Chlorophyll-a Mapping and Prediction in the Mar Menor Lagoon Using C2RCC-Processed Sentinel 2 Imagery

A research team from the University of Murcia has published a novel AI-driven methodology for predicting and mapping chlorophyll-a (Chl-a) concentrations in the Mar Menor, Europe's largest coastal lagoon. The work addresses a critical environmental challenge: severe eutrophication crises that cause harmful algal blooms. By integrating nearly a decade of Sentinel 2 satellite imagery—processed with the C2RCC atmospheric correction algorithm—with in-situ buoy measurements, the team has created a scalable monitoring solution that overcomes the spatial and temporal limitations of traditional water sampling. The core innovation is a depth-specific prediction model, moving beyond simple surface estimates to map the entire water column, which is vital for accurate ecological assessment.

The researchers trained and validated multiple machine learning (ML) and deep learning (DL) algorithms, including Random Forest (RF), XGBoost, CatBoost, and Multilayer Perceptrons, using cross-validation. Performance was depth-dependent: ensemble models and XGBoost achieved an R² of 0.89 for surface water (0-1m), CatBoost reached R²=0.87 for 1-2m depth, and accuracy remained robust down to 3-4m (R²=0.66 with RF). The generated maps successfully reproduced known eutrophication events like the 2016 crisis and a 2025 surge, confirming the model's robustness. This end-to-end framework, which combines specific multispectral band combinations with ML, offers a transferable blueprint for monitoring other turbid coastal ecosystems globally, enhancing early-warning capabilities for environmental managers.