Monthly and Interannual Variations in Upwelling and Chlorophyll off the Vietnam Coast in Summer Based on Deep Learning

Abstract

The deep learning technique is used to automatically identify upwelling regions off the Vietnam coast using Sea Surface Temperature (SST) data from satellite imagery. After delineating these upwelling areas, this study is pioneering in deriving the upwelling probability off the Vietnam coast, pinpointing two significant upwelling hotspots. To deepen the understanding of the individual impacts of wind and sea surface height anomalies on monthly and spatial variations in upwelling probability and chlorophyll distribution, the Empirical Orthogonal Function method was employed to analyze this data set. The upwelling probability demonstrates a close relationship with wind and sea surface height anomalies, while chlorophyll concentration correlates with the strength of the southwesterly wind. The southwesterly wind, Ekman pumping, eddy dipole, and northeastward jet have been identified as key drivers of the formation and spatial variability of upwelling in summer, albeit with varying contributions to different parts of the coastal region. The distribution of upwelling probability correlates with chlorophyll variation off the northern Vietnam coast. High chlorophyll concentration off the southern Vietnam coast is primarily influenced by the southwesterly wind that carries the Mekong River plume eastward. Furthermore, two abnormal scenarios in upwelling and chlorophyll concentration during 2010 and 2018 were analyzed, attributed to the abnormal southwesterly monsoon and coastal circulation influenced by an El Niño event and a weak La Niña event with a positive Indian Ocean Dipole, respectively. This study elucidates the dynamic processes underlying the intricate chlorophyll distribution over the continental shelf, which is influenced by both the river plume and upwelling.