Grading the Ensembles...

Every now and then, I check on some aspect of long-term forecasting to see how reliable such forecasts are, and if they are even worth considering. For example, for the month of February, I performed an analysis of the CPC's monthly forecast. I considered several random locations across the U.S., measured their monthly temps and precip against normal monthly values. Looking at the monthly forecast, I discovered that the CPC could have done better, but their forecast for the end of the month really helped their overall grade: C+

This month, I'll be looking at the CPC forecast for April. However, recently, I performed an analysis of the GFS Ensembles for 2m Temp Depart From Normal. On April 1, I looked at the Ensembles for April 2-12 and used Louisville, Indianapolis, Nashville, St Louis, Bowling Green, and Memphis for my analysis. I looked at all daily normal highs for each city during the period, considered the shaded regions indicating the departure from normal, and plugged in the actual highs against the forecast.

The results have been tallied. I thought that I was very generous in my grading process, especially since forecasting temps beyond 3 days can be very inaccurate. However, one notable problem I found with the ensembles during this time frame was the gross underestimation of above normal temps. The teleconnectors, NAO and PNA, both indicated above normal temps for this time period while the GFS ensembles mainly forecasted normal to below normal readings.

Here was a breakdown of the grades:
Louisville D
Indianapolis C
Nashville C+
Memphis B-
Bowling Green C

Overall grade for the region was: C

Again, I was very generous in the grading process. I would like to see the percentages come up just a bit. It could've been just a bad stretch. I will perform another analysis starting tomorrow, April 19, and run through the end of the month.

The teleconnectors I'm using, the NAO and PNA, are showing a cooldown toward the last part of the month. Possible below normal temps for a few days may occur. I'll see how the GFS Ensembles for 2m Temp Departures From Normal handle its forecast for this period...

I'll have updates on our impending severe weather chances this week soon. Plus, I'm going to discuss whether tornadoes 'skip' or not. Pretty interesting stuff. Come on back now, ya hear.
MS