Inverse modeling using the Bayesian least squares method is applied to better constrain the sources and sinks of atmospheric methyl chloride (CH3Cl) using observations from seven surface stations and eight aircraft field experiments. We use a three-dimensional global chemical transport model, the GEOS-Chem, as the forward model. Up to 39 parameters describing the continental/hemispheric and seasonal dependence of the major sources of CH3Cl are used in the inversion. We find that the available surface and aircraft observations cannot constrain all the parameters, resulting in relatively large uncertainties in the inversion results. By examining the degrees of freedom in the inversion Jacobian matrix, we choose a reduced set of parameters that can be constrained by the observations while providing valuable information on the sources and sinks. In particular, we resolve the seasonal dependence of the biogenic and biomass-burning sources for each hemisphere. The in situ aircraft measurements are found to provide better constraints on the emission sources than surface measurements. The a posteriori emissions result in better agreement with the observations, particularly at southern high latitudes. The a posteriori biogenic and biomass-burning sources decrease by 13 and 11% to 2500 and 545 Gg yr-1, respectively, while the a posteriori net ocean source increases by about a factor of 2 to 761 Gg yr-1. The decrease in biomass-burning emissions is largely due to the reduction in the emissions in seasons other than spring in the Northern Hemisphere. The inversion results indicate that the biogenic source has a clear winter minimum in both hemispheres, likely reflecting the decrease of biogenic activity during that season.