A comparison between integrated meteo-diffusive models: RAMS-AERMOD/RAMS-HYPACT

Dispersion models require information on meteorological variables that are not routinely measured; the process of estimating the quantities required from routinely available data is known as meteorological pre-processing. The main purpose of meteorological pre-processing is to estimate atmospheric stability and boundary layer height. The AERMOD modelling system consists of two pre-processors and a dispersion model. AERMET is a meteorological pre-processor and AERMAP is a terrain pre-processor. The dispersion model AERMOD has the capability of utilising more advanced meteorological information such as that produced by a mesoscale model such as the non-hydrostatic meteorological prediction model RAMS (Regional Atmospheric Model System). On the other hand, RAMS drives an advanced dispersion code called HYPACT, Hybrid Particle and Concentration Transport Model. This paper presents a comparison of pollutant concentrations predicted by AERMOD and by HYPACT. Both models were coupled to RAMS that has been run with activated microphysics and three nested grids, the finest grid having been set to 1 kilometre cell width. We developed a procedure to extract data from RAMS, ingest it into AERMOD meteorological pre-processor AERMET, and then run AERMOD to produce concentration output. In addition to the standard variables, namely wind, temperature, and moisture, we have extracted from RAMS many other parameters such as PBL heights, surface roughness, cloud fraction, friction velocity, and surface heat. This allowed us also to bypass the meteorological pre-processor AERMET by sending the pre-processed surface and upper air parameters as input to the dispersion model AERMOD. We have also considered a set of AERMOD and HYPACT runs using RAMS data as input and compared the results with AERMOD runs made using surface and upper-air data from Local Airbase. The discussion of results outlines possible advantages and limitations of using the output of mesoscale models as input data for dispersion models.