CALMET (California Meteorological Model), which is a meteorological subroutine for an air quality dispersion mode (CALPUFF; California Puff Model), closely related with the land surface structure. In this study, the sensitivity of micro-meteorological parameters including wind speed and roughness height, Albedo, Bowen ratio, soil heat flux, and leaf area index were closely evaluated with change of land-use in relation to urban development. As a result, although no consistent dependence of roughness height on surface wind was found, it showed that high value of surface roughness could lead to the increase of friction velocity, influence the Monin-Obukhov length and the mixing height. At the same time, the increasing Albedo reduced friction velocity and mixing height. Thus, it was concluded for the CALMET modelling that it is necessary to first define the roughness height, Albedo, and Bowen ratio according to land-use.

There are many pollutants emitted into the air. Some of these pollutants have a malodor. Unlike other pollutants, people are able to detect and feel discomfort when this type of pollutant becomes high peak concentration instantaneously. In this sense, the peak concentration has an important meaning in the odor management and modeling. In previous odor modeling, the peak concentration was calculated by correcting the one-hour average concentration using the correlation equation. This study was carried out to find appropriate method to predict the peak concentration using meteorological input data of high time resolution in the odor modeling. It show that the peak concentration could be directly calculated from the dispersion modeling without using the correction equation when fine time scales such as 1 min or less time intervals are used as the meteorological input.

In order to improve an accuracy of the real-time odor dispersion modeling system, a sensitivity of CALMET model with different input meteorological data was studied. The performance of CALMET model was tested by comparing the model predictions with the observations at the Daedeok Industrial Complex in Daejeon Metropolitan City. It is shown that the CALMET model with WRF (Weather Research and Forecasting) input data of GFS (Global Forecast System) depicts the measurements better than that of RDAPS (Regional Data Assimilation and Prediction System). The CALMET model could be further improved by selecting options of Divergence minimization, Froude number adjustment and Slope flows without choosing Kinematic Effects in the modeling procedure.