Distance-dependent depth-duration analysis in high-resolution hydro-meteorological ensemble forecasting: A case study in Malmö City, Sweden

•Ensemble high-resolution rainfall and runoff reforecasts of a cloudburst event.•Extreme rainfall intensities were forecasted in the region but with location errors.•A new approach: Distance-Dependent Depth-Duration in Forecasts - 4DF.•The 4DF-plot visualizes the spatial distribution of rainfall extremes in a region.•The associated “worst-case ensemble” of runoff forecast support risk assessment.

In 2014, a cloudburst hit Malmö City and parts of the city were severely flooded. In this study, we assess the possibility to forecast this event by high-resolution ensemble rainfall-runoff modelling. Meteorological reforecasts are generated by the convection-permitting HarmonEPS model and the hydrological response is simulated by the HYPE model. Although the meteorological forecasts manage to reproduce the observed extreme rainfall intensities, due to location errors only a weak, low-probability signal of a high runoff in Malmö City is obtained in direct forecasts. To describe and visualize the spatial uncertainty of local rainfall extremes in forecasts, and construct hydrological model forcing representing “worst-case ensembles”, an approach called Distance-Dependent Depth-Duration in Forecasts (4DF) is proposed. In this approach, a depth-duration analysis is combined with a neighborhood approach to describe rainfall severity as a function of distance from the basin. The results demonstrated the applicability of the approach for improved situation awareness.

Graphical abstract4DF-plot: Distance-Dependent Depth-Duration in forecasts: This plot illustrates the most severe Depth-Duration (D-AD) function for a certain basin size in successively larger forecast regions, defined by the radius R, as compared with an observed local relationship (OBS). The plot shows median functions from an ensemble forecast.Download high-res image (267KB)Download full-size image