The DWD climate predictions website: Towards a seamless outlook based on subseasonal, seasonal and decadal predictions

Abstract

The climate predictions website of the Deutscher Wetterdienst (DWD, https://www.dwd.de/climatepredictions) presents a consistent operational outlook for the coming weeks, months and years, focusing on the needs of German users. At global scale, subseasonal predictions from the European Centre of Medium-Range Weather Forecasts as well as seasonal and decadal predictions from the DWD are used. Statistical downscaling is applied to achieve high resolution over Germany. Lead-time dependent bias correction is performed on all time scales. Additionally, decadal predictions are recalibrated.

The website offers ensemble mean and probabilistic predictions for temperature and precipitation combined with their skill (mean squared error skill score, ranked probability skill score). Two levels of complexity are offered: basic climate predictions display simple, regionally averaged information for Germany, German regions and cities as maps, time series and tables. The skill is presented as traffic light. Expert climate predictions show complex, gridded predictions for Germany (at high resolution), Europe and the world as maps and time series. The skill is displayed as the size of dots. Their color is related to the signal in the prediction.

The website was developed in cooperation with users from different sectors via surveys, workshops and meetings to guarantee its understandability and usability. The users realize the potential of climate predictions, but some need advice in using probabilistic predictions and skill. Future activities will include the further development of predictions to improve skill (multi-model ensembles, teleconnections), the introduction of additional products (data provision, extremes) and the further clarification of the information (interactivity, video clips).