User-tailored climate predictions – the DWD climate predictions website

Abstract

<p>DWD publishes operational seasonal climate predictions since 2016. In the following years work towards a seamless climate predictions website commenced, with the aim to provide consistent climate predictions across all timescales, focused on the needs of national users.</p><p>Today, the DWD climate predictions website www.dwd.de/climatepredictions contains operational decadal and seasonal predictions. Next, we will add post-processed subseasonal prediction products, derived from the IFS forecasts provided by ECMWF, as a further step towards the seamless climate prediction approach.</p><p>The user-oriented graphical presentation of the climate predictions is identical over all timescales. It was co-designed with stakeholders from various sectors at user workshops on climate predictions and through surveys and individual user meetings to guarantee its comprehensibility and usability. This co-design always includes the aspect of how the available predictive power is clearly addressed and its limitations are presented in a way that is understandable to users.</p><p>As a result, the website offers different layers of information on a basic and an expert level. It includes maps, time series and tables of 1- and 5-year means (decadal) and 3-month means (seasonal) ensemble mean and probabilistic predictions of temperature and precipitation on a global scale and for Europe, Germany, and German regions. For the subseasonal scale, we will add corresponding figures for weekly means of temperature and precipitation.</p><p>The information on DWD’s climate predictions website is retrieved from post-processed model output of the German seasonal and decadal prediction systems based on MPI-ESM. A statistical recalibration is applied to improve the skill of decadal climate predictions. It performs a drift correction as well as a lead time dependent optimization of conditional bias and ensemble spread. To fit the needs of German climate data users of a high spatial resolution in Germany (~20 km) and for climate predictions for German cities (based on ~5 km simulations), the empirical-statistical downscaling EPISODES is applied. All predictions are displayed in combination with their skill.</p><p>We work on several extensions of the website: multi-year seasonal predictions (e.g. 5-year summer means), the prediction of extreme indices (e.g. drought indices) and El Nino Southern Oscillation predictions. In addition, a seamless time series combining observations, climate predictions and climate projections is in preparation.</p>