| Summary: | Estimating predation rates on juvenile salmonids is challenging due to the scarcity of reliable data. Opportunities to address the data gap now exist as new sensors are incorporated into telemetry packages that can be deployed across the Salish Sea. Here we report on the development of a field ready sampling scheme for a novel, head-mounted radio frequency identification (RFID) reader designed to detect tagged juvenile salmon as they are consumed by pinnipeds. Feeding trials were conducted in a controlled setting over 8 weeks with a wild-caught harbor seal (Phoca vitulina) and live coho salmon smolts (Oncorhynchus kisutch) carrying passive integrated transponder (PIT) tags. We calculated the passage rates of 4 PIT tag types to maximize fish detections and explored smart sub-sampling techniques to minimize costly RFID field power consumption by logging PIT IDs (8 Hz) and measuring head-strike acceleration (250 Hz) during prey capture attempts. Analyses of the archival instrument revealed that 100% of consumed smolts were detected (n = 505), that time in the RFID field was unrelated to smolt length (R2 < 0.20), and that passage rates were significantly different for each PIT tag class (F3, 501 = 317, p <0.001). Notably, PITs were detected for significantly shorter durations (means < 0.35 sec) during unsuccessful captures (n = 200, p < 0.001) suggesting that false positives may be easily screened from field data. By recording prey capture attempts in a laboratory setting, we developed a behavior based sampling scheme that would significantly extended future field deployments while maximizing the likelihood of detecting consumption of juvenile salmon. We also demonstrated that acceleration based triggers could be useful to activate a variety of on-board sensors or to moderate sampling rates for instrumented free-ranging marine predators.
|
|---|