A sliding-window-threshold algorithm for identifying global mesoscale ocean fronts from satellite observations

Abstract

Mesoscale ocean fronts play a significant role in marine fisheries, ecosystems, and climate change, and a front detection algorithm appropriate for satellite Earth observation use is a fundamental tool to observe and monitor long-term frontal activities. Since various ecological processes usually work on different spatiotemporal scales, it is vital to separate the fronts at different scales from the multiscale oceans to better study their scale-specific biophysical processes. A sliding-window-threshold algorithm (SWTA) is thus developed to extract the mesoscale frontal zones and lines from the multiscale frontal information existing in the high-resolution satellite images, which employs the sliding-window thresholds to sift for the candidate frontal pixels and a set of morphological operations to then generate more continuous frontal zones and lines. Compared with the two commonly used gradient and population-based algorithms, statistical analysis indicates that a great majority of fronts identified by the existing algorithms are detected by the use of a SWTA, while the SWTA detects more fronts that were ignored by them. Visual and statistical analyses, together with some known fronts, suggest that SWTA seems to be as insensitive to gradient variations as the population-based algorithms are and they have the same merits as the gradient-based algorithms in the ability to identify coastal or complex fronts. The application, based on 14,245 SST images from 1982 to 2020, shows the global distribution patterns of frontal occurrence and intensity, with some well-known persistent fronts reappearing in the extensions of the western boundary currents and Antarctic circumpolar currents that are not clearly defined in the global probability maps generated with the previous algorithms. In addition, similar seasonal variations in the frontal occurrence are found in the five western boundary current/extension areas and the four eastern upwelling systems, reaching a peak intensity during the warm season and a minimum intensity during the cold season. The information on the mesoscale fronts detected by SWTA would be potentially useful for better understanding of the ecological and climatic effects of front dynamics and in the predicting of the distribution of fish species and other marine organisms in a changing climate.