Using principal component analysis for marine soundscape research: A case study from Taiwan

Abstract

Recent years have seen enormous increases in research on marine soundscape, as underwater acoustic sensors are becoming more affordable, coupled with a growing interest and increased concerns on underwater noise impacts on marine fauna. However, when faced with large amounts of data, there are significant challenges regarding how to analyze large datasets effectively and accurately to make informed decisions. In this study, we used principal component analysis (PCA) to analyze the temporal, spatial, and spectral patterns of the soundscape from major contributing sources in a shallow water coastal environment at two locations in Taiwan in early and late summer of 2012. Acoustic intensity levels were analyzed in three frequency bands: 150–300 Hz, 1.2–2.4 kHz, and 3–6 kHz. The PCA results show higher acoustic intensity in the 1.2–2.4 kHz band at night than during daytime hours, and relatively high acoustic intensity levels overall in the 150–300 Hz frequency at all times. In addition, a large temporal relationship was observed between early and late summer in the 1.2–2.4 kHz band. The results show that PCA has the potential to provide a simple but reliable way to quantify the dynamics of marine soundscapes.