Research of Sound Speed Field Spatiotemporal Variations in the Central Philippine Basin

The Philippine Sea Basin is one of the world’s largest marginal sea basins, and the spatiotemporal variation characteristics of its sound speed field hold significant importance for deep-sea navigation and positioning as well as underwater acoustic detection. This study investigates the sound speed field in the central Philippine Basin (130.0–134.0° E, 17.5–20.5° N) using the Global Ocean Physics Analysis and Forecast product from the European Union’s Copernicus Marine Environment Monitoring Service (CMEMS), cross-validated with the U.S. HYCOM (Hybrid Coordinate Ocean Model), and independent verified against 69 Argo profiles. We systematically investigate the spatiotemporal variation characteristics of the sound speed field in this region. Temperature and salinity consistency between the two products is established (deviations of <0.5 °C and <0.05 ppt below 400 m), with CMEMS selected as the primary data source for its higher accuracy and greater temporal stability. Three sound speed formulae—Del Grosso, Chen–Millero, and TEOS-10—are intercompared, with TEOS-10 yielding the highest accuracy in cross-validation; it is therefore recommended for its rigorous thermodynamic consistency. Vertical sound speed profiles are evaluated using bi-exponential, Munk canonical, and fourth-order polynomial models. Among them, the bi-exponential model achieves the optimal balance between physical interpretability and fitting accuracy (RMSE = 2.77 m/s, inter-monthly correlation coefficient = 0.857). Its two exponential decay scales characterize the upper-ocean thermocline and the deep stratification, respectively, avoiding the physically unrealistic deep-water fluctuations exhibited by the polynomial model (RMSE = 2.68 m/s) and the poorer generalization of the Munk model (RMSE = 2.97 m/s). Horizontal gradient analysis reveals a cross-directional correlation of approximately 0.5 between sound speed gradients and ocean currents, reflecting the combined modulation of sound speed gradients by Kuroshio advection and thermodynamic stratification. The general gradient control scale is estimated at approximately 100 km × 100 km, confirmed by cross-method consistency between K-means and Gaussian mixture model clustering. Temporal analysis demonstrates that sound speed peak-to-peak variation attenuates rapidly with depth (from ~8.9 m/s at 50 m to <0.1 m/s at 4000 m), and EOF (empirical orthogonal function) analysis reveals that the first four modes explain over 99% of the total variance, with harmonic fitting identifying annual and semi-annual cycles as the dominant periodic components. Sound channel axis depth varies seasonally between 900 and 1125 m (deeper in winter, shallower in spring), with axis sound speed stable at 1480–1484 m/s (slightly higher in winter, slightly lower in spring) and axis thickness ranging from 225 to 450 m (wider in winter, narrower in spring). These results provide prior critical constraints for underwater acoustic positioning, AUV navigation, and long-range sound channel communication and navigation in the central Philippine Sea region.