The oxidation of hydrocarbons plays a key role in the atmosphere. All hydrocarbons including methane (CH₄) react in air to form carbon monoxide (CO) and then carbon dioxide (CO₂) through a series of reactions. CO and CH₄ are the largest sink for hydroxyl radical that controls the oxidizing capacity of the troposphere, and CO₂ and CH₄ are the most important greenhouse gases. Quantifying emission sources of these species and understanding the underlying sinks are urgently needed to support atmospheric chemistry research. Here we use a multi-species Bayesian inversion framework that jointly assimilates observations of related tracers within hydrocarbon oxidation to retrieve the distribution of CO and CH₄ fluxes at global and regional scales. By analyzing spatial and temporal details and comparing the Bayesian inversion results to bottom-up estimates, the key trends and drivers of emission fluxes are investigated to achieve robust and consistent estimates. For CO₂, we use the satellite measurement from OCO-2 retrievals to quantify fossil fuel CO₂ emissions at urban scale. I will introduce the Bayesian inversion framework built upon OCO-2 data and present preliminary inversion results of CO₂ emissions with a case study.