Hector Machado and Andrew Martin
Abstract – Meteorological and dispersion models use input data that samples reality. To simulate airflow and dispersion, the meteorological models use gridded terrain heights to fit wind vectors (and other parameters) to grid cells. However, due to limitations on the resolution of grid size, terrain heights will often vary substantially within each grid, so the model is producing wind fields based on smoothed terrain. In addition, the dispersion model will also use terrain heights as one determinant of plume centreline height.
Source and receptor heights can be defined as ground height plus height above ground. By default, some model interfaces will use the modelled terrain grid to assign ground height. Some modellers prefer to correct these for the actual terrain heights, so that the model accurately represents the heights of sources and receptors relative to each other. However, the modelled airflow and plume centreline, or puff centre, do not follow the actual terrain: they follow the modelled terrain. Hence, in some circumstances, it may be more appropriate to model heights accurately relative to ground level. For screening, a range of heights can be modelled to provide a conservative worst case.
This paper uses different hypothetical scenarios and the CALPUFF, AERMOD and GRAL models to compare how the models react to steep isolated hills and steep ridges. High resolution, vertical cross-section contours are used to illustrate predictions. Recommendations are provided on the importance of grid resolution and on how to treat source and receptor heights.
Presented at the The Clean Air Society of Australia and New Zealand (CASANZ) 28th International Clean Air and Environment Conference and the Asia-Pacific Odour Conference in Cairns, Queensland, on 31 August to 2 September 2026.
Read the full paper here.