Effect of delayed reporting of band recoveries on survival estimates

Brownie et al. (U.S. Fish and Wildl. Serv., Resource Publ. 131, 1978) presented 14 models based on an array of explicit assumptions for the study of survival in avian populations. These methods are replacing the life table methods previously used to estimate survival rates (e.g., Burnham and Anderson, J. Wildl. Manage., 43: 356-366, 1979). The new methods allow survival or recovery rates, or both, to be constant, time-specific, or time- and age-specific.

In studies to estimate survival rates for birds the data are often from recoveries of birds shot or found dead during the hunting season and reported to the Bird Banding Laboratory by sportsmen, conservation agency employees, or the general public. This note examines the bias in estimating annual survival due to a proportion of the recoveries being incorrectly reported a year late. Specifically, a few recoveries each year of, for example, adult male American Widgeon (Anas americana) banded in California are reported as being recovered in year i + 1 when in fact they were actually recovered the previous year i. Delayed reporting might typically be caused by people finding a band in their health clothing in the fall of the year and, being embarrassed about their failure to report the band when it was taken, report it a year late not mentioning the actual year of recovery. Heuristically, delayed reporting should bias estimated annual survival rates upwards because it appears from the data that the birds corresponding to the "delayed" recoveries actually lived an additional year.