Usnea hirta and Xanthoparmelia cumberlandia are commonly used as bio-monitors of air quality. In order to more accurately and efficiently determine the distribution of these two sensitive indicator species, we have developed a probabilistic distribution map as a function of 9 macroclimatic and topographic variables for the White River National Forest, Colorado using Non-Parametric Multiplicative Regression (NPMR) analysis. Furthermore, we also developed a logistic regression (LR) model for X. cumberlandia in order to evaluate the strengths and limitations of the NPMR model. The best model for U. hirta included four variables - solar radiation, average monthly precipitation, average monthly minimum and maximum temperature (log β = 3.68). The presence rate for U. hirta based on field validated test sites was 45.5%, 65.4%, and 70.4% for low, medium, and high probability areas, respectively. The best model for X. cumberlandia generated by both NPMR and LR involved the same variables - solar radiation, average monthly maximum temperature, average monthly precipitation, and elevation as the best predictor variables (log β = 5.10). The occurrence rate for X. cumberlandia using the NPMR model was 32%, 44.4%, and 20% for the low, medium, and high probability areas respectively while the LR model had 26%, 50%, and 38% for low, medium and high probability areas respectively. Although the LR model predicted a smaller high probability area compared to the NPMR model there was substantial overlap between the two. The U. hirta model performed better than the X. cumberlandia model. The reduced performance of our model especially for X. cumberlandia may be due in part to the absence of field measured data in the development of the model. Our study also suggested that the northeast and western part of the forest should be preferentially considered for establishing future air quality bio-monitoring reference sites. Finally, in the future a well defined sampling design with sufficient sampling sites, field measured predictor variables, and microclimatic data should be used in the development of predictive models.