On Friday 9 June ECMWF had a delay for 12z HRES forecast. This was the result of a failure in the Pressure Bias correction scheme applied to conventional surface observations. To minimize any further delays the scheme was temporarily turned off, until a problem with some specific observations was identified. These observations were then excluded and operations continued with the scheme turned back on Sunday morning starting from the 06z run.
During the period the bias correction scheme was switched off, more surface pressure observations than normal ware rejected by the data assimilation and we also found a degraded fit of the analysis to the (non-bias corrected) observations. We have now reached the point when we can verify the affected forecasts, and although there maybe local degradations we do not see any significant impact in the hemispheric forecast scores currently.