This study developed a heirarchical approach to improve the efficiency and reliability of
surveys for stream-breeding frogs in the montane forests of south-eastern Australia.
Areas with suitable climate for some of these species were first determined by
bioclimatic modelling. Landscape and habitat preferences were then determined for the
four species commonly found in far East Gippsland, along with an assessment of survey
techniques from which effective survey protocols were developed.
The climate of localities at which five stream-breeding species had been previously
found was modelled using BIOCLIM, and maps of the predicted potential distributions
having similar climate were developed. Litoria phyllochroa was found to occur in
coastal to alpine areas from Melbourne to Nowra, and potentially further inland, in areas
with relatively low temperatures and low summer rainfall. Litoria lesueuri had a
slightly larger known and predicted distribution, covering a wide range of values for the
temperature and precipitation parameters. Both species had predicted areas inland from
Melbourne and the Blue Mountains where they had never been recorded, probably
largely due to land clearing. The rare Heleioporus australiacus had a slightly more
northerly distribution, closer to the coast and adjacent ranges from Central Gippsland to
just north of Sydney, with a generally warmer climate and large range of average annual
precipitation values. Litoria citropa occupies an even narrower area along the coast,
extending slightly further north again, with a climatic profile of few extremes but with
relatively high rainfall particularly in summer. The rare Mixophyes balbus had a more
northerly known and predicted narrow range along the NSW coast, just extending into
Victoria and Queensland, characterised by the subtropical influence of higher summer
rainfall. The southern extension into far East Gippsland appears to have an extreme
sub-climate, and is hence considered highly susceptible to climate change. Future
monitoring sites for these species can now be systematically chosen to include their full
climatic range, so that any negative effects of climate change on amphibian populations
may have a greater chance of detection.
Four species were found commonly enough in the far East Gippsland study area for
more detailed study - the three stream-breeding specialist species, L. phyllochroa,
L citropa, and L. lesueuri, along with Crinia signifera which was not restricted to
streams but commonly found along them. Three species, L. phyllochroa, L citropa and
L. lesueuri, were not found at the few higher altitudes sites on the Errinundra Plateau,
however they may have been restricted by habitat requirements other than altitude.
Litoria phyllochroa was otherwise widespread, but preferred larger in-stream pools and
backwaters for breeding with overhanging vegetation for perch and call sites. The other
three species preferred stream reaches characterised by wide bedrock outcrops and
associated wide break in the tree canopy, which may allow their preferred perch sites on
rocky substrates to retain heat after sunset. Litoria lesueuri tadpoles were found in both
unconnected and connected pools, and the common association of this species with
bedrock outcrops may explain its range being largely restricted to granitic substratum
where these outcrops mostly occurred. Litoria citropa preferred reaches with bedrock
and boulder outcrops, although frogs used both rocky and vegetative substrates as perch
sites, with tadpoles caught in both unconnected pools and backwaters. Crinia signifera
tadpoles were mostly restricted to unconnected pools, with frogs generally only detected
within or very close beside these.
Habitat models varied in their predictive ability, from 67.5% for L. phyllochroa frogs to
92.7% for L. lesueuri tadpoles. Models developed from normally-distributed habitat
variables by discriminant function analysis were generally more predictive than those
from logistic regression analysis. The occurrence of frogs and tadpoles of
L. phyllochroa, L. citropa and L. lesueuri can be predicted by the measurement of five
habitat variables along a 50m reach: average width between the banks, average channel
width, an average of the maximum depth of channel cross-sections along the reach,
channel slope, and proportion of the stream length as pools. The occurrence of
C. signifera along reaches can be predicted with the additional measurement of the
proportion of the banks covered in ferns, the maximum cross-sectional channel depth
along the reach, and the number of unconnected pools.
Several different techniques for surveying frogs were compared for the four species
commonly found along the streams. Probability of detection models were derived for
each species for each technique, which allowed calculation of the statistical confidence
of detecting a species that actually occurred at a site for any given number of surveys.
For transect-based techniques, models were also developed which allowed
determination of the minimum number of surveys required without detecting a species
to be 95 percent statistically confident that the species did not occur there, for various
transect lengths. This level of confidence can be obtained for the four common species
by four surveys using the reliable night encounter technique along 500m stream
transects, during the period from Ocotber to March, with air temperatures above 10°C.
The application of habitat models to select favourable 50m reaches for survey also
generally improved the detectability of each species.
Timer-activated tape recorders, which automatically recorded the calls of frogs for many
consecutive nights at a site, were very efficient at detecting all species except the quietcalling
L. lesueuri. Night encounter surveys along stream transects detected all species
from a reasonable effort, and dip-netting for tadpoles was also effective for most
species. Active visual encounter surveys were less reliable, and day encounter surveys
detected few frogs but provided some additional data when other activities were being
performed along the streams.
The findings of this study have important implications for future amphibian surveys and
monitoring undertaken in East Gippsland and the south coast of New South Wales. If
the standardised survey techniques recommended by this study are used in
Environmental Impact Assessments, their results can be objectively assessed and
defended. The use of habitat modelling and improvement in survey reliability can also
be used more efficiently to find sites with populations suitable for monitoring. The
likely amount of effort required by a long-term monitoring program can also be
determined so as to largely overcome daily variations in the detection of each species.
| Identifer | oai:union.ndltd.org:ADTP/219026 |
| Date | January 1997 |
| Creators | Holloway, Simon, n/a |
| Publisher | University of Canberra. Resource, Environmental & Heritage Sciences |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | ), Copyright Simon Holloway |
Page generated in 0.0021 seconds