Present computer models used to predict vapor cloud length, concentration and duration at air force installations do not properly account for the effects of structures. Structures can act to trap the cloud and increase the cloud dimensions, thereby possibly reducing downwind concentrations and increasing the cloud's lateral extent. Recent research conducted by CPP has shown that the trapping and increased dispersion effect varies significantly with site geometry. Use of numerical models alone will not likely provide sufficiently accurate concentration estimates. To provide accurate estimates, CPP is proposing to develop a hybrid modeling approach. The hybrid model is developed using wind tunnel modeling for near- field dispersion estimates and standard regulatory or heavier-than- air models for farfield estimates. The farfield estimates are obtained by initializing the numerical models using the wind tunnel measurements at the point where the cloud is significantly larger than the structures associated with the installation.
