Research on mesoscale eddy-tracking algorithm of Kalman filtering under density clustering on time scale

Abstract

This article proposes the tracking algorithm based on density clustering of time scale and mesoscale eddy of Kalman filtering using the fused SLA data of altimeter. Firstly, the definitive density clustering based on time scale discovers the potential association pattern between data, and screens out the data set of mesoscale eddy trajectory. With regard to the data set with time scale conflict, it analyzes the Kalman filtering, eliminates the noise points and obtains the correct mesoscale eddy trajectory. Secondly, it turns the tracking process into an algorithm that supports batch processing by applying the data processing method to the mesoscale eddy-tracking algorithm, which solves the problem of single serialization and high time and space complexity of the traditional tracking algorithm. Based on the algorithm, this article selects the experimental data of the South China Sea for the mesoscale eddy-tracking test. The experiment turns out that the algorithm can better reveal the life course of mesoscale eddy and evolution rule of physical oceanography according to spatial scale, amplitude and eddy duration, etc.

Keywords

• mesoscale eddy
• tracking
• density clustering
• batch processing
• time complexity