Short‐term biweekly variations in the surface upwelling off northeastern Taiwan are well documented. However, due to limitations in the spatiotemporal resolution and coverage of the observed data, the lifecycle and associated dynamics are poorly understood. In this study, a gradient‐based edge detection algorithm is proposed to detect the surface upwelling. The sea surface temperature data of geostationary satellite Himawari‐8 were used to explore a complete short‐term process as a case study. The evolution was analyzed in terms of an upwelling index measured via the temperature difference between upwelled and surrounding water and the area and shape of the surface upwelling. The process lasted approximately 17 days as the upwelling moved northeastward under the advection of the Kuroshio Current and included two stages: intensification and decay. The last 8 years (2010–2017) of multisatellite combined sea surface temperature data were examined to verify the results obtained in the case study. A statistical analysis shows that most of the short‐term processes occur in the summer with an average lifecycle of approximately 15 ± 5 days. The trajectories of the upwelling center can be grouped into three types: quasi‐stationary near the shelf break, moving northward across the shelf break, and moving northeastward along the shelf break. A preliminary discussion suggests that the observed variations may be caused by the collective effects of multiple dynamical mechanisms, such as fluctuations in the Kuroshio Current, surface wind curl, and typhoons. This study furthers our knowledge of the surface upwelling off northeastern Taiwan and provides constraints for numerical simulations.

Oceanic frontal waves are frequently observed, but their life cycles are poorly understood because of the lack of time series data. In this study, the data of geostationary ocean color imager was used to explore the complete evolutionary process of submesoscale frontal waves off the southeast coast of China. Their evolution was analyzed in terms of both wave outline and ridge lines. The process lasted approximately 10 days as the waves propagated southward along the isobaths, accompanied by tidal oscillations. The life cycle comprised three stages: development, maturation, and decay. Scale estimation suggested that the onset of this process is caused by the collective effect of forced motion and unforced instability which is triggered by the passage of a tropical storm. The observed life cycle of frontal waves will provide an empirical basis for future theoretical investigations.