Abstract
The Army and Air Force teamed with the Alaska Department of Fish and Game (ADF&G) and the National Marine Mammal Laboratory (NMML) to investigate the use of real-time passive acoustic monitoring for belugas. SMRU Ltd, SMRU, LLC and Marine Instruments Ltd were contracted to conduct a trial of their PAMBuoy® system in Eagle River as part of a multi-year effort led by ADF&G and NMML to conduct applied passive acoustics research on Cook Inlet belugas (CIBA project).
A PAMBuoy® detection system was deployed in the mouth of the Eagle River, Alaska between 12 and 31 August 2012, close to the point at which the river flows out into Eagle Bay. The system was configured to detect both the echolocation clicks and the whistles of beluga whales (Delphinapterus leucas) known to be regular visitors to both the Bay and the River at that time of year.
The PAMBuoy® system utilizes a low power digital signal processing system to automatically detect certain types of sound and then sends that information in near real time over a wireless communications system to a remote base station.
The primary purpose of the deployment was to assess how effectively the system can detect beluga whales with the eventual aim of setting up a system which can automatically alert staff at the nearby Joint Base Elmendorf-Richardson (JBER) military base when animals approach or are in the river.
The system was effective at detecting both clicks and whistles during the trial period, successfully detecting all (100%) sightings from within the river and 430 out of all 432 (99.5%) recorded sightings at ranges varying from 10 to 1035m from both within and from outside the river. A negligible false alarm rate and high efficiency could be achieved by requiring multiple click detections within a short time window (e.g. 20 clicks in 10s). These results indicate that the system efficiency is more than adequate for the primary purpose of detecting whale presence within or close to the river.
When comparing individual detection events detected by the fully automatic system with those picked out by an experienced human operator, the automatic system was unable to perform as well as the human, successfully detecting only 75% of human detected events. However, these missed events were generally short, relatively quiet click trains which were often close in time to louder events, presumably from the same beluga encounter, which were successfully detected, thus not missing encounters. A recall of 98% with 98% precision was achieved using a 15 minute encounter interval.
A future long term monitoring program may require two PAMBuoy® systems, one deployed at the mouth of the river and one further upstream if there is a need to distinguish between whales in the bay, close to the river and whales in the river system itself. A second system was deployed outside the river mouth for a single day and the two systems operated simultaneously. Animals can clearly be detected on the system outside the river mouth, a short time before being detected in the river.
While the main PAMBuoy® system can be considered to be at a high level of “Technology Readiness” a long term monitoring installation would require some work to be put into the wireless communications system in order to transmit the signal all the way to the joint base and considerable investment may also be needed to develop an ice and debris proof mounting system for the hydrophone if monitoring is to take place for extended periods and during winter months.
Once information on costs and practicalities of long term hydrophone mounting and data transmission become available, the final decision as to hydrophone configuration and placement and operational alarm thresholds will require detailed discussions with the relevant representatives from the joint army and air force base and regulators before a final system can be specified.