Advancing Mapping and Monitoring of Canopy-Forming Brown Macroalgae Using Multispectral Remote Sensing

Joshua Mutter¹, Tom Fairchild¹, Rocio Hernandez-Clemente^2,1, Alberto Hornero^3,4, John Griffin¹

¹Swansea University, UK - ²Universidad de Córdoba, Spain - ³Consejo Superior de Investigaciones Científicas, Spain - ⁴The University of Melbourne, Australia

Canopy-forming brown macroalgae support diverse organisms and provide critical ecosystem services. Monitoring their forest cover, recognized as an Essential Ocean Variable, is increasingly feasible using publicly available satellite data. The Brown Algal Index (BAI) has shown promise in differentiating canopy-forming brown macroalgae on rocky shores, enabling rapid cover assessments. However, its large-scale application across shores, where species may have varying pigmentation and reflectance, remains underexplored. Additionally, while the overall cover is informative, the ability to differentiate between species or species groups would enhance ecological insights. We trained and tested predictive models using Sentinel-2 data and ground-truthed multispectral UAV surveys from varied rocky shores around Great Britain. Our results demonstrate that BAI can robustly estimate total brown macroalgal cover when parameterized with data that accounts for the natural heterogeneity of rocky shores. However, we caution that applying a single BAI regression model universally may weaken predictions mainly due to species-specific reflectance variations. Random Forest classifiers accurately predicted species-group cover during cross-validation within the same set of shores but require additional data to generalize to novel shores with differing conditions. These findings support the potential of multispectral satellite imagery to upscale intertidal macroalgal mapping and monitoring, with preliminary evidence for its use in species-level differentiation to improve ecological assessments.