I’m always on the lookout for the best forecast model. (Anyone into weather is probably on this ongoing search.) The issue is that some models do better in different weather patterns.

The only model that got close to predicting the thundershowers last night that affected the northwest suburbs and some of the Philadelphia area was a model that I haven’t until recently focused upon- the WRF with its two variants- the WRF-NMMB and the WRF-AFW.

The WRF (Weather Research Forecast) model is a short range (48 hour) model that has a high resolution. It’s only run twice a day for the US.

It did well last night and also with the storms last week. So let’s see how it does today.

The WRF is forecasting showers and thunderstorms to develop as early as 2:30 PM (earlier far northwest suburbs).  Most likely time for Philadelphia is 3-5 PM.  High temperature forecast for today is 84-85 before the showers move in.


Tonight’s models are doing extremely poorly predicting the thunderstorms that are on our doorstep at 11 PM.

NEXRAD radar image 11pm Saturday (weathertap.com)

This doesn’t bode well for Sunday’s forecast; if the models can’t get the 3 hour forecast correct, what confidence can one have for the 20 hour forecast?

There’s a wide range of times for the thunderstorms expected on Sunday.  They may start as early as 2:30 in the afternoon, or they may move through as late as 7 PM.  A low confidence forecast with tonight’s models.