Parameterization of the oceanic ambient noise using orthogonal decomposition

Area of Science:

• Oceanography
• Acoustics
• Data Analysis

Background:

• Ambient noise in the open ocean is crucial for various marine applications.
• Traditional methods for processing and predicting ocean noise face challenges with data gaps and outliers.

Purpose of the Study:

• To assess the utility of Empirical Orthogonal Function (EOF) decomposition for processing and predicting ambient noise in the open ocean.
• To evaluate EOF-based methods for data gap filling and noise spectrum prediction.

Main Methods:

• Empirical Orthogonal Function (EOF) decomposition applied to ambient noise observations in the Southern Ocean.
• Singular Value Decomposition (SVD) for analysis and pre-processing of noise data.
• Comparison of EOF-based gap-filling with linear interpolation.

Main Results:

• The first three EOFs explain 89%, 96%, and 98% of the ambient noise spectrum, respectively.
• EOF-based gap-filling achieved approximately 35% higher accuracy than linear interpolation.
• EOFs reduced regression parameters for noise prediction, enabling computationally efficient, continuous spectrum prediction.

Conclusions:

• EOF decomposition is a powerful tool for analyzing and predicting ocean ambient noise.
• EOF-based methods offer significant improvements in data processing and prediction accuracy.
• Accurate prediction of frequency-averaged noise magnitude is key to improving spectral prediction accuracy.