Accountability of wind variability in AERMOD for computing concentrations in low wind conditions

Commonly large shifts of wind direction in low-wind conditions are poorly understood and are not sufficiently captured by air-quality dispersion model AERMOD. In the low-wind conditions, the observed concentration distribution is multi-peaked and non-Gaussian due to the large variability in the wind direction. To account the variability in the wind direction, a segmented approach is used by assuming that a shorter time period (2 min) mean wind direction estimate the plume more closely than the hourly mean wind. For illustration, concentration measurements from the low wind diffusion experiment conducted at Idaho are utilized. The qualitative performance of AERMOD with all the three options (FASTALL, LOW-WIND1, and LOW-WIND3) using segmented approach is reasonably good in terms of explaining the key characteristics such as multiple peaks and large plume spread of the observed concentration and is relatively better than the hourly approach (using hourly mean wind). The statistical measures for all the three options of AERMOD using segmented approach are found good in agreement with the observations, and a quantitative analysis based on ANOVA (analysis of variance) shows that the results from all the three options of AERMOD using segmented approach are found to be comparable at 5% significance level.