A few issues were discovered in the climate-model data processing such as data preparation, bias-adjustment and calculation of climate impact indicators.
Inconsistency between daily mean, daily minimum and daily maximum temperature
Two projections (KIOST-KIOST-ESM, NOAA-GFDL-GFDL-ESM4) had cases where in the original data, the daily mean temperature could fall outside daily minimum and maximum temperature. One of them had to be excluded from the ensemble due to the severity of the issue. The other one could be handled by adjusting minimum or maximum temperature to be physically consistent with mean temperature in the bias-adjustment of the temperature variables. This projection only had few problematic cases and most of the cases were over the Sea and in high Northern latitudes.
High maximum temperature
Two projections (NCC-NorESM2-MM, MOHC-UKESM1-0-LL) showed an unphysically high maximum temperature (both in daily minimum, mean and maximum temperature) in the bias-adjusted data over parts of Siberia and Northern America in the future climate. The same high temperatures were not seen in the non-adjusted data. Cases with very high temperatures only were seen for the highest percentiles in the daily data series. The results for the calculated temperature CIIs were not affected by those high values.