Background Acoustic Noise Models for the IMS Hydroacoustic Stations
Abstract
Background acoustic noise levels in the ocean have been increasing for the past several decades yet many of our hydroacoustic detection assessment tools use noise models based on data from the 60's and 70's. In some ocean basins, noise levels in the monitoring band (1-100 Hz) have risen 15 dB since the 1960's. To address this issue and provide accurate noise models at each of the six International Monitoring System (IMS) hydroacoustic stations, noise models are constructed using historical data from the stations, many now in operation for over 5 years. The analysis procedure consists of computing a power spectral density (PSD) curve for each 2-hour time period and for each hydrophone sensor (28 in all) over the entire archived data history of the stations. There are nearly 20,000 2-hour spectra for some stations. The PSD's are instrument corrected, converted to units of dB relative to 1 micropascal, and accumulated in 1 dB wide bins at each 0.1 Hz increment for each individual hydrophone. This results in a "noise model" matrix for each sensor that can be viewed as hydrophone noise histograms for each 0.1 Hz increment from 1 to 100 Hz. The noise model becomes a probability density model by simply dividing the matrix by the total spectra count. The noise model is used to create maximum probability curves and 90% confidence curves for each sensor that can then be utilized as background noise levels in network capability assessments. The noise models do not support or refute that acoustic noise levels have risen significantly since the stations do not have a long history of measurements to compare with. They do show that noise variation between stations is significant and complex. The noise models document the existence of persistent noise sources at most stations as well as some notable differences in sensor noise within triads. Besides serving as input to network assessment codes, these noise models can also help track and assess the system health of individual sensor
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2010
- Accession Number
- ADA569498
Entities
People
- Philip E. Harben
- Terri F. Hauk
Organizations
- Lawrence Livermore National Laboratory