Predicting air quality near roadway intersections through the application of a Gaussian puff model to moving sources

With substantial health and economic impacts attached to many highway-related projects, it has become imperative that the models used to assess air quality be as accurate as possible. The United States (US) Environmental Protection Agency (EPA) currently promulgates the use of CAL3QHC to model concentrations of carbon monoxide (CO) near roadway intersections. This model uses steady-state and macroscopic methods to model the physical phenomena (e.g., emission rates, atmospheric dispersion, etc.) occurring at intersections. These methods are not straightforward and unintuitive for the users. Therefore, this project investigated the possibility of developing a model that is theoretically more realistic and flexible than CAL3QHC.; The new model entitled, Traffic Air Quality Simulation Model (TRAQSIM), uses a microscopic approach by modeling vehicle movements and dispersion in a simulation environment. Instead of steady-state plume equations used in the CAL3QHC model, TRAQSIM uses a discrete puff methodology that can be used to model time-based dispersion of pollutants. Most of the components incorporated into TRAQSIM have been drawn from existing methodologies and therefore, are not new. However, the combination of these different methods into a single integrated model is new and presents a novel approach to such a model.; Initial verification and sensitivity/trend studies of the model indicate that TRAQSIM uses reasonable/realistic traffic parameters and behaves intuitively correct. A validation study showed that TRAQSIM produces good results when compared to actual measured data with an overall R2 value of 0.605 for 11 scenarios comprising 264 data points. Although most statistical parameters showed CAL3QHC agrees better overall with measured data (R 2 value of 0.721), the comparisons were mixed on a scenario-by-scenario basis; that is, CAL3QHC showed better results for 6 scenarios and TRAQSIM showed better results for 5 scenarios. Additional tests with larger datasets, which were beyond the scope of this work, could be conducted to obtain more definitive conclusions and allow further development of TRAQSIM.; While CAL3QHC is a mature model that has been developed over many years, TRAQSIM is new and has much more potential for improvement. The physical parameters used in TRAQSIM allow it to be more directly (more logically) improved than the approximations used in CAL3QHC. In addition, although the fundamental-level modeling in TRAQSIM make it a more complex model internally, it is much more intuitive for the user to understand and use.