EGU22-8184, updated on 09 Jan 2023
EGU General Assembly 2022
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.

Operational impact-based flood forecasting in data-scarce environments: the Early Warning System of the Buzi watershed in Mozambique 

Andrea Libertino1, Alessandro Masoero1, Maria Laura Poletti1, Isac Filimone2, Matteo Darienzo1, Flavio Pignone1, Giacomo Fagugli1, Lia Romano3, Agostinho Vilanculos2, Lauro Rossi1, and Simone Gabellani1
Andrea Libertino et al.
  • 1CIMA Foundation, Savona, Italy (
  • 2DNGRH Direcção Nacional de Gestão de Recursos Hídricos, Maputo, Mozambique
  • 3WeWorld-GVC, Italy

Mozambique is one of the countries in Africa most frequently and most seriously affected by natural disasters such as floods, cyclones and droughts. In March 2019 the Cyclone Idai, one of the Southern Hemisphere’s deadliest storms, made landfall in the central part of the country, affecting about 1.7 million citizens, with devastating flooding in the central part of the country, especially in the Buzi and Pungwe river basins. Despite the existence of several studies aimed at the hydrological characterization of the area, the unexpected severity of the event undermined the local EW/EA system.

In the framework of the ECHO funded project “Building inclusive resilient communities and schools to face rapid-onset hazards in risk-prone areas in Mozambique affected by cyclone Idai, linking early warning to early action”, an operational flood forecasting system, up to real-time inundation mapping, have been implemented for the Buzi watershed (30’000 km2, in Manica and Sofala Provinces), with the aim of increasing the preparedness and response capacity to rapid onset disasters of the local and national levels of the EW/EA systems. For granting the sustainability and the maintenance of the tool, the operational chain has been implemented in co-operation with the local authorities (DNGRH) and is based on the use of open-source free software and models.

A preliminary collection of the available data has been carried out for the setup and the calibration of the CONTINUUM hydrological fully distributed model (Silvestro, 2013). Several existing studies have been considered for the development of the land data and the collection of hydrological measurements for calibration. Furthermore, the outdated level-discharge rating curves available have been reviewed and updated using an innovative approach (BayDERS, Darienzo 2021).

Stemming from the output of a long-term hydrological simulation fed with meteorological reanalysis conditioned with local rainfall data, dynamic flood scenarios have been developed for the Dombe flood prone community by setting up a hydraulic model with the Telemac-2D open system using the Copernicus DSM at 30 m resolution as topographical input. Outcomes obtained by simulating the Idai 2019 flood has been compared with satellite images, demonstrating good agreement and reliability of the implemented model. Modelled flood maps have been shared and commented with the local community in Dombe, with the dual objective of receiving feedback on map reliability and increasing flood risk awareness.

The full flood forecasting chain for the Buzi watershed has been then operationally implemented by means of the FloodPROOFS open-source modelling system (, fed twice per day by deterministic and probabilistic forecasts freely provided by NOAA (GFS and GEFS). Operational forecasts are made available to DNGRH officers through the EW platform, informing on potential flood events expected for the following 5 days, including their probability of occurrence, thus facilitating decision making in issuing early warnings and taking early action measures.

Finally, for the Dombe pilot-case the flood depth and water velocity maps are combined with the spatial distribution of the exposed assets, identified in collaboration with the community itself, resulting in real-time forecasts of the expected impacts. 

How to cite: Libertino, A., Masoero, A., Poletti, M. L., Filimone, I., Darienzo, M., Pignone, F., Fagugli, G., Romano, L., Vilanculos, A., Rossi, L., and Gabellani, S.: Operational impact-based flood forecasting in data-scarce environments: the Early Warning System of the Buzi watershed in Mozambique , EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-8184,, 2022.