Influence of marine weather and coastal geomorphology on seaweed wrack dynamics along Quebec coastsNovel Approaches

Annaëlle Anquet^1,2, Sébastien Dugas¹, Khouloud Baccara¹, Pascal Bernatchez¹, Thomas Burel², Fanny Noisette¹

Seaweed wracks are essential input of organic matter into coastal ecosystems, fueling temporary food webs and creating structure in often resource-scarce habitats. The accumulation of wrack on beaches depends on the combination of biotic and abiotic factors, including the physical forces (winds, currents, waves, ice, tides), the coastal morphology and geology, the species composition of the donor ecosystems and the life cycle of those species. Seaweed wracks remain highly unpredictable while characterizing the dynamics of these phenomena is essential for understanding their cycle and the functioning of beaches receiving these deposits. Using a unique network of stationary camera deployed along the shoreline of the St Lawrence system (Quebec, Canada), this study aims to detect and quantify seaweed wracks in association with marine weather and coastal geomorphology. Shore images taken every 15 minutes from 2014 to 2023 on 4 sites from the St Lawrence system were processed with a deep learning model to extract the occurrence of seaweed wrack and their cover. To identify the conditions that favor their accumulation and retention on shore, four physical variables were analyzed: waves, winds, surface currents and tides. Eight explanatory variables among these processes were considered for building linear models: wave height (m), period (s) and direction (°), wind speed (m s-1) and direction (°), surface current speed (m s-1) and direction (°), and tidal range (m). Preliminary results showed that seaweed wrack dynamics vary according to the sites studied and were significantly influenced by wave height, wind direction and surface current speed. The greatest seaweed accumulations were correlated to the highest wave heights observed. According to our observations, topography likely influenced seaweed retention on the beach. These results will be used to develop a predictive model of seaweed wrack and retention based on site-specific conditions. Camera detection, coupled with automated image processing using a deep learning model to detect and quantify seaweed wracks, offers promising perspectives for its characterization, with potential economic and management outcomes.