A stochastic approach on the extreme hydrological events: the case of Thessaly, Greece

In September 2023 storm Daniel struck the area of Thessaly, in the central part of Greece, causing extreme rainfall over four consecutive days. The aftereffects were devastating, as 17 people died, extensive damage -yet to be restored- was caused to infrastructure (including roads, bridges and the port basin of Volos) and the economic impact was also severe. This devastating disaster could have been limited if a reliable estimate of flood risk was available and efficient risk mitigation measures were adopted. To move a step forward towards such target, stochastic models serve as powerful tools for predicting floods and extreme rainfall incidents since they accurately simulate the inherent uncertainty that characterises natural processes like precipitation and river flows. In this work, we obtain historical precipitation data for the area of Thessaly and by applying the appropriate stochastic models and procedures we generate synthetic rainfall data. Then, by comparing the synthetic data to the historical, in stochastic terms, we test at what degree the stochastic models can effectively capture the variability of natural processes. The resulting synthetic data provide valuable insight in the likelihood of occurrence of extreme events, paving the way for incorporating stochastic tools in the development of flood early warning systems (FEWS). Furthermore, the application of stochastic models in extreme rainfall events will also guide the infrastructure design, in order to be resilient against extreme weather events, and will facilitate water resources management.