Bias from false-positive detections and strategies for their removal in studies using telemetry

Abstract

The use of radio and acoustic telemetry to study aquatic animals has flourished since the 1950s and 1960s (see Section 1). Electronic data-logging receivers are commonly used in both types of active telemetry to record the presence of transmitters in the detection field formed by one or more antennas or hydrophones. As described in Sections 5.1 and 7.1, the path of a transmitter signal to a telemetry receiver can be influenced by many factors and the received signal is not always detected or correctly assigned. It should be of no surprise to users of active telemetry systems that not all records in telemetry receivers are from tagged fish and not all tagged fish are recorded when present. Four types of observations are possible in data from telemetry receiving systems based on the binary nature of presence and absence (Table 1). True positives and true negatives are what one ideally expects from telemetry systems, but in most studies they are accompanied by false negatives (not recorded when present) and false positives (recorded when absent). False negatives arise due to a variety of causes, including insufficient detection area relative to transmitter pulse rate and fish travel speed, collisions between transmitters in the detection area, interference from ambient noise, or a received signal too weak to be recorded (see Sections 3, 5, and 7). The probability of false negatives can be calculated as (1–detection probability) and can be estimated with proper study design and incorporated into estimates of fish presence (see Sections 7.2 and 9.2).