Detecting capture-related mortality in radio-marked birds following release

摘要

A fundamental assumption of avian survival analysis is that the act of capture, handling, and marking birds does not affect subsequent survival. This assumption is violated when animals experience injury, physiological stress, or disorientation during capture and handling that increases their mortality risk following release. Such capture-related effects must be accounted for during analysis, typically by censoring individuals from the survival history, to avoid biasing the resulting survival estimates. We reviewed studies of radio-marked upland game birds to characterize researcher approaches for addressing short-term effects of capture on survival, and used data from a study of Ruffed Grouse (Bonasa umbellus) to illustrate an empirical approach for evaluating such effects and identifying time thresholds to censor individuals that die shortly following release. A majority of studies (65%) reported using some form of censoring for mortality that occurred within one to three weeks after release, although only 8% of studies reported an empirical approach to identify a threshold for censorship. We found that Ruffed Grouse mortality was greater from one to six days following release when compared with 7 to 30 days. This threshold, and the proportion of birds censored as a result of it, is consistent with a number of previous studies of radio-marked Ruffed Grouse. We also found that short-term mortality of Ruffed Grouse following release was reduced by checking traps twice each day and by adequately concealing traps. We recommend that future studies of radio-marked birds employ empirical methods for detecting postrelease mortality thresholds, which will allow for reduced bias while minimizing unnecessary censorship of birds that die for reasons unrelated to capture.