Linking remote sensing-derived environmental gradients to in situ traits to predict Posidonia oceanica morphological...

Posidonia oceanica meadows serve as sensitive bioindicators of coastal pollution, yet the environmental drivers of their morphological variability remain poorly quantified. We integrated shoot-level measurements of 13 morphological traits with satellite-derived environmental variables to examine morphological plasticity across three sites along the western Algerian coast (Cap Blanc, Cap Carbon, Sidi Lakhdar). In situ sampling campaigns were conducted during 2012 and 2017, while environmental data- thermal-optical, hydrodynamic, and trophic indices-were extracted via Google Earth Engine (GEE) for the preceding years (2011 and 2016) to capture cumulative pre-seasonal conditions. Principal Component Analysis revealed two dominant axes of morphological variability-leaf elongation/surface area (PC1: 37.8 %) and leaf abundance (PC2: 16.8 %)- Sea Surface Temperature (SST) and Photosynthetically Active Radiation (PAR) (>140 μmol photons m-2 s-1). Hierarchical clustering further identified three phenotypic syndromes, reflecting gradients in foliar development and leaf number. Lower-development syndromes were more common with elevated turbidity and chlorophyll-a, while higher-development syndromes prevailed under clearer, sheltered conditions. Partial Least Squares Regression identified PAR, wind speed, and SST as key predictors, while Random Forest and Gradient Boosting models achieved R2 values up to 0.49 for surface area traits. Redundancy Analysis attributed 55.8 % of constrained variance to SST, PAR, and hydrodynamic gradients. These findings demonstrate that P. oceanica morphological plasticity is primarily driven by light and temperature gradients, modulated by hydrodynamics and trophic conditions. Key traits-leaf elongation, surface area, and shoot density-were predictable from satellite-derived variables, offering a scalable and transferable framework for large-scale, trait-based monitoring and early detection of stress in seagrass ecosystems worldwide.