PROJECTS
Introduction & Objectives
Passive acoustic monitoring using autonomous recorders deployed on the seafloor is an effective method for long-term monitoring of marine mammals (Mellinger et al. 2007, Van Parijs et al. 2009). Autonomous recorders have been used to investigate the distribution, occurrence, and acoustic behaviors of a variety of marine mammals in diverse habitats and geographic locations (Clark et al. 2002, Clark and Clapham 2004, Baumgartner et al. 2008, Johnston et al. 2008, Sousa-Lima et al. 2013). Recently, researchers have analyzed autonomous recorder data to investigate the effects of anthropogenic noise, such as seismic airguns and sonar, on the calling behavior of baleen whales (Nieukirk et al. 2004, Di Iorio and Clark 2010, Castellote et al. 2012, Melcón et al. 2012, Risch et al. 2012). Changes in vocal behavior in response to anthropogenic noise have been studied in several species of odontocetes. For example, beluga whales (Delphinapterus leucas), killer whales (Orcinus orca)and Pacific humpback dolphins (Sousa chinensis)have been documented to change call rates and time‑frequency characteristics of their calls in response to vessel noise (Au et al. 1985, Lesage et al. 1999, Van Parijs and Corkeron 2001, Foote et al. 2004). Much less is known about the behavioral responses of odontocetes to mid-frequency active sonar (MFAS).
This study is a collaborative effort involving researchers at Cornell University, Bio-Waves, Inc., and St. Andrews University to develop robust statistical methods that can be used to analyze vocal behavior before, during, and MFAS activities on a species-by-species basis when possible. The primary objectives were to develop analytic methods to evaluate vocal responses based on passive acoustic monitoring techniques, and evaluate behavioral responses by cetaceans exposed to MFAS from U.S. Navy training and testing activities.
Technical Approach
In order to further investigate potential changes in behavior in response to MFAS exposure, sounds of marine mammals were recorded in the vicinity of three U.S. Navy training exercises before, during, and after the use of MFAS in Onslow Bay, North Carolina (July 2008), and near Jacksonville, Florida (September–October, and December 2009). The resulting passive acoustic data were analyzed using automated signal-detection software to detect individual sonar transmissions (“pings”), and sounds from a variety of cetaceans including striped dolphin , short-beaked common dolphin, bottlenose, Atlantic spotted dolphins, short-finned pilot whales, sperm whales, minke whales, fin whales, and North Atlantic right whales. Statistical analysis and methods development was performed using generalized estimating equations (GEEs), and the other used hidden Markov models (HMM).
Progress & Results
Initial methods development has been completed. The results obtained from the initial methods development and analyses provide a good foundation for the further development of statistical methods for evaluating changes in cetacean acoustic behavior in response to MFA sonar, however larger sample sizes and the inclusion of covariates representing a cumulative effect of sonar on vocalization behaviors are necessary to make these analyses more robust. Methods and results are detailed in two final reports and continued model development and analysis is being conducted through the incorporation of additional data. A final report covering the results from this continued work is now available.
Project Information
Timeline
2014-2016
Location
Cherry Point and Jacksonville Range Complexes
Methods
PAM
Performing Organizations
Principal Investigators
Julie Oswald - BioWaves, Inc
Russ Charif - Cornell University
Len Thomas - University of St. Andrews
Sponsor(s) & Funding
FY13 - $335K, FY14 - $50K, FY15 - $45K, FY16 - $50K
Point of Contacts
Project manager
Dan Engelhaupt
HDR Inc
Program Manager
Joel T. Bell
NAVFAC Atlantic
Environmental Conservation, Marine Resources Section
Data, Reports, & Publications
Technical Reports
+
Increasing Sample Size for Examining Changes in Vocal Behavior in Relation to Navy Sonar Activity.
Oswald et al. 2016.
More details
Collapse details
Type:Ocean:Region:Species:Methods:
The aim of this study was to develop and refine robust statistical methods to evaluate the effects of Navy sonar on the acoustic behaviors of cetaceans. Data collected using High-frequency Acoustic Recording Packages (HARPs) deployed off Jacksonville, Florida (n=1), Onslow Bay, North Carolina (n=2), and Cape Hatteras, North Carolina (n=1) were combined with data previously analyzed from Marine Autonomous Recording Units (MARUs) off Jacksonville and Onslow Bay (Oswald et al. 2015).
Development of Statistical Methods for Examining Relationships Between Odontocete Vocal Behavior and Navy Sonar Signals.
Oswald et al. 2015.
More details
Collapse details
Type:Ocean:Region:Species:Methods:
The first step in this analysis was to identify delphinid acoustic encounters to be used as inputs in the species classification analysis. A delphinid acoustic encounter was defined as a continuous portion of a recording with no more than a 30 minute (30 min) gap between sounds produced by delphinids. Delphinid sounds were defined as one of three mutually exclusive types: 1) whistles, 2) clicks or 3) burst pulses. ROCCA was used to extract and measure whistle features and to classify encounters containing whistles to species. The second step in this study involved the detection and characterization of MFA sonar events. This was conducted using the program SonarFinder (Bio-Waves 2013). SonarFinder is a Matlab-based program designed to automatically detect sonar pings and measure their acoustic features. Sonar events were defined as a series of sonar pings with no more than a 30 min gap between individual pings
Development of Statistical Methods for Assessing Changes in Whale Vocal Behavior in Response to Mid-Frequency Active Sonar.
Charif et al. 2015.
More details
Collapse details
Type:Ocean:Region:Species:Methods:
In order to further investigate potential changes in behavior in response to MFAS exposure, sounds of marine mammals were recorded in the vicinity of three U.S. Navy training exercises before, during, and after the use of MFAS in Onslow Bay, North Carolina (July 2008), and near Jacksonville, Florida (September–October, and December 2009). The resulting passive acoustic data were analyzed using automated signal-detection software to detect individual sonar transmissions (“pings”), and sounds of North Atlantic right, minke, fin, and sperm whales. In addition, putative right whale “gunshot” sounds that had been detected in an earlier, separate analysis of the Jacksonville recordings were reviewed by two experienced marine mammal acoustic analysts.
Publications
+Presentations
+Associated Projects & Blog Posts
Blog Posts
+