ONE OF THE ONLY DIRECT ANTHROPOGENIC IMPACTS IN THE UPPER TROPOSPHERE ARE THE ICE CLOUDS WHICH CAN FORM IN THE AIRCRAFT ENGINE EXHAUST UNDER CERTAIN ATMOSPHERIC CONDITIONS. THESE CONTRAIL CIRRUS PERSIST FOR UP TO 24 HOURS SPREADING FROM LINE-SHAPED TRAILS INTO CIRRUS CLOUDS WHICH CAN BE KILOMETERS WIDE. ALTHOUGH EACH INSTANCE TYPICALLY LASTS FOR LESS THAN A DAY CONTRAIL CIRRUS HAVE BEEN ESTIMATED TO CAUSE CLIMATE IMPACTS EQUIVALENT TO THOSE ASSOCIATED WITH ALL CO2 EMITTED FROM AIRCRAFT IN THE PAST 80 YEARS. HOWEVER DUE TO SIGNIFICANT UNCERTAINTIES IN OVERALL COVERAGE AND OPTICAL PROPERTIES OF CONTRAIL CIRRUS THERE REMAINS AN ORDER-OF-MAGNITUDE UNCERTAINTY IN THEIR OVERALL IMPACT. MEANWHILE EFFECTS ON ATMOSPHERIC COMPOSITION AND AIR QUALITY HAVE YET TO BE INVESTIGATED. THIS PROJECT AIMS TO ADDRESS THESE ISSUES BY (I) VALIDATING AND REFINING THE CONTRAIL EVOLUTION AND RADIATION MODEL (CERM) THROUGH AN OBSERVATIONAL APPROACH; AND (II) APPLYING CERM AND ATMOSPHERIC CHEMICAL MODELING TO ASSESS PAST CURRENT AND FUTURE CONTRAIL COVERAGE AND ITS ATMOSPHERIC IMPACTS ON CLIMATE AND SURFACE AIR QUALITY. FOR THIS PURPOSE WE WILL FIRST LEVERAGE A NOVEL MACHINE LEARNING-BASED APPROACH TO QUANTIFY CONTRAIL CIRRUS COVERAGE OVER THE CONTINENTAL UNITED STATES BASED ON A FULL YEAR OF DATA FROM THE GOES-16 GEOSTATIONARY SATELLITE. BY TRAINING A CONVOLUTIONAL NEURAL NETWORK ON PREVIOUSLY-CLASSIFIED IMAGERY WE WILL PRODUCE AN OPEN-SOURCE GENERALIZABLE MODEL OF CONTRAIL CIRRUS IDENTIFICATION. THIS MODEL WILL BE USED FOR A FIRST MACHINE-LEARNING BASED OBSERVATIONAL STUDY TO IDENTIFY CONTRAIL CIRRUS COVERAGE AND ITS SPATIAL DIURNAL AND SEASONAL TRENDS AT HIGH RESOLUTION (~2KM). THIS DATA WILL THEN BE USED TO CALIBRATE CERM WHICH ESTIMATES CONTRAIL CIRRUS COVERAGE AND PROPERTIES AT THE REGIONAL TO GLOBAL LEVEL. ONCE CALIBRATED AND VALIDATED CERM WILL BE APPLIED TO QUANTIFY GLOBAL CONTRAIL CIRRUS COVERAGE AT A SPATIAL RESOLUTION OF ~50 KM FOR EACH HOUR IN THE CALIBRATION YEAR. THIS COVERAGE ESTIMATE WILL BE MADE PUBLICLY AVAILABLE ALONG WITH AN ESTIMATE OF THE TOTAL GLOBAL AND REGIONAL RADIATIVE FORCING RESULTING FROM CONTRAIL CIRRUS. FOLLOWING THIS INITIAL ESTIMATE OF CURRENT-DAY COVERAGE A MULTI-DECADE SIMULATION WILL BE EMPLOYED TO DERIVE THE FIRST CONTRAIL CIRRUS INVENTORY FOR THE YEARS 1980 TO 2050. PAST METEOROLOGICAL AND CLIMATOLOGICAL CONDITIONS WILL BE TAKEN FROM THE EXISTING MERRA-2 METEOROLOGICAL DATA SET AND FUTURE CONDITIONS WILL BE ESTIMATED WITH THE COMMUNITY EARTH SYSTEM MODEL UNDER DIFFERENT ECONOMIC SCENARIOS. USING HISTORICAL AVIATION SCHEDULE DATA AND PROJECTED FUTURE AVIATION ACTIVITY COVERAGE AND PROPERTIES (E.G. LIFETIME OPTICAL DEPTH) OF CONTRAIL CIRRUS WILL BE SIMULATED FOR THE ENTIRE PERIOD. THE DATA WILL BE ANALYZED WITH REGARD TO DEPENDENCE ON TRAFFIC REGIONAL VARIATIONS IN FORMATION LIKELIHOOD AND THE EFFECT OF CLIMATE CHANGE ON CONTRAIL CIRRUS. FINALLY WE WILL QUANTIFY THE IMPACTS OF CONTRAIL CIRRUS ON CLIMATE ATMOSPHERIC COMPOSITION AND AIR QUALITY OVER THE SAME 70 YEAR PERIOD. THIS WILL BE ENABLED BY DIRECT SIMULATION OF ATMOSPHERIC COMPOSITION IN THE GEOS-CHEM HIGH PERFORMANCE CHEMISTRYTRANSPORT MODEL USING THE CONTRAIL INVENTORIES PRODUCED IN PREVIOUS STEPS. WE EXPECT THIS PROJECT TO HAVE SIGNIFICANT IMPACTS FOR SCIENTISTS DECISION MAKERS AND THE PUBLIC. IN PARTICULAR IT WILL (I) FOR THE FIRST TIME CALIBRATE AND VALIDATE CONTRAIL MODELS THROUGH OBSERVATIONAL DATA; (II) PRODUCE THE FIRST SET OF COMPREHENSIVE CONTRAIL INVENTORIES FOR USE IN ATMOSPHERIC MODELS; (III) REDUCE THE LEVEL OF UNCERTAINTY IN CONTRAIL COVERAGE FOR PAST PRESENT AND FUTURE CONDITIONS; (IV) ADVANCE ALGORITHMIC KNOWLEDGE REQUIRED FOR NEAR REAL-TIME CONTRAIL IDENTIFICATION; AND (V) PRODUCE THE FIRST ESTIMATES OF IMPACTS ON ATMOSPHERIC COMPOSITION AIR QUALITY AND HUMAN HEALTH RESULTING FROM CONTRAIL CIRRUS.