WE PROPOSE TO STUDY THE CHEMISTRY AND EMISSIONS OF NOX A FAMILY OF CHEMICALS WITH IMPORTANT INFLUENCES ON AIR QUALITY AND CLIMATE. OUR PROJECT WILL ADVANCE METHODS FOR USING SPACE-BASED NO2 OBSERVATIONS AND APPLY THE ADVANCES TO DEVELOPING A DEEPER UNDERSTANDING OF PROCESSES GOVERNING EPISODIC EMISSIONS FROM LIGHTNING AND SOILS. WE WILL ASSESS TRENDS IN THE EMISSIONS AND IN THE LIFETIME OF NOX IN URBAN CENTERS. IN ADDITION WE WILL PURSUE OPTIMIZED METHODS FOR COMPUTATIONALLY EFFICIENT RETRIEVALS THAT BUILD ON THE BEHR V3.0B RETRIEVAL; A RETRIEVAL THAT INCLUDES DAILY HIGH RESOLUTION (12 KM) A PRIORI PROFILES. TAKING FULL ADVANTAGE OF CURRENT HIGH RESOLUTION CHEMICAL SENSORS SUCH AS TROPOMI AND THE SOON TO BE LAUNCHED FLEET OF HIGH RESOLUTION GEOSTATIONARY SENSORS (E.G. TEMPO) WILL REQUIRE EVEN HIGHER RESOLUTION A PRIORI PROFILES. TO SUPPORT THE INTERPRETATION OF OBSERVATIONS FROM THESE NEW INSTRUMENTS WE WILL PURSUE MODERN DATA SCIENCE APPROACHES TO COMBINING THE OBSERVATIONS WITH HIGH RESOLUTION HOURLY WEATHER MODELS AND A SINGLE YEAR OF A HIGH RESOLUTION INSTANCE OF A CHEMICAL-WEATHER MODEL. MANY DAYS HAVE SIMILAR A PRIORI PROFILES AND WE ARE CONVINCED THAT ARTIFICIAL NEURAL NETWORKS AND OTHER MODERN TOOLS SUCH AS THOSE USED FOR EFFICIENT FACIAL RECOGNITION CAN IDENTIFY SIMILARITIES IN PROFILES THAT ALLOW ACCURATE ESTIMATION OF TRACE GAS COLUMNS WITHOUT EXPLICIT COMPUTATION OF A PROFILE AT THE RESOLUTION OF THE OBSERVATION. THESE RETRIEVAL IDEAS WILL BE COUPLED TO RESEARCH ON USE OF DATA ASSIMILATION TO SUPPORT IMPROVED WEATHER AND AIR QUALITY FORECASTS.