Characteristics of Global Ocean Abnormal Mesoscale Eddies Derived From the Fusion of Sea Surface Height and Temperature Data by Deep Learning
Yingjie Liu, Quanan Zheng, Xiaofeng Li
Abstract
Abstract Recent satellite sea surface height (SSH) and sea surface temperature (SST) observations have shown that abnormal eddies, that is, warm cyclonic eddies and cold anticyclonic eddies occur sporadically in some regions, which triggers an essential question on the spatiotemporal distribution of abnormal eddies in the global ocean. In this study, a deep learning framework was developed to systematically mine information from the synergy of satellite‐sensed global SSH and SST data over the 1996–2015, 20‐year period. Abnormal eddies account for a surprising one‐third of total eddies and are active along the Equatorial Current and high unstable currents. Normal (abnormal) eddies are stronger in winter (summer) in the North Hemisphere and vice versa in the Southern Hemisphere. The annual mean amplitudes of normal eddies are larger than that of abnormal eddies. Crucially, the daily number of normal (abnormal) eddies increased (decreased) 9.68 (11.80) every year.