Classifying and inferring behaviors using real-time acceleration biotelemetry in reproductive steelhead trout ().

Movement behaviors are central to ecology and conservation. Movement sensing technologies can monitor behaviors that are otherwise difficult to observe under field conditions and may enhance the ability to quantify behaviors at the population scale. We monitored steelhead trout () spawning behaviors in a seminatural enclosure using accelerometer telemetry tags while simultaneously observing behaviors with underwater cameras. Behavioral assignments from visual observations were compared to acceleration histories to develop assignment criteria for acceleration data, including for a key behavior (oviposition). Behavioral events independently classified using acceleration data prior to reviewing video were compared to video scoring and 97% of holding behaviors, 93% of digging behaviors, and 86% of oviposition/covering behaviors were correctly assigned using acceleration data alone. We applied the method to steelhead in spawning tributaries. Acceleration records revealed putative spawning and oviposition in female steelhead, and time budgets for steelhead were similar to those monitored within enclosures. The use of similar movement sensing tags and classification approaches offers a method for monitoring movement behavior, activity budgets, and habitat use in a broad array of aquatic and terrestrial taxa, and may be especially useful when behaviors are cryptic.