DEVELOPMENT OF MESOSCALE MODELS FOR MINICOMPUTER

WE HAVE DEVELOPED A MESOSCALE NUMERICAL PREDICTION MODEL WHICH RUNS ON A MICROCOMPUTER, CURRENTLY AN HP-9000/520. FOR OPERATIONAL APPLICATIONS OF THIS MODEL IN REAL TIME, WE WILL REQUIRE A DATA COMMUNICATIONS LINK TO ACCESS FAA CIR- CUIT 604. THE DATA RECEIVED VIA THIS CHANNEL CONSIST OF SYNOPTIC RADIOSONDE AND SURFACE OBSERVATIONS. BEFORE THESE DATA CAN BE ASSIMILATED INTO OUR NUMERICAL PREDICTION MODEL THEY WILL HAVE TO UNDERGO QUALITY CONTROL PROCEDURES. SUCH PROCEDURES HAVE NOT YET BEEN DEVELOPED FOR MINI- AND MICRO- COMPUTER ENVIRONMENTS. WE DEEM IT ESSENTIAL THAT A MESO SCALE MODEL FOR "INDIVIDUALIZED" WEATHER PREDICTIONS CAN RUNLOCALLY AT WFOS (WEATHER FORECASTING OFFICES) WITH ORIGINAL DATA AND NOT ONLY WITH OBJECTIVE ANALYSIS PRODUCTS RECEIVED FROM A CENTRAL FACILITY. IN OUR EXPERIENCE, SUCH ANALYSES TEND TO BE SMOOTHED EXCESSIVELY TO THE POINT WHERE THE CAUSES FORM IMMINENT MESOSCALE WEATHER DEVELOPMNENT HAVE BEEN OBLITERATED. UNDER PHASE-I, WE PROPOSE TO DEVELOP AN OPTIMUM-COST DESIGN OF HARDWARE AND SOFTWARE FOR DATA RECEP-TION, DATA QUALITY CONTROL, AND MODEL INITIALIZATION. THIS DESIGN WILL INCLUDE THILITY OF USING ORIGINAL OBSERVA TIONAL DATA AS WELL AS IMPORTED OBJECTIVE ANALYSES OF METEOROLOGICAL DATA FIELDS.