Abstract
Passive Acoustic Monitoring (PAM) technologies are commonly used to monitor echolocating marine mammals in tidal channels, but their detection efficiency is hampered by a series of factors in high flow environments (e.g., flow noise, ambient noise) that can ultimately impede monitoring efforts. In partnership with The Pathway Program, the Fundy Ocean Research Center for Energy (FORCE) conducted an assessment for two ‘stand-alone’ (i.e., CPOD, FPOD) and three ‘conventional’ (i.e., AMAR, SoundTrap, icListen) PAM instruments to understand the operational limitations of these ‘off the shelf’ technologies. The PAM instruments were mounted to a subsea platform and deployed at the FORCE tidal demonstration site. A series of passive drifts were then conducted over the platform from a vessel across a range of tidal flow conditions while playing synthetic clicks (‘pseudo clicks’) emitted from an icTalk. This data was supplemented with that collected from real harbour porpoise transiting the FORCE site.
Pseudo clicks were insufficiently similar to real harbour porpoise click trains to be classified by either the CPOD or FPOD; limiting comparisons of these devices to real harbour porpoise clicks collected during the deployment. ‘Conventional’ PAM technologies (AMAR, SoundTrap, icListen) detected pseudo clicks over short ranges (median detection range: ~40m) due to the lower source level of the icTalk (~130 dB re 1μPa at 1m) relative to real porpoise clicks (~160 dB re 1μPa at 1m). The low source level of the pseudo clicks necessitated increasing the sensitivity of the detectors, which in turn increased the frequency of false-positive detections. True detections of pseudo clicks decreased with increasing flow speed, with few detections above current velocity of 2 m/sec. The icListen detected more pseudo clicks than the other conventional PAM technologies assessed in this study and showed a lower measurement of ambient noise at high frequencies. However, this second result may have resulted from differences in instrument calibration. Considering the detection metrics on the per-minute scale, both the icListen and AMAR had nearly identical performance. While all five PAM technologies were able to detect real harbour porpoise clicks, the false positive detection rates for the three ‘conventional’ instruments were higher than the ‘stand-alone’ instruments, creating additional post-processing steps. Reducing the sensitivity of the ‘conventional’ instruments decreased the instance of false positive detections, suggesting that further efforts on the classification of detections could reduce the rate of false positive detections while keeping recall high.
The choice of which PAM device to use depends on the scientific questions being asked. A primary objective of The Pathway Program is to establish a regulator-approved monitoring solution that can be used by tidal energy developers for monitoring the near-field (0 - 100m) region of their tidal energy device at the FORCE demonstration site. To that end, if the appropriate sensitivity settings are coupled with a very good classifier, ‘conventional’ PAM instruments could be used for monitoring tidal turbines at the FORCE site. From an acoustics perspective, the icListen and AMAR are functionally equivalent at detecting harbour porpoise at the FORCE site. However, additional considerations (e.g., costs and logistical constraints) area also important for identifying which ‘conventional’ PAM instrument to use.