Downscaling Models for Early Warning Forecast of Extreme Hydrological Events

EasyChair Preprint no. 4734

Abstract

The development  of downscaling methods to link regional or basin scale  extreme precipitation to large scale atmospheric circulation features is crucial for achieving reliable  early warning forecast of potentially  dangerous flooding events, as well as, for assessing  the probable  future  entity of hydraulic risk  of flooding under global warming scenarios.

In this note two downscaling approaches, namely non-homogenous hidden markov model ( Cioffi et al. 2016,2017.2020)  and event synchronization model ( Conticello et al. 2018,2020), are combined together in a model system  able to calculate, as a function of the fields of  two atmopheric predictors - geopotential at 850 hPa and the Integrate Vapor Transport (IVT) -   the probability of occurence of daily rainfall amount exceeding a prefixed  threshold reputed enough high to generate dangerous  floodings.

The model sistem is developped and validated for the case study of St. Lucia Island ( Caribbean) as part of an Early Warning System  of flooding.

Keyphrases: Event Synchronization, Eventi Estremi, NHMM, Previsioni a medio termine

@Booklet{EasyChair:4734,
  author = {Francesco Cioffi and Tommaso Lapini and Federico Rosario Conticello},
  title = {Downscaling Models for Early Warning Forecast of Extreme Hydrological Events},
  howpublished = {EasyChair Preprint no. 4734},

  year = {EasyChair, 2020}}