Recent Development to Improve NAEFS SPP

The NAEFS (Northern American Ensemble Forecast System) Statistical Post Process (SPP) includes bias-correction and downscaling. The bias correction in the current NAEFS SPP is mainly first moment adjustment by applying a Kalman Filter (KF) technique with most recent forecast/observation information. Therefore, there is no second moment adjustment or correction. Recently, we have tested the method to improve NAEFS 1st moment correction, and implement 2nd moment adjustment. In order to enhance the performance of bias correction, 26-year GEFS reforecast has been applied in additional to current operational NAEFS bias correction process. Several different methods have been examined to optimize the usage of past 26-year reforecast information. For 2nd moment adjustment, the same KF technique has been adopted to accumulate past ensemble forecast uncertainty information for future reference of 2nd moment correction. A full year (2012) NCEP operational GEFS forecast has been studied to demonstrate the forecast improvement in term of forecast bias, reliability, probabilistic forecast and et al. Several forecast elements are investigated in this study, such as surface temperature, 500hPa geopotential height and et al, but we will mainly focus on the discussion of surface temperature up to 16 days.