UNDERSTANDING THE PHENOMENON OF Sargassum spp. MASS DYNAMICS USING FRACTAL GEOMETRY THROUGH ORBITAL PRODUCTS

Since 2011, extreme strandings of Sargassum spp. have been recorded off the west coast of Africa, the southwest Atlantic, the Caribbean Sea and the coast of Florida, resulting in significant ecological and economic impacts, such as increased trophic level, development of associated and harmful algae, competition for oxygen, changes in habitats and decline in food chains. This work investigated the behavior and structural complexity of these masses using multifractal analysis applied to georeferenced images from the SargAlert (ANR) consortium. The images obtained are products classified as 0 (no Sargassum spp.) and 1 (Sargassum spp.), covering two sets: CWA, with 167 daily images for some months from 2020 to 2024, and MNQ, with 263 images with five-day observations in the period from 2018 to 2022. The analyses included pixel quantification (particle count, average particle area and total coverage) and multifractal parameters such as lacunarity(Λ), width of the multifractal spectrum(Δ_α), spectrum asymmetry (f(Δ_α)) and dominant singularity (α₀). Regions with vortices, identified visually in the images, were isolated and checked separately, subdivided into east (VE) and west (VW). The images were also partitioned into spatial squares of 8 km for NMQ and 93.2 km for CWA, allowing the creation of distribution maps of the multifractal parameters. The results indicated that the images of both areas have geometric multifractal properties, with generalized dimensions following D₀>D₁>D₂. The singularity spectra of the CWA are more symmetrical than those of the NMQ. In NMQ, there was a high correlation between the particle count and the total coverage area, while Λ and the average area showed a negative correlation with Δ_α, f(Δ_α) and α₀. In 2019, there was high coverage and concentration of particles at the start of the year, followed by an increase in Λ and the average number of particles. In 2021, the images showed high Λ and average particle area, followed by an increase in total coverage and particle count, and then a return to high Λ and average particle area values at the end. For CWA, there was a positive correlation between α₀ and coverage area, with a negative correlation between Λ and f(Δ_α). The 2020 images stood out with the highest Λ and f(Δ_α) values. In vortices, VE and VW showed similar behavior: the total area correlated positively with the average particle size, as did α₀ and particle count, as did Δ_α and f(Δ_α). The distribution maps showed areas of greater dispersion, complexity and symmetry. In the CWA, the vortices showed low values for f(Δ_α), high values for α₀ and moderate values for Λ. The multifractal approach was shown to be effective in assessing the structural complexity of Sargassum spp. and provides new information that can support management decisions and mitigate the impacts of these algal masses.