I check the NWS forecast every morning.   I also check the “Forecast Discussion” section – where the forecasters give the details behind the forecast on the main page – because sometimes there’s something like this:

.NEAR TERM /UNTIL 6 PM THIS EVENING/...
***TODAY/S FORECAST IS OF VERY LOW CONFIDENCE GIVEN CONSIDERABLE
  MODEL DISAGREEMENT***

…

PRETTY REMARKABLE MODEL DISAGREEMENT FOR A 12 HOUR FORECAST.  THE
MODELS ARE STRUGGLING WITH A PLUME OF TROPICAL MOISTURE STREAMING
NORTHWARD FROM THE MID ATLANTIC STATES. WHENEVER YOU ARE DEALING
WITH TROPICAL MOISTURE AND VIRTUALLY NO THERMAL GRADIENT...ITS
GOING TO PLAY HAVOC WITH COMPUTER MODEL FORECASTS AND PLACEMENT OF
QPF.

AT THIS POINT...THE BEST THING WE CAN DO IS KEEP THE THEME OF THE
FORECAST GOING FROM THE LAST FEW DAYS...NOT JUMPING ON WILD MODEL
SWINGS...

Three cheers for truth in advertising.  Admission of uncertainty is a good thing.

PS  Speaking generally, a good way to deal with model uncertainty is Bayesian Model Averaging.   I’m sure implementation for weather forecasting would be non-trivial but it’s a good way of “harmonizing” results from different models – each of which could be the right one – and avoiding wild swings in predictions.