Comparing Internal Variabilities in Three Regional Single Model Initial-condition Large Ensembles (SMILE) over Europe

The increasing number of single model large ensembles and the growing number of publications in this field shows the great possibilities that lie in this concept, where the internal variability component of a model is used to estimate climate variability. Nevertheless, the underlying assumption always is that the internal variability of the chosen model is a good approximation of natural climate variability. Only few studies have undertaken the effort to validate this assumption properly. In this study, we compare the internal variabilities of three regional climate model large ensembles (16 members of an ECEARTH-RACMO ensemble, 21 members of a CESM-CCLM ensemble, 50 members of a CanESM-CRCM ensemble) for four European domains (British Islands, France, Mid-Europe, Alps). To the knowledge of the authors, this is the first comparison of regional large ensembles. Comparison includes the validation of interannual variability against E-OBS and inter-comparisons between the models, and comprises seasonal temperature and precipitation, as well as indicators covering extremes (heavy precipitation, dry periods and heat waves).

Results show a large consistency of all three models with E-OBS data for most indicators and regions, validating the abilities of these models to represent natural variability on the interannual scale. As a second step, the interannual variability in the models is compared to the inter-member variability of the ensembles. For the selected models, this assumption holds for all models, indicators and most regions (deviations only appear in the Alps). In a third step, the consensus of all three models in equal representations of internal variability is tested together with the analysis of the temporal evolutions of internal variability till the end of the 21^st century. There is no clear pattern (indicators and regions over time) for the accordance of the models, but they show similar variabilities in over 2/3 of the cases and mostly the same direction of change in variability.

Overall, the study validates three regional climate model large ensembles against E-OBS and between each other, showing mostly similar representations of natural variability in the test metrics. This strengthens the concept of large ensembles, and is of great value for the further development of such ensembles.