Quantifying Uncertainty in Aquatic Telemetry: Using Received Signal Strength to Estimate Telemetry Error

Abstract

Telemetry error is not regularly considered in aquatic studies; however, when it is considered, it is generally treated as a static value despite high variation that can occur even within a single tracking event. We describe a simple procedure for using received signal strength (RSS) to estimate telemetry error. We recorded the RSSs of ground- and air-based detections of large (11 g; 11 × 59 mm) and small (8 g; 11 × 43 mm) 172-MHz radio transmitters across a range of distances. Received signal strength was an excellent predictor of the distance to transmitter for ground tracking detections (large transmitter: r2 = 0.98; small transmitter: r2 = 0.97) and a fair predictor for aerial detections (large transmitter: r2 = 0.49; small transmitter: r2 = 0.57). We also manipulated transmitter antenna lengths and unexpectedly found that both transmitters performed better with shorter antennas relative to factory lengths. With calibrated models relating RSS to the distance to transmitter, RSSs from field-collected data can be used to approximate telemetry error and draw spatial confidence areas around location estimates for analysis and interpretation. During a concurrent movement study, we estimated telemetry error for 2,436 detections of fish at large. Ground tracking error estimates ranged from 1 to 131 m (median = 24 m), and aerial error estimates were most often less than 300 m but were as high as 1 km. The benefits of representing telemetry data as spatial confidence areas guided by the RSS of each detection are discussed. With appropriate caution, this method will provide a more robust alternative to the assumption that error is constant, negligible, or both.