catchments targeted in
both provinces include waterways that supply water to the residents of Vanua Levu.
The descriptions of the sampling stations are summarized in Table 3 and their
locations shown in Figure 22. Photographs of the habitats of the sampling stations
are given in Appendix 12.
Table 3: Macroinvertebrate sampling localities and methods used at each River/Stream Site Code Description Survey type Nasealevu village
Surber & Kick-netting
Next to village
Next to bridge-confluence Kick-netting
Next to village
Surber & Kick-netting
Waisali river upper
Savusa-Savutagitagigagone VL10 Upstream-above waterfall Kick-netting
VL12 Upstream-above waterfall Hand-picking
VL14 Upstream-above waterfall Hand-picking
VL13 Next to village
VL15 Roadside spring
VL16 Next to current logging site Hand-picking
Water physicochemistry Water physicochemical parameters were measured at each sampling station using a
calibrated multi-parameter water quality meter (Aquaread AP 1000). Parameters
measured included temperature, dissolved oxygen (DO), conductivity (milisiemens
per centimeter (mS/cm), pH, Total Dissolved Solids (TDS), turbidity (Nephelometric
(NTU)) and salinity. Water Quality was taken only at major
sampling stations where Surber sampling or kick-netting was carried out.
Macroinvertebrate sampling Macroinvertebrate samples were collected using both quantitative and qualitative
survey methods to allow an assessment of macroinvertebrate density at selected
stations and to compile a list of taxa present at each site. The quantitative and
qualitative sampling methods were adapted from Stark et al. (2001) and modified to
suit the time period and objectives of this particular survey.
Quantitative assessment – This is a quantitative method that provides a measure of
macroinvertebrate density, adapted and modified from Protocol C3 (Stark et al.,
2001). Three replicate Surber samples (area 0.1 m², 0.5 mm mesh) were collected
from riffle habitats at stony streambed sites. A riffle is a shallow area (water depth
≤ 0.5 m) where water flows swiftly over stones, creating surface turbulence. Samples
were collected by placing the Surber sampler over a defined area of streambed in
riffle habitat and disturbing the habitat by washing the particles with the water
flowing through the net to collect dislodged macroinvertebrates. Surber sampling
was only carried out for two sites; Nasealevu village [VL1] and Waisali village [VL9]
due to time constraints.
Qualitative assessment – a single sample was collected from each sampling station
via 3-minute kick-netting over five metre riffle and run habitats, or hand-picking
using thumb forceps (opportunistic collection) where necessary. Typical habitats
sampled included runs, riffles, chutes, pool edges, woody debris, leaf litter, stream
edges, and tree roots along banks, stream bank vegetation and sand/silt substrates.
The purpose of multi-habitat sampling is to provide a list of taxa at the selected
station. Kick-netting was carried out at all main stations (VL1-VL11 and VL13),
therefore it will be used for the majority of the data analysis. For the remaining sites
(VL12 and VL14-VL16), opportunistic collection was conducted for taxa of interest.
Macroinvertebrate samples collected were placed into 250ml specimen jars with 70%
ethanol for sorting and identification by the author (Bindiya Rashni). Crustacean
(prawn and shrimp) specimen identification was confirmed by Laura Williams,
crustacean specialist at the School of Marine Studies, USP. The guides referenced in
the identification process included; Haynes (2009), Haynes (in prep.), Haase et al. (2006), Williams (1980) Winterbourn et al. (2006), and
Marquet et al. (2003),
(1983; 1991). Identified macroinvertebrates were preserved in 100% ethanol for long
Data analysis Community composition and structure: the combined Surber and kick-net data set was
used to calculate the relative abundance of the main taxonomic groups.
Macroinvertebrate density: an assessment was made of macroinvertebrate density in
riffle habitats at selected stony streambed sites based on quantitative Surber sample
data by multiplying the mean Surber sample abundance data (per 0.1 m
) by a factor
of ten to give abundance/m
Status & distribution of taxa: taxa were classified as endemic and native to Fiji, native
to other regions (e.g. Pacific, South Pacific, Indo-Pacific, and South East Asia),
introduced tropical species or other (i.e. unknown for new records).
Functional feeding group (FFG) assessment – FFGs represent the mode by which
macroinvertebrate taxa feed (i.e., collector-filterer, scraper, grazer, predator or
shredder). The FFG assessment involved calculating the number of taxa within each
FFG and the relative abundance each group made up across sampling sites.
Taxa of interest: macroinvertebrate taxa of potential interest suspected to be a new
record for Vanua Levu or Fiji or to Science. 6.3 Results Water physicochemistry The water physicochemistry parametres measured at the different stations are
summarised in Appendix 13. Waterways sampled ranged from almost neutral to
slightly acidic. The freshwater macroinvertebrate communities described in this
survey are unlikely to be significantly affected by pH values within this range.
