Forest fires can significantly contribute to unhealthy air pollution levels. For this reason, their impacts are frequently quantified using multidimensional air quality modeling systems. Recently, research efforts focused on characterizing the emissions from forest fires. However, little attention has been given to the sensitivity of modeled surface pollutant concentrations to meteorological inputs required by gridded photochemical air quality models. In this work, we explore these sensitivities by modeling the impacts of wildland fires on air quality for real fire episodes affecting an urban location in the Southeastern US. The simulations were performed using the Community Multi-scale Air Quality modeling system (CMAQ) and driven by meteorological inputs produced by the Weather Research and Forecasting Model (WRF) and emissions processors. The sensitivities of pollutant surface concentrations to different meteorological inputs were quantified. Specifically, the effect of plume rise, and the resulting vertical distribution of emissions, on ground level pollutant concentrations was analyzed. Plume rise calculations are reflective of fire characteristics as well as atmospheric stability conditions, and require adequate numerical representations of these processes to achieve realistic results. Additionally, high uncertainty in short-term wind forecasting is expected with weather prediction models. Therefore, sensitivities of ground level pollutant concentrations to wind speed and direction were calculated. Finally, an assessment of the sensitivity of surface pollutant levels to modeled atmospheric mixing height was carried out. Analyses involved brute force and decoupled direct methods to quantify sensitivities to the different air quality model inputs. Through this study, the significance of different meteorological inputs required in regional air quality modeling is evaluated. The results provide valuable information about the degree to which current simulations of fire impacts on air quality are constrained by the uncertainties in meteorological fields produced by prognostic weather models and emissions processors. Findings of this works are potentially applicable to simulations of the impacts from other emissions sources with distinct plumes.