Accurate Underwater Ambient Noise Modeling via Source Level Calibration Using a Nelder-Mead-Based Algorithm

Underwater ambient noise is a persistent background sound field in the ocean, directly influencing the performance of underwater tasks such as detection and localization. Its characteristics depend on sea surface noise sources and their propagation paths, shaped by factors such as water depth, sound speed profile, and topography. The interplay of these factors complicates the numerical modeling of underwater ambient noise. Furthermore, the incomplete understanding of sea surface noise source mechanisms presents additional challenges for the development of accurate modeling algorithms. This study develops a time-varying ambient noise model for the South China Sea and introduces the Nelder-Mead optimization algorithm to calibrate an empirical sea surface noise source model, achieving high-precision noise simulation. Experimental results confirm the superior accuracy of the proposed model.