N. Agatheris, Extreme-oriented rainfall modelling on global scale using knowable moments, Diploma thesis, 140 pages, Department of Water Resources and Environmental Engineering – National Technical University of Athens, Athens, October 2019.
Assessment of extremes in hydrological processes is crucial in a variety of tasks from engineering design to risk management. Using classical moments to express important attributes of such assessment, proves to be efficient only for low order of moments. However, extreme rainfall events are better modelled using high-order moments. Whilst L – moments can be reliably estimated even for those higher orders, they fail in accounting for long-term dependence bias which exists in most large hydrological records. Thus, the newly introduced knowable (K) moments are used to model extremes, as they provide better grounds for prediction based on high orders, whilst retaining precision of classical moments for low orders. This study’s findings may improve knowledge on how to correctly model and predict such extreme rainfall events, providing comparison between the effectiveness of K – moments and classic methods. As this is a global study using data from the GHCN – Daily database, an attempt is made at constructing the basic framework for correlating a distribution’s fitting parameters and regional climatic characteristics.
Full text (8942 KB)