In this study, a Bayesian framework was applied to update a model of pathogen fate and transport in the indoor environment. Distributions for model parameters (e.g., release quantity of B. anthracis spores, risk of illness, spore setting velocity, resuspension rate, sample recovery efficiency, etc.) were updated by comparing model predictions with measurements of B. anthracis spores made after one of the 2001 anthrax letter attacks. The updating process, which was implemented by using Markov chain Monte Carlo (MCMC) methods, significantly reduced the uncertainties of inputs with uniformed prior estimates: total quantity of spores released, the amount of spores exiting the room, and risk to occupants. In contrast, uncertainties were not greatly reduced for inputs for which informed prior data were available: deposition rates, resuspension rates, and sample recovery efficiencies. This suggests that prior estimates of these quantities that were obtained from a review of the technical literature are consistent with the observed behavior of spores in an actual attack. Posterior estimates of mortality risk for people in the room, when the spores were released, are on the order of 0.01 to 0.1, which supports the decision to administer prophylactic antibiotics. Multivariate sensitivity analyses were conducted to assess how effective different measurements were at reducing uncertainty in the estimated risk for the prior scenario. This analysis revealed that if the size distribution of the released particulates is known, then environmental sampling can be limited to accurately characterizing floor concentrations; otherwise, samples from multiple locations, as well as particulate and building air circulation parameters, need to be measured.