Research on underwater ambient noise prediction in the Eastern Indian Ocean based on the LightGBM algorithm

Ocean ambient noise is a type of interference background field in underwater acoustic channels, which significantly affects the analysis of ocean ambient characteristics, as well as ocean ambient noise modeling and forecasting. It is also one of the main factors limiting active and passive sonar systems. Therefore, monitoring and accurately predicting ocean noise is crucial. Based on in-situ ocean ambient noise measurements from the Eastern Indian Ocean and meteorological data from publicly available reanalysis databases, this paper analyzes the impact of ocean ambient parameters on ambient noise. A noise prediction method based on the LightGBM (Light Gradient Boosting Machine) algorithm is proposed and validated. The results show that above 400Hz, the correlation coefficient between wind speed and ocean ambient noise spectrum levels exhibits minor fluctuations but remains generally stable, with values consistently greater than 0.52. A LightGBM-based ocean ambient noise prediction model was established, and validation results indicate that in the 400-10000Hz range, the root mean square error between predicted and measured values is less than 1dB. This method can achieve accurate noise spectrum level predictions.