Comparison of Lake-Breeze Model Simulations with Tracer Data

Abstract

During the 1980s there were numerous occurrences of ozone exceedances of federal standards in the Lake Michigan area. An intensive field program was undertaken in order to gain insight into the dispersion patterns in the area. A field experiment, which included a tracer release, was conducted on 16 July 1991 and provided the data used to test the skill of a meteorological and dispersion modeling system. Specifically, the Regional Atmospheric Modeling System (RAMS) provided meteorological input to a Lagrangian Particle Dispersion Model. The RAMS model was run at a variety of grid spacings and configurations, and the output was compared to observations for each simulation. It was found that a minimum ?x of 4 km and a 3D setup were needed to accurately simulate the meteorology of 16 July 1991. The dispersion model was then run for each meteorological simulation and compared to aircraft data. Again the 3D simulations produced the best correlations to tracer data. It was also found in the 2D simulations that the highest correlation to meteorological variables was derived from the finest-resolution meteorological simulation. Finally, a commonly used Gaussian plume model, the Industrial Source Complex Model, was run using the best RAMS produced meteorology. It was clear from the results that this model is not applicable to this case day and, in fact, failed to produce a nonzero correlation to the aircraft data.