DERISC: Deep learning based Extreme Rainfall and flood warnIngs through Seamless foreCasting​

DERISC project

The project DERISC aims to use deep learning (DL) to generate rapidly updating, probabilistic seamless precipitation forecasts and reliable early warnings in a consistent framework.

By co-developing these with hydrological impact modelers, we will ensure that the resulting system can generate reliable impact-based warnings that allow stakeholders such as local authorities and emergency services to make better-informed decisions to mitigate risk due to extreme precipitation. By increasing the forecast reliability and sharpness at longer forecast lead times, we ensure that mitigating actions can be made in a more timely and effective manner at all decision levels, thereby drastically reducing the impact of these extreme events. The improved precipitation forecasts can also promote the sustainable exploitation of resources by enabling more efficient water management such as smart water buffering.

To reach these goals, we will go beyond the state of the art in seamless prediction of extreme events, by using innovative DL techniques for observation fusion and model combination, downscaling and calibration. While the focus of the project is on extreme precipitation events, the methodology developed will be transferable to other weather variables.

The DERISC project is a collaboration between RMI, VUB and KU Leuven with contributions by VMM, financed by the BRAIN-be 2.0 framework of BELSPO, the Belgian Science Policy Office.


Cookies saved