Evaluation of ECMWF's forecasting system for probabilistic urban flood prediction: a case study in Mexico City. Issue 1 (15th November 2021)

• Record Type: Journal Article
• Authors: López López, Marco R.; Pedrozo-Acuña, Adrián; Severiano Covarrubias, Marcela L.
• Is Part Of: Journal of hydroinformatics. Volume 24:Issue 1(2022)
• Page Start: 38
• Page End: 55
• Publication Date: 2021-11-15
• Subjects: ensemble forecast -- flood forecasting -- forecast verification -- probabilistic forecast -- urban floods; Hydrology -- Data processing -- Periodicals; Geographic information systems -- Periodicals; Geographic information systems; Hydrology -- Data processing; Electronic journals; Periodicals; 551.480285
• DOI: 10.2166/hydro.2021.072
• Languages: English
• ISSNs: 1464-7141

Abstract

Abstract: As the world continues urbanizing, including efforts to forge a new framework of urban development is necessary. Recent studies related to flood prediction and mitigation have shown that Ensemble Prediction Systems (EPSs) constitute a valuable and essential tool for an Early Warning System. However, the use of EPS for flood forecasting in urban zones has yet to be understood. This work has the objective to investigate the potential use of the Operational EPS, issued by the European Centre for Medium-Range Weather Forecasts (ECMWF), for probabilistic urban flood prediction. In this research, a precipitation forecast verification was carried out in two study zones: (1) Mexico Valley Basin and (2) Mexico City, where for the latter, forecasts were compared against real-time observed data. The results showed good forecast reliability for a rain threshold of up to 20 mm in 24-hourly accumulations, with the first 36 h of the forecast horizon being the most reliable. The EPS has sufficient resolution and precision for flood prediction in Mexico City, which represents a further step toward developing a flood warning system at the local level based on ensemble forecasts. HIGHLIGHTS: ECMWF's forecast evaluation from real-time measurements and at a temporal resolution of less than 12 h. Implementation of a predictive model to predict the occurrence of a flood event. To extend the use of EPS to include them in the chain for emergency and decision-making. This work seeks to shorten the gap of previous research and direct its efforts to the evaluation of urban and flash floods.