Improving Weather Warnings for Flood Forecasting Centers by Enabling Specific Individualization in the DWD Warning Portal

Expert users of weather warnings often have specific and diverse needs. These needs depend on the type of application, as the relevant warning thresholds for weather conditions and forecast probability can vary significantly.

To meet these individual requirements, the German Meteorological Service (Deutscher Wetterdienst, DWD) is developing a new tool called the warning portal (“DWD-Warnportal”). It is a web app that allows users to customize, receive, and visualize probabilistic weather warnings based on their specific needs. In its current version, users can define an area of interest, select weather elements and choose from a set of pre-configured thresholds for these elements. Based on these settings, the occurrence probabilities for each threshold are calculated and presented both spatially on a map and temporally in a bar chart.
The warning portal is part of the broader RainBoW program (“Risk-based, Application-oriented and Individualizable Provision of Optimized Warning Information”), which aims to renew the weather warning system with a strong focus on the end users.

One key user group of the warning portal includes experts from the hydrological sector. To ensure that the application meets their needs, we are following a co-design approach. This involves close collaboration with flood forecasting centers through joint workshops and meetings to gather feedback and ideas. These insights are integrated into our agile development process and directly shape the design of new features. Features inspired by the requirements of the hydrological sector include warnings based on aggregation measures over individual areas, such as river catchments. This enables rainfall warnings based on spatial averages, maximum, minimum, or specific percentiles, providing a measure for areal rainfall that can serve as a rainfall-based signal for flood risk. We are also working to include extreme value analysis of historical rainfall observations for user-specific catchments. This allows the determination of return periods for forecasted rainfall events, helping to assess and communicate their statistical extremity and thus enabling faster detection of potentially critical events.

Here, we will give an overview of the current development of the warning portal, focussing on the relevant features for flood forecasting centers. We welcome feedback from hydrological experts to help us better understand their needs and improve the warning portal.