The Atmospheric Release Advisory Capability (ARAC) provides real-time dose assessments for airborne pollutant releases. ARAC is currently in the process of developing an entirely new suite of models and system infrastructure. Diagnostic and dispersion algorithms are being created in-house and a prognostic model NO-RAPS, imported from the Naval Research Laboratory, Monterey, is currently being adapted to ARAC`s needs. Diagnostic models are essential for an emergency response capability since they provide the ability to rapidly assimilate available meteorological data and generate the mass-consistent three-dimensional wind fields required by dispersion models. The resulting wind fields may also serve to initialize and validate prognostic models. In general, the performance of diagnostic models strongly correlates with the density and distribution of measurements in the area of interest and the resolution of the terrain. problem, data can be extracted from user-specified databases within a region defined by a metdata grid. Typically the data collection region will cover a geographic domain significantly larger than the area involved in the dispersion simulation in order to provide the most complete set of meteorological information relevant to the problem. This also permits the user to redefine the problem grid size and location, within limits, without reaccessing the meteorological data …