Abstract
To establish an effective and accurate method for measuring the locations of breeding Northern Goshawks (Accipiter gentilis) in central Hokkaido, we first took bearings from a transmitter placed in known locations. These sites were all located in a flat area in order to avoid topological effects on bearings. The average bearing error was 7.7 degrees using this method.
We then took bearings from a radio-tagged male goshawk from two fixed receiving stations situated 2.5km apart, and the goshawk's locations were calculated by means of triangulation. These locations were compared with goshawk's actual locations based on direct observations. This method revealed that the average bearing error was 29.6 degrees and the average telemetry error was 1.9km, indicating that telemetry errors using normal triangulation methods are too great to be of use in identifying habitat use in a mosaic landscape.
A further method was employed that involved tracking a radio-tagged goshawk by car until within 50m of it, and then attempting to maintain that distance. By direct observation and triangulation within this 50m range, goshawk locations were accurate to within a few meters.
During a 20-hour study period, we were able to maintain this range for more than half of the time, indicating the effectiveness of this method as a means of taking accurate locations of goshawks.
