NCEP Service Center Evaluations HPC Pete Manousos

• Real-Time Parallel Run Evaluation Comments:

• HPC forecasters experience has subjectively indicated the parallel NAM QPF does seem to outperform (by a small margin) the operational version.  Differences in mass fields between the two versions do not typically become manifest until about f60 and tend to be subtle. Precipitable water values in the parallel version seem to be about .1" higher than the operational version (which was in the noise level compared to the observations). Additionally, the parallel NAM has been a little more progressive by about a half degree lon (and slightly weaker by about 30dm) at 500mb than the operational version with cut off lows moving across the CONUS.  This latter tendency met with mixed review internally at HPC. Objective statistics generated for QPF indicated for light amounts (less than 1 inch), there are no real difference in threat scores.  There is some improvement for thresholds of 1 inch or greater in the day 1 time frame.  However, differences are minimal between the two versions for days 2 and 3 at the same thresholds.  The parallel version did exhibit a slightly higher bias (closer to 1) than the operational version, but by day 3 differences in bias are negligible