A synthesis of the VALUE perfect predictor experiment

0478788 - ÚFA 2018 DE eng A - Abstrakt
Maraun, D. - Gutierrez, J. - Widmann, M. - Hertig, E. - Soares, P. - Kotlarski, S. - Fischer, A. - Wilcke, R. - Huth, Radan - San-Martin, D. - Bedia, J. - Herrera, S.
A synthesis of the VALUE perfect predictor experiment.
EMS Annual Meeting Abstracts, Vol. 14. Berlín: European Meteorological Society, 2017. EMS2017-845.
[EMS Annual Meeting and European Conference for Applied Meteorology and Climatology. 03.09.2017-07.09.2017, Dublin]
Institucionální podpora: RVO:68378289
Klíčová slova: climate change * downscaling * precipitation extremes
Kód oboru RIV: DG - Vědy o atmosféře, meteorologie
http://meetingorganizer.copernicus.org/EMS2017/EMS2017-845.pdf

VALUE is an open European network to validate and compare downscaling methods for climate change research. A key deliverable of VALUE is the development of a systematic validation framework to enable the assessment and comparison of both dynamical and statistical downscaling methods. VALUE’s main approach to validation is user-focused: starting from a specific user problem, a validation tree guides the selection of relevant validation indices and performance measures. We consider different aspects: marginal aspects such as mean, variance and extremes; temporal aspects such as spell length characteristics; spatial aspects such as the de-correlation length of precipitation extremes; multi-variate aspects such as the interplay of temperature and precipitation; and process-oriented aspects.

Several experiments have been designed to isolate specific points in the downscaling procedure where
problems may occur. A perfect predictor experiment has been conducted to isolate downscaling skill in present climate. In this experiment, downscaling methods are driven with ERA-Interim reanalysis data to eliminate global climate model errors, over the period 1979-2008. As reference data we use observations from 86 meteorological stations distributed across Europe. With more than 40 contributing methods, this study is the most comprehensive downscaling intercomparison project so far.

This presentation will give an overview on the results. They clearly indicate that for several aspects, the
downscaling skill varies considerably between different methods. For specific purposes, some methods can
therefore clearly be excluded.
Trvalý link: http://hdl.handle.net/11104/0275104