Workflow 6

Workflow 6: Forecasting future phytoplankton dynamics using predictive models

This workflow is designed to analyse and forecast long-term phytoplankton dynamics by integrating biological and environmental data and applying advanced statistical and machine-learning approaches. Starting from cleaned and harmonised time-series datasets, the workflow explores temporal and seasonal patterns in phytoplankton density and investigates their relationships with key abiotic drivers. It applies both predictive and forecasting models—including regression-based methods, machine-learning algorithms, and seasonal time-series models, to capture nonlinear relationships, temporal dependencies, and seasonal variability. By comparing model performance and identifying the most influential environmental predictors, the workflow supports the interpretation of phytoplankton responses to environmental change and provides tools for anticipating future dynamics, with applications in ecosystem assessment, management, and early warning of potentially harmful events. Thanks to its modular structure, this workflow is flexible and easily adaptable. Users can customize the workflow by editing the code and parameters and by adding or removing cells. Furthermore, the workflow can be tailored to specific research needs by performing additional analyses.