METHANE IS THE SECOND MOST IMPORTANT ANTHROPOGENIC GREENHOUSE GAS BUT ITS EMISSION ESTIMATES ARE HIGHLY UNCERTAIN BECAUSE SPACE-BASED OBSERVATIONS LACK SUFFICIENTLY FINE RESOLUTIONS TO UNAMBIGUOUSLY DISENTANGLE SOURCE CATEGORIES. RECENT ADVANCEMENTS IN AIRBORNE REMOTE SENSING SUCH AS FROM THE AIRBORNE VISIBLE/INFRARED IMAGING SPECTROMETER (AVIRIS-NG) AND THE HYPERSPECTRAL THERMAL EMISSION SPECTROMETER (HYTES) THAT USE METHANE ABSORPTION FEATURES AT 2.3 MICRONS AND 7.6 MICRONS ENABLE QUANTITATIVE RETRIEVALS OF ABUNDANCES AT 1-5M SPATIAL RESOLUTION ALLOWING FOR THE DIRECT DETECTION OF METHANE PLUMES AND THEIR ORIGIN. HOWEVER ACCURATE INVERSION FROM DETECTED METHANE CONCENTRATION TO THE ACTUAL EMISSION RATE AT THE SOURCE HAS SO FAR REQUIRED SIMULTANEOUS GROUND-BASED MEASUREMENTS OF BOUNDARY-LAYER CONDITIONS SUCH AS WIND SPEEDS AND SURFACE HEAT FLUXES. THIS HINDERS THE ABILITY TO QUANTIFY METHANE POINT SOURCES FROM REMOTELY-SENSED DATA ALONE. USING LARGE EDDIES SIMULATIONS (LES) TO SIMULATE REALISTIC METHANE DISTRIBUTIONS IN THE ATMOSPHERIC BOUNDARY LAYER I INITIALLY FOUND THAT THE SPATIAL PLUME PATTERNS ARE DISTINCT UNDER VARYING WIND AND HEAT FLUX CONDITIONS. IN MY OPTIMEEM PROJECT I PROPOSE TO APPLY MODERN MACHINE LEARNING (ML) ALGORITHMS TO IDENTIFY AND CHARACTERIZE PLUME SHAPES AND FLUX RATES BASED ON REAL NASA OBSERVATIONAL DATA AND A LARGE TRAINING DATASET OF LES SIMULATIONS UNDER VARYING BOUNDARY LAYER CONDITIONS. THE POWERFUL MACHINE LEARNING APPROACHES HAVE THE POTENTIAL TO COMPLETELY RECONSTRUCT FLUX RATES ONLY BASED ON PLUME SHAPE AND METHANE ENHANCEMENTS WITHOUT THE NEED OF IN SITU SURFACE MEASUREMENTS. I WILL FIRST TEST AND REFINE THE APPROACHES BASED ON SIMULATED BEFORE APPLYING THEM ON REAL DATA OBTAINED IN RECENT NASA AIRBORNE CAMPAIGNS. A SPECIFIC FOCUS WILL BE TO EVALUATE THE POTENTIAL OF BOTH SHORT-WAVE AS WELL AS THERMAL RETRIEVALS BY TAKING THE INSTRUMENT AVERAGING KERNELS AND RANDOM NOISE INTO ACCOUNT. I ENVISION THAT OPTIMEEM WILL BE A CRUCIAL STEPPING STONE TOWARDS THE AUTOMATION OF THE IDENTIFICATION AND QUANTIFICATION OF LOCAL POINT SOURCES OVER WIDE GEOGRAPHICAL AREAS FROM NASA AIRBORNE HIGH-RESOLUTION METHANE REMOTE SENSING BUT ALSO BEING APPLICABLE TO FUTURE HIGH-RESOLUTION SPACE-BORNE MISSIONS AS HIGHLIGHTED BY THE 2018 NAS DECADAL SURVEY CONSENSUS REPORT.