Inversion studies statistically combine carbon dioxide observations and atmospheric models to infer fluxes of carbon between the atmosphere and the Earth's surface. The covariance between carbon dioxide and carbon monoxide has been shown to provide a powerful tool for improving the accuracy of these inverted fluxes. However, this approach is not as useful for studies of the growing season when the biosphere fluxes are in question. It has been hypothesized that carbonyl sulfide could provide a novel tracer for extracting information about carbon dioxide. Here we demonstrate how carbonyl sulfide observations can be exploited in a four dimensional variational data assimilation framework to obtain improved carbon dioxide surface flux estimates during the growing season over North America. We use aircraft observations from the NASA INTEX-A experiment along with the STEM regional atmospheric transport model. In the future, observations of carbonyl sulfide made along with carbon dioxide could provide a great benefit to unraveling the processes associated with carbon dioxide fluxes.