Annual survival of Snail Kites in Florida: Radio telemetry versus capture-resighting data

We estimated annual survival of Snail Kites (Rostrhamus sociabilis) in Flor ida using the Kaplan-Meier estimator with data from 271 radio-tagged birds over a three-year period and capture-recapture (resighting) models with dat a from 1,319 banded birds over a six-year period. We tested the hypothesis that survival differed among three age classes using both data sources. We tested additional hypotheses about spatial and temporal variation using a c ombination of data from radio telemetry and single- and multistrata capture -recapture models. Results from these data sets were similar in their indic ations of the sources of variation in survival, but they differed in some p arameter estimates. Both data sources indicated that survival was higher fo r adults than for juveniles, but they did not support delineation of a suba dult age class. Our data also indicated that survival differed among years and regions for juveniles but not for adults. Estimates of juvenile surviva l using radio telemetry data were higher than estimates using capture-recap ture models for two of three years (1992 and 1993). Ancillary evidence base d on censored birds indicated that some mortality of radio-tagged juveniles went undetected during those years, resulting in biased estimates. Thus, w e have greater confidence in our estimates of juvenile survival using captu re-recapture models. Precision of estimates reflected the number of paramet ers estimated and was surprisingly similar between radio telemetry and sing le-stratum capture-recapture models, given the substantial differences in s ample sizes. Not having to estimate resighting probability likely offsets, to some degree, the smaller sample sizes from our radio telemetry data. Pre cision of capture-recapture models was lower using multistrata models where region-specific parameters were estimated than using single-stratum models , where spatial variation in parameters was not taken into account.