An Acoustic Propagation‐Guided Scaling Clustering Method of Ocean Sound Speed Gradients and Environmental Effect Analysis

The use of unsupervised clustering methods to classify hydroacoustic profiles with similar structures in the study area, and subsequently divide them into distinct marine regions based on acoustic propagation characteristics, facilitates targeted research on the environmental effects of underwater acoustics in different sea areas. However, many existing methods are limited to clustering based solely on sound speed magnitude, which leads to clusters with limited practical utility and poor robustness across different marine regions. This paper proposes a scaling strategy for ocean sound speed gradients based on acoustic propagation characteristics, combined with a clustering algorithm from machine learning. Clustering analysis was conducted using the climatic sound speed field data from the World Ocean Atlas 2023 (WOA23) North Indian Ocean region, and the Bellhop ray‐tracing model was employed to simulate acoustic propagation effects under different environmental conditions. The results indicate that the sound speed profiles in the study area fall into two broad categories: shallow‐sea and deep‐sea, each of which can be subdivided into five distinct classes based on their sound propagation characteristics. Our method shows enhanced accuracy and provides more intuitive guidance for environmental acoustics compared to conventional clustering techniques.