Suchergebnisse

Some factors that may affect the aerial counts of dugongs, dolphins and turtles were examined experimentally.
Tidal influence did not affect the counts of dugongs or dolphins, but those of turtles increased around high
tide. A combined doubling of survey altitude and transect width reduced observed density of all three classes
of animals by 50%. The counts of four observers did not differ significantly, but further data counsel caution.
Overcast weather depressed counts of dugongs and turtles. Dolphin counts were affected by water surface
condition, counts being lowest in choppy seas. A mark-recapture model was used on tandem observations
to derive correction factors for groups of animals missed on the surface of a transect. Observers missed between
33% and 75% of dugong groups on the surface, the probability of detection decreasing with increased survey
altitude and transect width. A similar range and pattern of probabilities was found for dolphins and turtles.
Dugongs were censused in the coastal waters of the Northern Territory between the Daly River and Millingimbi
in December 1983, an area of 28 746 km*2. Sampling intensity was 7.6%. A minimum population of
2953 � 530 (standard error) was estimated, an overall relative density of 0.11 � 0.02 km-2. A theoretical
correction for submerged dugongs not seen yielded a total population estimate of 38 000, an overall density
of 1.46 km-2. The distribution of dugongs in the survey area was patchy, the highest densities being
associated with shallow coastal waters, sheltered bays, and large islands.

Feral and domestic livestock (buffalo, cattle, horses and donkeys) were censused by fixed-wing aerial survey over the 'Top End' of the Northern Territory (above the 16th parallel) in March-April 1985, encompassing 233 672 km2 (16.7% of the Territory). The mean sampling rate on the major coastal lowland (38 281 km2) was 7.3%, and that for the rest was 3.7%. The accuracy of population estimates was improved by applying habitat-specific visibility correction factors to animals counted in groups.

The cane toad (Bufo marinus) has received considerable attention because of its rapid spread in Australia and
the potential threat it may represent to native species. Although the introduction of pathogens from native
populations is now being considered to control this species, population estimates based on comparable
methods that demonstrate that native populations are in fact less dense than the introduced ones are not
available. Accurate population estimates are necessary to evaluate potential techniques for the control of the
cane toad. We estimated population densities of cane toads over a wide range of habitat types and climate
conditions by means of mark-recapture data. The capture history and location of toads each night were
analysed to explore the validity of some of the assumptions of mark-recapture models. Because migrations,
deaths and recruitment over three nights appear to be unimportant, populations may be legitimately regarded
as closed for that period. However, cane toads seem highly sensitive to disturbance effects due to trapping
andlor handling. Consequently, density estimates based on removal methods seem the most reliable because
they are not sensitive to handling and trapping effects. Similarly, analyses of residuals of regressions between
1-night counts and density estimates suggested that toad nightly activity is affected by the air temperature
during sampling. However, only 60% of the variation in estimated densities can be predicted by 1-night
counts and air temperature. Estimates of population density over a wide range of habitats in South America
were one order of magnitude lower than estimates in Australia. We speculate on the possible factors that may
account for the lower densities in populations in the native range of the species.

The effect of harvest of magpie goose clutches on nest numbers and juvenile recruitment was investigated
experimentally within a single breeding colony, by comparing the densities of nests and offspring (eggs
and goslings) produced in harvested and unharvested blocks in each of two successive years. When
adjusted for initial differences in nest density, by covariance analysis, there were significantly higher
densities of nests in harvested blocks than in unharvested blocks, possibly due to re-laying in new
nests by those geese whose clutches were harvested. There were no differences in densities of offspring
produced between harvested and unharvested blocks although there were differences in densities of
offspring produced between years.

