Yes we look at model data but more importantly we look at what's actually happening on satellite and radar.
Earlier this afternoon I saw a large shield of rain in Mississippi tracking eastward. I knew that the air was dry across Alabama but was it dry enough to cause the precipitation to evaporate as it arrived? It turns out it had little impact on the precipitation reaching the ground and what you saw on radar matched up with what you experienced when stepping outside.
So I had to scramble and change the once dry forecast to a wet one. Model data is there for guidance and as you can see, it did not pick up on the northern extend of the rain.
Here's what the RPM showed the radar looking like at 6:30pm:
Oh, here's the 6:30pm radar snapshot:
Since the RPM was so far off, I couldn't show it on TV and decided to show the HRRR which was the closest. The HRRR updates every 1 hour, which makes it the most useful model in the short term forecasting. The model only goes out 15 hours, which means your accuracy beyond that time frame decreases dramatically.
Sometimes you just have to toss aside model data and time out the precipitation yourself. It looks like the back edge of rain will be near Tuscaloosa by 9pm, near Pell City by 10pm and exiting east by midnight. There are additional showers behind the main shield that will likely arrive on Friday morning.
Models have a hard time with small scale features and today is a perfect example! I expect these complications and challenges to continues through Saturday.
Keep up with the latest forecasts that will may have to be adjusted on the fly!
Jill Gilardi WBRC First Alert Certified Meteorologist