It is difficult to make accurate forecasts of warm season rainfall along the Southeast Coast of the U.S.  The difficulty arises from the complex processes that drive the convection. The thunderstorms are produced from collisions and other interactions between the large scale synoptic flow and mesoscale sea-breeze boundaries.  The rising air parcels generated by these collisions may encounter large variations in the environmental temperature, moisture and wind fields.  The resulting daily rainfalls can be a serious obstacle to outdoor activities, whether for commercial or recreational interests.  The transistion from text products to high resolution gridded forecasts demands a new approach to the prediction of warm season convection to provide our customers with the temporal and spatial resolution they need to plan their activities.

            Past research in forecasting convective rainfall over Florida used “flow regimes”, which relies on synoptic pattern recognition to predict the most likely locations for diurnal convective initiation.  The flow regime concept has been used to generate spatial and temporal rainfall climatologies for the WFO Tallahassee's area of responsibility.

Gridded daily (1200 UTC to 1200 UTC) multi-sensor quantitative precipitation estimates (QPE) were obtained from the NWS Southeast River Forecast Center for the months of June-August, 2003-2008.  A sea-breeze regime was objectively assigned to each day using the Tallahassee rawindsonde data.  Whenever the QPE for any 4 km grid point was greater than 0.01 inches, then that day was determined to be a rain day for that point.  (Daily rainfalls of 0.01 inches or less were excluded to provide a first pass filter on the dataset.)  The grid of rain days for each regime was summed, and divided by the total number of days in that regime, to determine the frequency distribution or Probability of Precipitation.

Software procedures were developed to enable the forecasters to initialize the precipitation grids with the appropriate daily climatological rainfall probabiliies by specifying the synoptic sea breeze regime category for that day.

Work is underway to increase the temporal frequency of the rainfall climatologies from 24 hours to six hours, while extending the data back through the summer of 1996.  In addition, research has begun to adjust the climatological rainfall probabilities based on forecast thermodynamic parameters.  The greater temporal frequency and longer dataset will allow more specific precipitation timing within a forecast package.  The results of this research will allow the Tallahassee NWS office to provide a much improved and higher resolution gridded forecast of summer thunderstorms.