Conductivity is a measure of the total ions in water and ranged between 1.110
mS/cm in the Nasealevu village waterway (VL1) and 0.054 mS/cm in the Savusa-
Turbidity (NTU) is a measurement of particles in the water column and provides an
indication of water clarity. Turbidity values ranged between 0 NTU in the majority
of sites (VL2-VL5, VL7, VL8, VL9, VL10, and VL13) to 2.4 NTU in the Nasealevu
village (VL1). Turbidity in Nasealevu village stream was higher due to heavy rainfall
a few nights ago prior to surveying. Turbidity above 5 NTU signifies poor water
quality; all the sampling stations had turbidity values less than 5 NTU. In the
majority of waterways surveyed turbidity values were 0 NTU, which signifies
excellent water quality for macroinvertebrate survival as well as the absence of
sediment-raising activities in the catchment, or at least not within the range of the
Dissolved oxygen concentrations ranged from 8.97 g/m
in Waisali village stream
(VL9) to 8.24 g/m
in Vunidogoloa-Wai Koroalau stream (VL13). All dissolved
oxygen concentrations were above the level considered sufficient for
macroinvertebrate survival (i.e. >5 g/m
). Waterway hydrology at sites surveyed
was unaltered except for the upper Dreketi (VL2) which had a culvert and Doguru-
Suweni river (VL8) which had a bridge, but these do not seem to have affected the
DO levels required for survival of macroinvertebrates, although alteration of flow is
highly possible. Salinity measurements at the survey stations demonstrated levels
that are expected in the waterways of any tropical inland river or stream.
Taxa richness and abundance A total of 70 distinct macroinvertebrate taxa were collected across all sampling sites
during the surveys (Appendix 15 and Appendix 16). Macroinvertebrates were
distributed among the taxonomic groups as shown in Table 4. The most diverse
group was Insecta with 48 taxa and representing 69% of the total number of taxa
recorded. Of the 48 insect taxa, fourteen were dipterans (true flies), eleven were
caddisflies and seven were mayflies. The next most diverse taxonomic group was
Crustacea (14 taxa) followed by Mollusca (6 taxa) and Annelida (2 taxa).
Table 4: Number of macroinvertebrate taxa recorded in each of the taxonomic groups across all sampling sites. Higher group Order / Class Common name Number of taxa Insecta Trichoptera
The number of macroinvertebrate taxa recorded from sites ranged between nine taxa
from the upper Doguru (VL3) and Vunidogoloa-Wai Koroalau (VL13) and 26 taxa
from the Nasealevu village (VL1) and Waicacuru (VL7). The Nasealevu village
waterway (VL1) supported a diverse insect fauna (22 insect taxa) while Waicacuru
supported seventeen insect fauna and six distinct crustacean fauna.
The Upper Doguru (VL3) and Vunidogoloa-Wai Koroalau (VL13) had riparian
vegetation removed (burning & cutting down of trees) and easy access to farming
areas. The Upper Doguru (VL3) site supported low taxa richness, most likely due to
changes in habitat characteristics as this site was dominated by chute habitats
supported by huge rocks and deep pools unlikely to support aquatic insects. The
Vunidogoloa- Wai Koroalau (VL13) site was next to a village with sluggish gravel
dominated uniform run habitat reflecting poor aquatic habitat conditions and
general absence of stable aquatic habitats such as run-riffle-pool sequence, woody
debris and overhanging stream bank vegetation.
The Surber samples were just taken from the riffle habitats and it was only carried
out for two sites while kick-net samples were consistent throughout the sites
covering multiple-habitats and hence kick-net data has been used for the majority of
the analysis, including taxa richness. The difference in taxa richness recorded from
the different sampling methods is shown in Figure 23.
Figure 23: Comparison of the number of macroinvertebrate taxa recorded from Kick-net and Surber Samples
Surber samples for Nasealevu village (VL1) site showed lower taxa richness (16
taxa) than kick-net samples (26 taxa) of the same site. However, Surber samples from
Waisali village (VL9) had slightly higher taxa richness (16 taxa) than the
corresponding kick-net samples (14 taxa). The Surber samples of the Waisali village
site (VL9) had an additional two insect fauna than were sampled by kick-netting.
Macroinvertebrate density A summary of the freshwater macroinvertebrates collected and their abundance is
presented in Appendix 14. The abundance is given as numbers of individuals, and is
also grouped into abundance categories as follows: very abundant (>100); abundant
(20-99); common (5-19); few (2-4) and very few (1). The overall (all taxa) abundance
ranged from 2730 individuals/m
at Waisali village site (VL9) to 4550 individuals/m
in Nasealevu village site (VL1). It is worth noting that only Surber samples (two sites
only) were used to calculate density (Appendix 15).
Insect larvae/nymphs were the most dominant taxa at all sites. This was strongly
represented by caddisfly, mayfly and dipteran larvae. This result is typical of the
headwaters of tropical inland streams. Insect larvae are well adapted to fast flowing
waters of stream/river headwaters, compared to crustaceans and molluscs which are
found in higher numbers in lower reaches of streams/rivers with swifter flows.
