An automatic identification method of marine magnetic anomalies based on the sliding window correlation coefficient method

Abstract

The identification of marine magnetic anomalies has fundamental importance to the study of plate tectonics and geodynamics. The traditional approach to identifying marine magnetic anomalies is by visual method and much depends on the experience of experts. Though the identification results are generally reasonable, they lack quantitative evaluation basis. Therefore, we proposed the sliding window correlation coefficient (SWCC) method to automatically identify marine magnetic anomalies and provide a quantitative and objective evaluation for the identification results. The Pearson correlation coefficient (PCC), Spearman rank correlation coefficient (SRCC) and Kendall rank correlation coefficient (KRCC) are compared in the SWCC method. The different skewness, spreading rates and random noises to the identification results are tested. The results show that the SWCC method is most optimal for identifying fast-spreading magnetic anomalies. The absolute values of the correlation coefficients at the same sliding steps are usually PCC > SRCC>KRCC, but the resolving abilities for different polarity chrons are usually KRCC>SRCC>PCC. Applications in the southwest Pacific verified the feasibility and effectiveness of the SWCC method, and suggest that short theoretical windows usually have limited feature information; therefore, combined neighbouring polarity chrons can improve the identification results of the SWCC method.