A numerical approach for understanding the temporal and spatial correlation scales of sound velocity in long term: a case study for the source and surrounding region of the Kuroshio Current

Abstract

To accurately understand the temporal and spatial distributions of sound velocity fields and their relevant characteristics, numerically investigated the sound velocity auto-correlation index at the source and adjacent region of the Kuroshio Current using 1911-2001 observation data. Auto-correlation index was fitted using a Gaussian function for the temporal and spatial scales of the investigated regions. The correlation scales in time and spatial domain in spring, summer, fall, and winter were 14.2/145.3, 14.2/135.9, 13.5/133.8, and 13.7/143.3 days/km, respectively. These scales are applied to the reconstructed sound velocity model in terms of empirical orthogonal functions. The error of the sound velocity constructed was minimal (3 m/s) compared with those constructed using other methods (up to 4.3 m/s). This method improves the precision of the sound velocity reconstruction compared with that based on monthly data. The use of temporal and spatial scales enhances the accuracy of sound velocity reconstruction.