Aerial surveys of coastal waters (27 216 km2) in the western Gulf of Carpentaria during the dry season
of 1984 and wet season of 1985 indicated no major seasonal changes in distribution and relative
abundance of dugongs. Minimum population size in the dry season was estimated as 16 816 � 2946
(standard error), with a relative density of 0.62 k 0.11 km-2, and that for the wet season 16 846 + 3257,
with a relative density of 0.62 � 0.12 km-2. The estimates exclude 5% of observations which could
have been either dugongs or Irrawaddy dolphins, and were corrected for submerged dugongs and those
missed on the surface. Dugongs were unevenly distributed over the study area, and neither mean group
size nor proportion of calves varied between seasons. Dugong abundance was correlated with area of
available seagrass. The catch rate of dugongs by Aboriginal people off Numbulwar decreased six-fold
between the 1960s and 1985 (60 to 10 p.a.), possibly due to excessive hunting.

Results of mail-card surveys of registered shotgun owners in the Northern Territory during 1984 and 1985 are
reported. Responses were used to generate estimates of seasonal and daily bags of waterfowl, the level of
hunter activity, and a crude estimate of hunter numbers. In 1984 the mean seasonal bag was 73.2 birds taken
over 10.3 hunter-days, which fell to 21.4 birds taken in 6.3 hunter-days during the 1985 season. The magpie
goose was taken most frequently, constituting 84% of 1984 and 79% of 1985 bags. Hunting activity was
concentrated in the Darwin area at small, readily accessible sites. Estimates of total seasonal harvests are
derived to provide an indication of the scale of hunting mortality. All estimates are based on relatively low
returns and improved targeting of surveys is desirable to increase sample size and reliability.

Annual aerial counts of kangaroos within randomly selected blocks of the western plains of New South
Wales showed that the numbers of kangaroos doubled between 1975-76 and 1982, and that the
widespread drought of 1982 reduced the populations on average by 43%. Localized reductions of similar
magnitude occurred after regional droughts in 1977 and 1980 within parts of the monitored area. The
observed trends in kangaroo numbers, with eastern and western blocks treated separately, were
correlated with annual rainfall with a time lag of 6 months in the response. The relationships show that
kangaroos reach their maximum rate of increase following rainfall 100 mm above the annual average in
the east and approximately 50 mm above the annual average in the west. At average annual rainfall
kangaroos increase at 25% (greys) and 35% (reds) per annum in the east and at 25% (greys) and 30% (reds)
per annum in the west. Rate of increase is zero when rainfall is 100 mm below average in the east and
approximately 60 mm below average in the west. When rainfall is below these values, kangaroo
numbers decline.

A mark-recapture experiment was used to estimate the total population of crocodiles in four 10-km sections
of the Adelaide River; a downstream and upstream section of the mainstream, and two sections of tidal
side creeks. The Petersen Estimate yielded 135 � 5 (SE) crocodiles in the downstream river section, 54 � 6
in the upstream section, and 44 � 2 in the two side creeks combined (approximately 22 crocodiles per
10 km of side creek). For the two mainstream sections and the combined creek sections, respectively,
66%, 59%, and 35% of crocodiles present were sighted in a typical spotlight survey. The probability of
sighting a crocodile decreased with increasing density of bank vegetation, with decreasing stream width,
and with increasing stream sinuosity. Detection of crocodiles in the downstream river section was also affected
by size-related wariness: the probability of detecting hatchlings was large (P = 0.69), and for crocodiles
longer than 3.0 m, small (P = 0.15).
Diurnal counts of crocodiles from a helicopter were calibrated to spotlight counts from a boat at low
tide in a 40-km mainstream section and four tidal side creeks of the Adelaide River system. In river habitats
with large exposed mudbanks, aerial observers missed 20-33% of sightable crocodiles in their field of view;
and in side creeks lined with dense mangroves, 33-75%. When corrected for this observer bias, helicopter
surveys in the mainstream at low spring tide produced similar counts to spotlight surveys. In the mainstream,
helicopter counts at low neap tide were 37% lower than spotlight counts at low spring tide, but in side
creeks were 45% higher. In comparison with spotlight counts, only 18% and 29% of crocodiles were seen .
from a helicopter in two densely vegetated freshwater billabongs. A feasibility survey showed that in remote
areas helicopter surveys produced counts of crocodiles similar to those by spotlight surveys from boats,
and cost one-quarter as much ($7 km-1 compared with $28 km-1 of habitat surveyed).