A MESOSCALE STATISTICAL THUNDERSTORM PREDICTION SYSTEM

A PROJECT IS PROPOSED TO DEVELOP A MESOSCALE STATISTICAL THUNDERSTORM PREDICTION SYSTEM (MSTPS) FOR THE KENNEDY SPACECENTER (KSC). IT WILL INGEST PREDICTORS OBTAINED FROM NUMERICAL FORECASTS GENERATED BY A MESO-BETA SCALE DYNAMICALMODEL AND OBSERVATIONS TAKEN FROM AN ARRAY OF ATMOSPHERIC OBSERVING PLATFORMS. THE OUTPUT FROM THIS SYSTEM WILL BE AN ESTIMATE OF THE PROBABILITY OF THUNDERSTORM OCCURRENCE DURING A SPECIFIED TWO-HOUR TIME PERIOD WITHIN A 10 KM SQUARE AREA CENTERED ON KSC. THE SYSTEM WILL BE DESIGNED TO PERMIT UPDATED FORECASTS TO BE PREPARED FROM 24 HOURS TO 1/2 HOUR PRIOR TO THE VERIFICATION PERIOD. THE THUNDERSTORMPROBABILITY WILL BE DETERMINED THROUGH THE USE OF A MULTIVARIATE DISCRIMINANT FUNCTION WITH THE PREDICTORS BEINGSELECTED ACCORDING TO THEIR RELATIVE PREDICTIVE POWER. SEPARATE DISCRIMINANT FUNCTIONS WILL BE DEVELOPED FOR EACH FORECAST PREPARATION TIME SO THAT A GRADUAL SHIFT BETWEEN THE MODEL AND OBSERVATIONAL PREDICTORS CAN BE ACHIEVED AS ONE APPROACHES THE FORECAST VERIFICATION TIME. THE DYNAMICAL MODEL SELECTED FOR THIS PROJECT IS THE MESOSCALE ATMOSPHERIC SIMULATION SYSTEM (MASS) MODEL. IT WILL BE INITIALIZED WITH A MESOSCALE 3-DIMENSIONAL ANALYSIS AND RUN FOR A 24 HOUR PERIOD TO PRODUCE MODEL PREDICTORS. THE OBSERVATIONAL PREDICTORS WILL INCLUDE TRADITIONAL DATA SUCH AS STANDARD SURFACE WEATHER REPORTS AND CONVENTIONAL RADAR ECHOES AS WELL AS DATA FROM EMERGING HIGH TECHNOLOGY SYSTEMSSUCH AS VHF DOPPLER WIND PROFILERS, THE NEXRAD DOPPLER RADARNETWORK, AUTOMATED SURFACE MESONET OBSERVATIONS, SATELLITE IMAGERY AND SATELLITE-DERIVED SOUNDINGS.