INTENSE CONVECTIVE STORMS ARE SUSTAINED BY STRONG UPDRAFTS THAT GENERATE A VARIETY OF WEATHER HAZARDS INCLUDING HAIL TORNADOES DAMAGING WIND AIRCRAFT TURBULENCE AND ICING IN ADDITION TO CLIMATE IMPACTS VIA THEIR INJECTION OF MOISTURE AEROSOLS AND TRACE GASSES INTO THE UPPER TROPOSPHERE AND LOWER STRATOSPHERE (UTLS). GRAVITY WAVE BREAKING OCCURS ATOP STORMS WITH TROPOPAUSE-PENETRATING UPDRAFTS IN ENVIRONMENTS WITH SUFFICIENT WIND SHEAR WHICH CAUSES CIRRUS CLOUD INJECTION INTO THE STRATOSPHERE ABOVE THE TROPOSPHERIC ANVIL CLOUD TOP. THESE ABOVE-ANVIL CIRRUS PLUMES (AACP) HAVE BEEN LINKED TO THE HIGHEST WV EVER OBSERVED IN THE MID-LATITUDE UTLS AND ARE ROUTINELY FOUND ATOP THE MOST SEVERE STORMS ON ALL CONTINENTS. AN AACP IS THE STRONGEST GEOSTATIONARY SATELLITE-OBSERVED INDICATOR OF A SEVERE STORM DOCUMENTED TO DATE AND SEVERE WEATHER OUTBREAKS CAN FEATURE DOZENS OF LONG-LIVED AACP-PRODUCING STORMS. TROPOPAUSE-PENETRATING UPDRAFTS OFTEN CALLED OVERSHOOTING TOPS (OTS) AND AACPS EXHIBIT UNIQUE PATTERNS WITHIN THE LATEST GENERATION OF GEOSTATIONARY (GEO) IMAGERY BEING COLLECTED WITH HIGHER SPATIO-TEMPORAL RESOLUTION THAN EVER BEFORE. RAPIDLY RISING AIR WITHIN OTS GENERATES COLD TEMPERATURES TEXTURE IN VISIBLE IMAGERY AND HIGH LIGHTNING FLASH RATES THAT CAN BE DETECTED BY THE NEW GOES-16/17 GEOSTATIONARY LIGHTNING MAPPER AND OTHER FUTURE GEO LIGHTNING IMAGERS. AACPS ARE NOT RISING AND RESIDE IN THE LOWER STRATOSPHERE ALLOWING THEM TO ADJUST TO AN ENVIRONMENT THAT CAN BE 20 K WARMER THAN THE UNDERLYING ANVIL TOP MAKING AACPS EVIDENT TO THE HUMAN EYE AND DETECTABLE BY PATTERN RECOGNITION ALGORITHMS. THOUGH PREVIOUS EFFORTS HAVE BEEN DEVOTED TO AUTOMATED OT AND AACP DETECTION NEW OPEN-SOURCE NEURAL NETWORK-BASED DEEP LEARNING METHODS HAVE RECENTLY EMERGED THAT ARE OPTIMAL FOR SPATIAL PATTERN RECOGNITION. SUCH METHODS COULD SIGNIFICANTLY IMPROVE OUR ABILITY TO DETECT AND ANALYZE THESE IMPORTANT CLOUD-TOP PATTERNS. PREVIOUS RESEARCH HAS IDENTIFIED 10 000 S OF STORM CELLS WITH OTS AND AACPS TRACKED USING GEO AND WEATHER RADAR DATA OVER THE U.S. WE BELIEVE THAT DEEP LEARNING METHODS WHEN TRAINED WITH A LARGE DATABASE OF KNOWN OT AND AACP EVENTS OFFER THE BEST POSSIBLE OPPORTUNITY TO DETECT THESE PATTERNS. PIXEL-SCALE OT AND AACP IDENTIFICATIONS ARE NEEDED TO IDENTIFY PROCESSES THAT GENERATE AACPS AND TO ADVANCE OUR KNOWLEDGE OF THEIR FREQUENCY GEOGRAPHIC LOCATION ASSOCIATION WITH SEVERE WEATHER AND AVIATION WEATHER HAZARDS CLOUD HEIGHT AND IMPACT ON UTLS COMPOSITION. THE PRIMARY OBJECTIVE OF THIS PROPOSAL IS TO USE GEO SATELLITE DATA GRIDDED WEATHER RADAR VOLUMES MODEL DATA AND HUMAN-BASED AACP IDENTIFICATIONS TO DETECT AND IMPROVE ANALYSES OF HAZARDOUS CONVECTION. THIS NOVEL COMBINATION OF GEO VISIBLE IR AND LIGHTNING DETECTION DATA PROCESSED WITHIN STATE-OF-THE-ART DEEP LEARNING METHODS WILL PROVIDE 1) THE MOST COMPREHENSIVE UNDERSTANDING OF THE STRENGTHS/LIMITATIONS OF DEEP LEARNING FOR ANALYSIS AND DETECTION OF HAZARDOUS CONVECTION 2) OPEN-SOURCE AUTOMATED STORM DETECTION SOFTWARE THAT CAN BE LEVERAGED BY THE EARTH SCIENCE RESEARCH COMMUNITY AND INDUSTRY AND 3) NEW INSIGHTS INTO THE PHYSICAL PROCESSES THAT LEAD TO CLOUD TOP SPATIO-TEMPORAL EVOLUTION DEPICTED BY CURRENT AND FUTURE GEO IMAGERS. REMOTE SENSING DATA FUSION ENABLED BY STORM CELL TRACKING AND GOES STEREOSCOPIC ANALYSES PROVIDES THE OPPORTUNITY TO ANALYZE CLOUD TOP HEIGHT AND MICROPHYSICS IN NEW WAYS TO ADDRESS UNCERTAINTIES IN HOW CONVECTIVE STORM PROCESSES IMPACT CLOUD-TOP APPEARANCE. IN ADDITION TO ADDRESSING SEVERAL NASA AND SCIENCE COMMUNITY RESEARCH GOALS A DIVERSE SET OF INTERNATIONAL AND INDUSTRY PARTNERS SEEK TO EVALUATE DETECTION PRODUCTS IN REAL TIME FOR OPERATIONAL FORECASTING AND AVIATION INDUSTRY APPLICATIONS AS WELL AS TO SET REQUIREMENTS AND BECOME EARLY ADOPTERS FOR THE HAZARDOUS STORM DETECTION SOFTWARE.