The small Fluviopupa (<4 mm) snails (spring snails) were also recorded as abundant
at two sites: Doguru village (VL5) and Upper Doguru (VL3). During an
opportunistic collection hand-picking), these snails were highly abundant in an
intact spring (VL14) within Savusa catchment; within the forest reserve. These
particular gastropods are usually catchment endemic and found in higher densities
in headwaters with narrow channels, swift flows and very clean water. They have
been found to be only present in streams undisturbed from cattle/horse grazing.
The damselfly nymph (Nesobasis spp.) was also abundant at two stations: Doguru
village site (VL5) and Waicacuru (VL7). They are known to be found in higher
densities in streams with overhanging vegetation, streamside root mass, open-partial
canopy shading and good water quality; hence there abundance in these streams.
The macroinvertebrate communities documented were typical of inland tropical
stream headwaters. The streams/rivers sampled provided suitable habitats for
diverse taxa composition. The sites surveyed had coarse stony streambed substrates
and a high proportion of turbulent riffle/chute habitats, which resulted in caddisflies
(Trichoptera) and mayflies (Ephemeroptera)being the most dominant group at the
majority of stations. These groups combined to give 95% (VL9), 87% (VL2), 84%
(VL5), 81% (VL11), 75% (VL4), 69% (VL10) and 62% (VL8) of the total species
recorded (Figure 24).
Figure 24: Community composition by major taxonomic group An exception to this pattern was at sites VL1, VL6, & VL13. At VL1, the Diptera
group was more abundant than the Ephemeroptera, and together with the
Trichoptera comprised 80% of species composition. At station VL6, the Crustacea
group was the second most abundant and together with the Ephemeroptera
comprised 75% of species composition. At VL13, the Diptera group was the second
most abundant, and together with the Ephemeroptera comprised 96% of species
The most abundant caddisfly taxon (Figure 25) recorded was the net-spinning filter-
feeder Abacaria fijiana. This species was most abundant in riffle habitats at Doguru
Co m m u n ity co m p o si tion (%) Sites Trichoptera
village and Nasealevu village site (VL1) where they represented between 55% and
31% of total abundance respectively. Other caddisfly larvae such as A. ruficeps, Hydrobiosis spp., Oxyethira spp. and Odontoceridae (case) were also common or
abundant but generally represented less than 9% of total abundance, except at sites
VL1 whereby Odontoceridae represented 21% and at VL9 and VL3, Oxyethira spp.
represented 15% and 10% of the total abundance respectively.
Figure 25: Macroinvertebrate community composition by taxa 0%
C o mm u n ity c o mpo si ti o n ( %) Sites Abacaria fijiana Abacaria ruficeps Anisocentropus fijianus Odontoceridae (case)
Hydrobiosis spp. Apsilochorema sp. Oxyethira spp. Trianodes fijiana Goera fijiana Pseudocloen spp. Cloeon spp. Caenis sp. Nymphula spp.
Chironomus sp. Corduliidae
Caridina spp. Muscidae
Macrobrachium spp. Ferrissia sp. Melanoides spp. Fluviopupa spp. Nesobasis spp. Physastra nasuta Unidentifiable species
Another common caddisfly recorded, the leaf-case Anisocentropus fijianus, was
present in highest proportions in the Waicacuru (VL7) and Upper Doguru 2 (VL4),
representing 27% and 22% of the total abundance respectively. Mayflies were also a
dominant taxonomic group recorded at survey sites and represented 86% of the
community in the Vunidogoloa stream (VL13) and 76% in the Waisali village stream
The most abundant mayfly taxon was Pseudocloeon spp. This is because Pseudocloeon spp.has a dorso-ventrally flattened body that allows it to graze on thin algal films
covering the surfaces of large boulder/cobble substrates in turbulent riffle/chute
habitats. In contrast, Cloeon spp. mayflies which are mostly associated with gentle
flowing habitats and are more common along stream margins and runs were
recorded in much lower proportions across the sites. Therefore many Cloeon spp. were part of the opportunistic collection. Another commonly recorded mayfly taxon
was Caenis sp. but represented just under 10% of the total abundance except at sites
Savusa- Savutagitagigagone (VL10) and Upper Dreketi (VL2), where it represented
16% and 11% of the total abundance, respectively.
Conservation status and distribution of taxa A total of six macroinvertebrate taxa recorded as part of the survey were endemic to
Fiji and represented 10% of the total number of taxa recorded. A total of 31
macroinvertebrate taxa were Endemic/native (taxa that are known to be endemic to
Fiji but the species are yet to be scientifically named) and represented 51% of the
total number of taxa recorded (Figure 26). Apart from a few unique specimens (~10),
many of the endemic taxa recorded are common throughout the headwaters of Fiji
Island streams. The remaining 39% of taxa were either native to Fiji, the Pacific or the
Indo-Pacific region, or introduced tropical species or unknown species.