Author(s): Ana Rueda; Andrea Del Pozo; Laura Cagigal; Paula Camus; Alberto Luceno; Fernando Mendez

Linked Author(s): Ana Rueda

Keywords: Rainfall inundation; Tropical cyclones; Metamodel; Satellite data; Lisflood-FP

Abstract: Rainfall associated with Tropical Cyclones (TCs) can cause severe damages to buildings and infrastructures and the loss of life in the worst cases. Accurate and fast estimations of the expected rainfall and associated inundation for a forecasted event can help on the minimization of the losses. Climatology models such as r-clipper are normally used as a baseline to forecast TC associated rainfall, however, we have observed these models systematically underestimate the total rainfall. To improve the rainfall estimation for any given TC track, an alternative parametric model is developed based on satellite rainfall data and data mining techniques. Afterwards, the associated inundation is obtained based on a metamodel of Lisflood-FP for each watershed in the domain. The results show good agreement between observed and modeled rainfall and inundation extents with a significant reduction of the computational times. These suites of models have been used for probabilistic assessment of TC-induced rainfall forecasts in Pacific Island countries.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920222014

Year: 2022