Complete Time-series Analysis for extreme rainfall estimation

Código

I-EBHE0213

Autores

Elena Volpi, Aldo Fiori

Tema

WG 1.10: Hydrologic Design - Solutions & Communication

Resumo

Understanding the probability of extreme precipitation events for various time scales is critical for infrastructure design and risk management. Extreme rainfall estimation generally relies on the Intensity-Duration-Frequency (IDF) curves, which are used e.g. to design urban drainage systems for stormwater control. Common approaches to estimating IDF curves need to be revised to account for available information and, potentially, non-stationarity due to global climate change. Complete time series analysis (CTA) is discussed here as a tool for extreme rainfall analysis, as an alternative to the classical approach based on annual maxima (AM), which is the basis of traditional extreme value analysis. Indeed, AM are usually analyzed to capture the tail of the distribution of the parent process, where the latter is the process of interest. The rationale behind CTA is to exploit all of the information provided by the observational data, with the goal of better estimating the return period over a wider range of values, not just at the largest extremes that are the focus of extreme value theory. Small to moderate return period values are still of interest in several practical problems, such as flooding.