Wednesday, June 26, 2013

Recent Fails In Storm Prediction From The SPC

Pardon the "grind my gear" bend of this upcoming article; however, I might as well be "that guy" and question the recent performance in severe thunderstorm outlooks from the Storm Prediction Center.

First, let me preface all of this by stating I completely understand that severe thunderstorm prediction in the summer is not an easy task.  Lots of mesoscale factors, weak disturbances, and atmospheric caps that either hold tight or pop too easily can lead to overperforming severe weather or a bust situation...or on scattered thunderstorm days such as the past two days, a little of both.

Looking at the pattern, though, and the potential for at least scattered thunderstorms, and one could have probably thought a rogue storm or two could potential go severe when forecasters were putting together the outlooks on Monday or Tuesday mornings.  However, both mornings featured a less than five percent risk threshold that was upgraded during the day to slight risk status.



The two graphics up above are Monday's initial severe weather forecast followed by the midday upgrade.  The SPC correctly upgraded the midday risk level to slight (15% chance of severe within 25 miles of a point, which sorta verified) as thunderstorms were firing up across the region during the midday hours on Monday.

Given the atmospheric dynamics were pretty similar yesterday to Monday, you'd think the early morning forecast for severe would have at least alluded to the possibility of severe weather in isolated locations. Um, it didn't.




But, by 9 AM the northern and western burbs were in a slight risk and then by 4 PM we were not only in a slight risk in Philadelphia, but were bumped from the 5% brown shade to the 30% potential of severe weather.

Dynamics and severe parameters don't change this quickly to result in such poor performing forecasts on the day of.  It's a hot and humid airmass with atmospheric disturbances around.  I can see upgrading to slight risks during the day but there should have been at least some mention of strong storms that could perhaps go severe from the Storm Prediction Center.

Upgrading storm risks during the day is one thing...while I'm not a big fan of it, I know that borderline situations do warrant upgrades or if you have a MCS or squall line upgrades are certainly within the realm of realistic expectations.  However, going from no risk whatever to a high end slight within the day is poor performance.

This isn't the first time the Storm Prediction Center has performed poorly...the derecho a year ago was another example of poor forecasting day of, partly due to modeling really crapping the bed across the Midwest and Great Lakes.  However, the past two days have shown scattered strong storms popping in high resolution modeling at enough of a clip to support at least a "see text" or 5% mention in severe weather outlooks.

If anything, the past two days show that severe weather forecasting does have a ways to go to get accurate...but the Storm Prediction Center should do better in this regard.  Nobody's perfect -- we certainly are not in forecasting weather -- but when it comes to predicting severe weather risks we should strive for better forecasts and better predictions from the Storm Prediction Center as the performance the past couple of days has been less than stellar.