Technical Abstract: Chambers temporarily sealed to the soil surface are an important and for many purposes the only means of measuring trace gas emissions to the atmosphere. However, past interpretations of chamber data systematically underestimated actual emission rates in most applications because they ignored or poorly modeled the effect of the chamber on the gas exchange process. To address this issue, we introduce a time-dependent diffusion model applicable to non-steady-state chamber observations and evaluate its performance, as well as that of previously published models, using concentration data simulated by an independent, 3-dimensional numeric diffusion model. The results demonstrate that commonly cited past measurement practices underestimate emission rates 15% to 25% under most circumstances and that the error varies with chamber height, soil air-filled porosity and flux model used to interpret the concentration data. In contrast, our model proved accurate and readily applicable over the entire range of soil and chamber dimensions evaluated.