Storm features can explain sub-daily extreme precipitation statistics

Código

I-EBHE0094

Autores

Eleonora Dallan, Francesco Marra, Hayley J. Fowler, Marco Borga

Tema

WG 1.10: Hydrologic Design - Solutions & Communication

Resumo

Quantifying hydrometeorological extremes is critical for managing hydrological hazards and effective climate change adaptation strategies. Typically, these extremes are assessed through frequency analysis of precipitation data. The aim of this work is to go beyond mere statistical extrapolation to enhance physical comprehension by investigating the relation between the statistics of extremes and their underlying physical processes. We employ here a non-asymptotic extreme value approach based on the concept of storm objects and ordinary events. The analysis is based on a network of about 150 rain gauges and temperature stations in a complex-orography region in Northeastern Italy. We estimate precipitation extremes from sub-hourly to daily durations and return periods for up to 100 years. Several storm features are extracted (e.g. peak and average intensity, duration, seasonality, temporal profile, peakedness, temperature, etc.) and their relations with the parameters of the non-asymptotic extreme value distribution are investigated. First results show variations of the statistical parameters depending on topography and precipitation duration. Heavier tails in extreme precipitation distribution emerge at sub-hourly durations in mountains and part of the lowland, while the scale parameter seems generally higher in the pre-alpine band. Storm characteristics also show patterns varying with topography, precipitation duration, and extremeness of the events. Summery front-loaded storms seem prevailing at the short durations where we find generally heavier tails, but more in-depth analysis and considerations are needed for assessing the relationship between storm properties and statistical properties. By identifying key proxies influencing precipitation extremal behavior and understanding the formation processes underlying these, this research could contribute to the development of more robust predictive models for precipitation extremes and a better understanding of their changes in a changing climate.