A proof-of-concept study was conducted to locate and quantify fugitive emissions within a gas processing plant with inverse micro-scale dispersion modelling. A synthetic wind field and receptor observations were generated for a simplified model of a gas processing plant. Source reconstruction was performed with an objective function that measured the misfit of candidate emission source distributions. The objective function gradient was evaluated with adjoint sensitivity analysis, and the candidate emission source distribution that minimized the objective function was found using the
L-BFGS-B optimization algorithm. The inverse dispersion model was tested under non-ideal conditions such as multiple emission sources, unfavourable receptor coverage, and the addition of observational noise. In an effort to mitigate the prediction of false emission sources, objective function regularization and emission source filtering were investigated. It was found that a combination of regularization and filtering achieved the best estimates of the emission locations and total emission rates.