Fine-Scale Spatial Patterns of Gelatinous Zooplankton in the Northern Gulf of Alaska

Abstract

The Northern Gulf of Alaska (NGA) is characterized by high variability across spatial and temporal scales. In the NGA, zooplankton are a crucial link between primary production and higher trophic levels. Understanding the mechanisms that structure zooplankton assemblages is important to our overall understanding of ecosystem functioning. Nonetheless, thorough description of zooplankton abundance and distribution patterns is challenging due to the inherent variability and complexity of the marine environment. The study of gelatinous zooplankton is further complicated by the limitations of traditional plankton net sampling methods that are inefficient for the collection of high-resolution spatiotemporal data and often inflict damage on these fragile bodied organisms. In the NGA, and many other ocean systems, this has historically left gelatinous zooplankton under sampled and poorly studied in comparison to cooccurring crustacean zooplankton. To address these challenges, recent advances in imaging technology and computing power were leveraged by deploying an In Situ Ichthyoplankton Imaging System Deep-Focus Particle Imager (ISIIS-DPI) in the NGA from 2022-2023. The ISIIS-DPI is a towed vehicle capable of collecting vast amounts of high-resolution imaging and oceanographic data. An analysis pipeline with convolutional neural network (CNN) architecture was employed to automate the identification of zooplankton images and expedite processing time, allowing for description of fine-scale distributional patterns of gelatinous zooplankton and their associations with surrounding biophysical drivers. Evidence is presented that ctenophore, hydromedusae, and siphonophore aggregations are concentrated around frontal features and track with the surrounding variability in their ocean environment. Several first records in the NGA of previously undetected species are also presented. These novel datasets demonstrate the previously underestimated prominence of gelatinous zooplankton in the NGA and improve our understanding of ctenophore, hydromedusae, and siphonophore abundance and distribution patterns in the context of their oceanographic environment. This work is the first adaptation of in situ imaging and machine learning technologies in the NGA and presents the opportunity to more accurately describe the role of gelatinous zooplankton in marine ecosystem function.