Environmental Engineering Reference
In-Depth Information
supported by the data. If any test rejected the assertion that the data were well
described by Poisson properties it would indicate that the patterns in collisions were
not random. By testing specifi c aspects the approach would also highlight specifi -
cally which aspects (if any) of the data deviated from that expectation.
The tests were:
Graphical confi dence interval test: A visual examination of whether the 95 %
confi dence level of number of collisions each year overlapped;
Chi
2
test for differences between years;
Laplace U test: to assess whether the collision rate was constant over time.
A negative result would suggest a higher rate during early surveys, whilst a posi-
tive result would suggest the rate was higher during later surveys;
Mann test: tested for patterns in the time between events. Clumping of mortali-
ties over time might suggest a seasonal pattern or that collision events were
related; and
Wald-Wolfowitz test (WW): to test for “runs” in the time between mortality
effects. Like the Mann test, a run of longer or shorter than expected intervals may
indicate seasonal patterns or related collisions.
Any clustering in the collision data might lend support to some possible theories
for why eagles collide with turbines at these sites.
Two of these theories are:
Cascade theory. When one eagle is lost, increased territorial interactions between
remaining resident and nearby eagles results in further collisions; and
Display period theory. Eagles pay less attention to turbines when displaying
(particularly during the period at the start of the breeding season), increasing
collision risk.
Eagle Breeding Success
The purpose of these surveys was to determine if the wind farms had a disturbance
impact on eagles breeding on site (see Hydro Tasmania
2013
for a full discussion of
the history of these surveys).
Two known WTE nests at BPWF, one at SBWF, and one WBSE nest at SBWF
and seven WTE nests and four WBSE nests in the surrounding region (within
15 km) were surveyed from 2002 to 2009, inclusive. Nest sites were visited during
September and November each year to document evidence of breeding activity,
using a methodology consistent with that described in Forest Practices Authority
(
2013
) and whether there was a chick present at the latter part of the breeding of the
season. Given that there were only a limited number of “treatment” nests (those on
the wind farms) and it was not possible to increase this number, it has not been pos-
sible to design a robust survey to assess the effect of the wind farms on breeding
success rate. Further, breeding success in eagles is potentially affected by a variety
