Introduction & Objectives

In the Pacific the Marine Mammal Monitoring on Navy Ranges (M3R) program maintains systems that automatically detect, classify and localize marine mammals in real‐time on the U.S. Navy’s deep‐water Southern California Antisubmarine Warfare Range (SOAR) in Southern California and Pacific Missile Range Facility (PMRF) off Hawai’i.  Long‐term archive data collected on these ranges allows for numerous types of studies on species inhabiting the ranges, including the monitoring of abundance and distribution, behavioral responses to naval activities, and habitat usage.  They also provide the opportunity to study ambient noise and soundscapes.  

The goals of the monitoring effort included the following: 

1. Collect M3R archives at both SOAR and PMRF to inform long‐term distribution and abundance estimates for Cuvier’s beaked whales (Ziphius cavirostris, Zc) and Blainville’s beaked whales (Mesoplodon densirostris, Md)

2. Analyze the accuracy of the installed low‐frequency detector algorithm at SOAR in coordination with the Naval Information Warfare Center (NIWC).

3. Validate the “sprinkle analysis” method for ambient noise calculation using the Fast Fourier Transform (FFT)‐based archives by comparison to ambient noise calculated from broadband recordings by NIWC

4. Support real‐time monitoring of on‐water tagging operations at SOAR and PMRF

The long‐term detection archives recorded at SOAR and PMRF allow for analysis of trends in detections over time.  Changes in relative detections could indicate changes in foraging behavior, changes in prey quality or density, or changes in animal abundance.  Passive acoustic methods of calculating abundance allow for the relatively low‐cost collection of archive data to provide insights on populations of species on Navy ranges. 

Technical Approach

The M3R system runs nearly continuously year‐round, archiving data from all range hydrophones simultaneously in real‐time, when there are no range activities that would preclude its operation. Detection, classification, and localization (DCL) reports are stored to binary archive files for later playback and analysis.   The M3R system employs three detector/classifiers:  a Fast Fourier Transform (FFT)‐based detector, a Class‐Specific Support Vector Machine (CS‐SVM) detector/classifier, and a Blainville’s beaked whale foraging click matched filter. The CS‐SVM classifier currently has four classes at SOAR:  Cuvier’s beaked whale foraging and buzz clicks, sperm whale clicks, and ‘generalized dolphin’ clicks. At PMRF there are six CS‐SVM classes: Blainville’s beaked whale foraging and buzz clicks, Cuvier’s beaked whale foraging and buzz clicks, sperm whale clicks, and ‘generalized dolphin’ clicks. Archives which included FFT‐based detections were first collected at SOAR in 2006, when the system was first installed.

Progress & Results

The M3R program had four areas of focus for SOAR and PMRF: 

1. Long‐term data collection and the evaluation of the distribution and abundance of Cuvier’s beaked whales at SOAR and Blainville’s and Cuvier’s beaked whales at PMRF. 

2. Accuracy analysis of the M3R low‐frequency detector algorithm at SOAR in coordination with the Naval Information Warfare Center (NIWC).

3. Validation of the “sprinkle analysis” method to extract ambient noise from M3R archives by comparison with broadband recordings analyzed by NIWC. 

4. Support of on‐site field exercises at SOAR and PMRF with real‐time monitoring using the M3R system.