Environmental models are frequently used within regulatory and policy frameworks to estimate environmental metrics that are difficult or impossible to physically measure. As important decision tools, the uncertainty associated with the model outputs should impact their use in informing regulatory decisions and scientific inferences. In this paper, we present a case study illustrating a process for dealing with a key issue in the use and application of air quality models, the additional error in annual mean aggregations resulting from imputation of missing data from model data sets. The case study is based on the US Environmental Protection Agency’s Multi-layer Model, which estimates the hourly dry deposition velocity of air pollutants based on hourly measurements of meteorology and site characteristics. A simulation was implemented to evaluate the effect of substituting historical hour-specific average values for missing model deposition velocity predictions on annual mean aggregations. Sensitivity studies were performed to test the effects of different missing data patterns and evaluate the relative impact of the substitution procedure on annual mean SO2 deposition velocity estimates. The substitution procedure was shown to result generally in long-term unbiased estimates of the annual mean and contributed less than 20% additional error to the estimate even when all data were missing. Consequently, it may be possible to use the historical record of deposition velocities to provide reasonably accurate and unbiased annual estimates of deposition velocities for years without meteorological measurements.