Bird Ecology and Conservation A Handbook of Techniques

then it may be possible to correct the incomplete samples by fitting an asymp-
totic regression.
Other approaches to this incremental analysis involve use of random instead of
consecutive locations, and estimation of stability by overlap of outlines with

Approaches| 153

Fig. 6.3The increase in range area as consecutive locations are added, estimated by
(a) minimum convex polygons on peripheral locations, (b) ellipses encompassing
99% of the utilization distribution, (c) polygons round locations grouped by cluster
analysis and (d) kernel contours for 99% of the utilization distribution. Lines show the
mean and bars give the range of estimates for four Northern Goshawks Accipiter gentilis
that were located 5–6 times daily. All methods indicate that range areas changed little
after the first 10 days of tracking.

(a)

Percentage of maximum area

Percentage of maximum area

Percentage of maximum area

(b)

(c)

(d)

MCP, 100% fixes

Jennrich-Turner ellipse, 99% density

Cluster Polygons, 99% fixes

Percentage of maximum area

Number of locations

Reference Kernel, 99% density

(^204060)
20 40 60
20 40 60
(^204060)
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0