Nordic and European Summer Droughts in a Past, Current and Future Perspective

On this project, we investigate the relation between non-linear trends in atmospheric pressure and meteorological dryness, and apply empirical orthogonal function to link the large-scale atmospheric systems with the local responses in northern European streamflow. Additionally, we develop a data-driven spatiotemporal prediction model of northern European wildfire danger by using tree-based machine learning algorithms (Random Forest, Decision Tree and AdaBoost) with a wide range of indices as potential predictors. The resulting model can be used to map wildfire danger and highlight the importance of predictors not included in traditional fire danger indices.