Methane is the second-strongest anthropogenic greenhouse gas, whose concentrations have more than doubled since pre-industrial times. Due to its short atmospheric lifetime, it is promising for climate change mitigation. This thesis focuses on estimating methane emissions from time-averaged satellite-measured concentration maps. The current method to do that systematically underestimates emission rates by 40%, and has an error estimate larger than 100% for most sources. I developed a new mass-balance method to reduce these errors, taking into account emissions moving from the source region into the surroundings. Its performance was tested using a year-long atmospheric transport model run over the United States. The new method has a lower emission underestimation of only 14-19%, while the errors are reduced by 60-70% compared to the old method. Direct application of the new method on TROPOMI data over two known large emission regions gives emission rates consistent with recent literature.