Last Saturday unorganized convection was dominating in a low-shear environment, but a few cells became severe and some of them produced large hail according to the ESWD. One outstanding development occured in the border region of the southernmost Netherlands and neighbouring Germany.
Interestingly, COSMO DE-EPS produced nearly perfect forecasts of this development in three consecutive initialization times (see example for 06 UTC model initiation):
![20160604 8_maxsdi](http://www.essl.org/testbed/blog/wp-content/uploads/2016/06/20160604-8_maxsdi-300x200.png)
Supercell Detection Index (maximum SDI of all ensemble members) for 14 UTC
![20160604 8_upssdi](http://www.essl.org/testbed/blog/wp-content/uploads/2016/06/20160604-8_upssdi-300x200.png)
Fraction of 20 ensemble members exceeding a threshold for the SDI: more than 90 % in the correct 40×40 km box, while the rest of the domain was correctly modelled free of such an intense development.
![20160604 Radar](http://www.essl.org/testbed/blog/wp-content/uploads/2016/06/20160604-Radar1-300x272.png)
High resolution visible satellite and radar overlay of reflectivity and Mesocyclone Detections at 14:05 UTC.
An interesting example for the potential of high resolution ensembles that we found in the very first hours of this years’ Testbed.