Estimating detection probabilities of tagged fish migrating past fixed receiver stations using only local information

Abstract

We developed a method to predict the probability of detecting acoustic tags crossing a receiver station using only detection information at that station. This method is suitable for acoustic or radio telemetry studies in which individually tagged animals migrate past fixed stations (where a station may consist of one or more receivers). It is based on fitting attenuation models to sequences of detections and missed transmissions of individually coded tags in fish migrating past stations of the Pacific Ocean Shelf Tracking Project (POST). We used estimated attenuation model parameters from detected fish at each station to predict the number of fish that crossed the station undetected, which in turn was used to calculate the local detection probability. This estimator was correlated (r = 0.54–0.81 in river and coastal habitats) with mark–recapture estimates of detection probability (pmr) that use nonlocal detection information at stations further along migration routes. This local detection probability estimate can be used as a covariate of pmr in mark–recapture models and can predict approximate values of pmr at final detection stations where pmr is not estimable because of the lack of recaptures further along migration routes.