Roberson, A.M. 2001. Evaluating and developing survey techniques using broadcast conspecific calls for northern goshawks in Minnesota. M.S. Thesis, University of Minnesota. 41 pp.
Broadcast surveys are effective at detecting northern goshawks (Accipiter gentilis) during their breeding season. Goshawk surveys are generally conducted using the alarm call during the nestling phase and the juvenile food-begging call during the fledgling-dependency phase (referred to as "standard calls"). However, goshawks are more vocal during their courtship phase than at any other time of the year. Because the probability of detecting goshawks declines after a nest has failed, a reliable means of surveying goshawks during the courtship phase would be more likely to detect nesting pairs than surveys conducted later in the breeding season. To evaluate the influence of breeding phase, distance, and call type on the probability of detecting goshawks, I estimated goshawk detection rates during the courtship, nestling, and fledgling dependency stages of the breeding season, using three conspecific calls (adult alarm call, male contact call, and juvenile food-begging call) at distances of 100, 150, 225, and 325 m from active nests in northern Minnesota. Unlike previous studies, broadcasts were conducted at only one distance per trial to better describe the relationship between distance and probability of detection and to estimate effective area surveyed per broadcast station. In 1999, 85 broadcast trials were conducted in 9 active nest areas.Detections occurred during 18.9% of broadcast trials and during at least one broadcast trial at 9 of 11 nests. In 2000, 132 broadcast trials were conducted in 16 active nest areas. Detections occurred during 48.5% of broadcast trials and during at least one broadcast trial at all 16 nests. When using standard calls during broadcast trials and when pooled over all distances, detection rates were highest during the courtship (70.4%) and fledgling-dependency phases (67.6%). Detection rates were lowest during the nestling phase (28.1%), when there appeared to be higher variation in the likelihood of response among individuals. During the courtship and fledgling-dependency phases, detection rates decreased with distance from goshawk nests, whereas during the nestling phase, detection rates were highest at 225 m from nests. Differences in detection rates among breeding phases highlight the importance of incorporating knowledge about local breeding phenology into survey design. The male contact call did not improve detection rates over the alarm and juvenile food-begging calls. Results from this study will assist in designing systematic landscape-level surveys with a known probability of detection in the Western Great Lakes Region (WGLR) and in defining a distance-detection relationship for calibrating results of extensive surveys. I used probability of detection as a function of distance in the courtship and fledgling-dependency phases to calculate the effective area surveyed per broadcast station using the survey techniques recommended in this paper. The effective area surveyed was 39.8 ha during the courtship phase and 34.4 ha during the fledgling-dependency phase. These results indicate that in Minnesota broadcast stations may be spaced 712 m and 662 m, respectively, when conducting systematic surveys during these two breeding phases. Calculation of the effective area surveyed could be applied to other regions where the probability of detection as a function of distance is known.