In the past wind tunnel modeling has not been used directly to obtain hourly concentration estimates for assessing regulatory compliance. It has been used indirectly by determining adjustment factors to correct models such as ISC or AERMOD/PRIME or for model validation and development purposes. The primarily reasons for this are: 1) the time and expense to evaluate all meteorological conditions in the wind tunnel; 2) the wind tunnel can not accurately simulate buoyant plume rise at a reasonable model scale and hence will significantly overestimate maximum concentrations; 3) the wind tunnel has not been shown to adequately simulate the stable or convective boundary layer; and 4) the wind tunnel has not been evaluated against EPA field data bases. This paper will describe a hybrid approach where wind tunnel modeling combined with the latest dispersion and boundary-layer theory (HYWINDMOD) can be used to overcome these problems. First, the concentration as a function of wind speed and wind direction is defined in the wind tunnel at the locations of concern due the source and building/terrain configuration of interest. A function similar to the Gaussian Plume Equation is used to fit the concentration data. Next this functional relation is used to predict hourly concentration levels at all receptors by applying theoretical correction factors to the wind-tunnel predictions to account for plume buoyancy and atmospheric stability effects. To show the validity of the method, wind tunnel studies of the Bowline Point Power plant and an industrial facility were conducted and wind tunnel predictions were compared to hourly field observations. Scale models of both facilities were constructed and evaluated in an environmental wind tunnel. The Bowline Point observational data base is one of the EPA data bases used to validate AERMOD for cases where building downwash effects are important. AERMOD showed good agreement with this data base. The industrial facility was unique in that the monitor and stack were in line with an upwind boiler building corner. When the wind blows along a building corner, two vortices are generated that tend to suppress plume rise by intensifying building downwash and therefore increasing ground-level concentrations. AERMOD was found to underpredict concentrations at this monitor location by about a factor of two. Current AERMOD building downwash equations do not account for corner effects; hence, the observed model underprediction is expected. It should also be noted that the monitor was showing that the 1-hr SO2 NAAQS is exceeded while AERMOD was showed compliance. Concentrations were measured in the wind tunnel at the receptor locations where field observations were obtained. Q-Q plots were then generated documenting the performance of the wind-tunnel predictions versus AERMOD and monitor observations. The comparison demonstrated that the wind tunnel overpredicts ground-level concentrations when buoyancy effects are not included. Including buoyancy and atmospheric stability effects improves the wind tunnel performance and similar overall maximum concentrations as observed in the field are predicted. The wind tunnel showed very good agreement with the field observations for the industrial facility with the corner vortex issue. The Q-Q plots for the two sites are shown below and demonstrate that the HYWINDMOD approach provides concentration estimates that are as good if not better than AERMOD. Since AERMOD was underpredicting at the monitor location for the Industrial Facility, the dispersion model could not be used to evaluate mitigation measures to bring the facility into compliance at the monitor. After demonstrating good agreement between the wind tunnel predictions and the field observations, HYWINDMOD was used to evaluate various mitigation options, including increased stack volume flows and the use of taller stacks. Multiple mitigation measures were identified to ensure that monitored concentrations will fall below the 1-hr NAAQS after the stack modifications are complete.