Sub-grid scale terrain effects are difficult to account for in atmospheric air pollution models. This is because emissions sources (area, roads, biogenic, pt source stacks) may be at substantially different actual elevations in nature than model "smoothed" elevations. This can result in emissions being attributed to inappropriate model layers, mis-calculation of bouyant emissions plume rise, application of inappropriate temperatures to temperature-sensitive emissions such as VOCs, etc. It can also result in poorer-than-necessary comparisons against surface-based monitor data.

In meteorology for example, comparisons against surface temperatures attempt to account for these effects using reasonable (inferred) lapse rates. However, no such approach has been generally described for the critical emissions and monitor-observation components of air quality forecast models.

In this paper, we describe a generalized mass-conservative framework that accounts for these effects. We utilize actual point source stack heights and their elevations above mean sea-level to illustrate the scheme. Further, we apply the scheme to surface monitors affected by large variations in elevation at scales sub-grid relative to the forecast model. We compare results of representative forecasts with and without the scheme, and offer some initial insights based on its application.