On the following topic : Multi-scale variability of Sargasso in the tropical North Atlantic Ocean
Jury
Director of these Mr Léo BERLINE Mediterranean Institute of Oceanology - MIO
Co-Director of these Mr Christophe LETT MARBEC
Rapporteur Mr Ronan FABLET Information, Communication and Knowledge Sciences and Techniques Laboratory
Rapporteur Mr Jean-Olivier IRISSON Villefranche Oceanography Laboratory, LOV
Chairman Mr Frédéric MéNARD Institut Méditerranéen dOcéanologie - MIO
Examiner Ms Laure PECQUERIE Institut Universitaire Européen de la Mer, LEMAR
Summary of the thesis
Pelagic Sargassum seaweed, Sargassum fluitans and Sargassum natans, is both a refuge for biodiversity in open oceans and, for more than a decade now, an ecological, economic and health scourge for coastal areas affected by its massive stranding. These harmful impacts on ecosystems and human activities, spread from the Gulf of Guinea to the Gulf of Mexico, have prompted an international research effort to understand and predict the invasion of these algae. This thesis is part of this effort, focusing on the spatial and temporal distribution of Sargassum seaweed and its high inter-annual variability. This work aims to map and then model their seasonal cycle, taking into account both their passive drift and their growth. Firstly, the remote sensing of sargassum is addressed, using MODIS satellite imagery to provide large-scale quantitative and daily mapping of its spatial distribution. The contribution of this thesis in this area focuses on the filtering of detection errors using an automated learning algorithm based on the spatial characteristics of these algae. Secondly, sargassum drift, which has been poorly quantified until now, is observed in successive satellite images by extracting trajectories of sargassum aggregates. Based on the evaluation of the resulting aggregate speeds, a detailed model of Sargassum drift is proposed as a function of the surrounding wind and currents. In the final part of this thesis, this new drift model, used in a Lagrangian approach, is combined with a biogeochemical model of Sargassum growth in order to reproduce annual changes in the quantities of algae and their distribution. Analysis of the results of this drift and growth model reveals the spatialisation of Sargassum growth and degeneration, and also sheds light on the main environmental limiting factors. This analysis also highlights a crucial climatic feature which, through the warming of ocean surface temperatures north of Brazil, halts the seasonal development of Sargassum biomass. The temporality of this climatic feature could therefore be monitored to better predict the annual increase in Sargassum and to alert countries to future risks of groundings.