Skill and added value of the MiKlip regional decadal prediction system for temperature over Europe. Issue 1 (1st January 2019)

• Record Type: Journal Article
• Authors: Feldmann, Hendrik; Pinto, Joaquim g.; Laube, Natalie; Uhlig, Marianne; Moemken, Julia; Pasternack, Alexander; Früh, Barbara; Pohlmann, Holger; Kottmeier, Christoph
• Is Part Of: Tellus. Volume 71:Issue 1(2019)
• Publication Date: 2019-01-01
• Subjects: regional decadal predictions -- added value of downscaling -- forecast recalibration -- MiKlip; Dynamic meteorology -- Periodicals; Oceanography -- Periodicals; 551.5
• DOI: 10.1080/16000870.2019.1618678
• Languages: English
• ISSNs: 0280-6495

Abstract

Abstract: In recent years, several decadal prediction systems have been developed to provide multi-year predictions of the climate for the next 5–10 years. On the global scale, high decadal predictability has been identified for the North Atlantic sector, often extending over Europe. The first full regional hindcast ensemble, derived from dynamical downscaling, was produced within the German MiKlip project ('decadal predictions'). The ensemble features annual starting dates from 1960 to 2017, with 10 decadal hindcasts per starting year. The global component of the prediction system uses the MPI-ESM-LR and the downscaling is performed with the regional climate model COSMO-CLM (CCLM). The present study focusses on a range of aspects dealing with the skill and added value of regional decadal temperature predictions over Europe. The results substantiate the added value of the regional hindcasts compared to the forcing global model as well as to un-initialized simulations. The results show that the hindcasts are skilful both for annual and seasonal means, and that the scores are comparable for different observational reference data sets. The predictive skill increases from earlier to more recent start-years. A recalibration of the simulation data generally improves the skill further, which can also be transferred to more user-relevant variables and extreme values like daily maximum temperatures and heating degree-days. These results provide evidence of the potential for the regional climate predictions to provide valuable climate information on the decadal time-scale to users.