Rainfall amounts, especially from thunderstorms, can vary widely in time and space. Meteorologists have been hopeful that a very new technique called "mesoscale modeling" will eventually be of great assistance.
Yesterday, we had an example of the promise and peril of mesoscale modeling.
The Actual Rainfall
I certainly see this as a "glass three-quarters full" situation. The amounts of rain were quite accurate and the locations were close enough to be useful -- but, the human meteorologist was required in this situation to monitor the radar and place any flood forecasts (with the drought, little flooding expected).
So, while this is an excellent start, we still have a ways to go before these forecasts are good enough to be used on their own.