A Simple Dispersion Model of Ground Level Ultrafine Particle Concentrations Downwind of a Major Airport

摘要

Introduction: Recent studies have observed a broad footprint of increased ultrafine particle number concentrations up to and possibly beyond 10 km downwind of major airports. Here we compare mobile monitoring measurements downwind of LAX with predictions from a set of relatively simple plume dispersion models. Methods: Mobile monitoring of ultrafine particle number concentrations took place on six days during daytime hours along seven N-S transects ranging from 2 to 9 km downwind of LAX. Background concentrations were first derived using the values obtained by applying a moving boxcar filter to the observations. Cross-wind integrated concentrations (ClCs) of estimated ultrafine particle background concentrations were then computed each sampling day for each transect. ANCOVA was used to compare observations with dispersion model predictions with a random effect of sampling day. The Aermod dispersion model was run separately with hourly average wind data averaged from LAX one minute ASOS data. The source configuration was either a volume-line source to represent the descent path emissions, a surface area source to represent ground level sources at the airport, or a surface trapezoidal area representing cross-wind, ground level transport of the aircraft vortices generated near the ground during descent. The volume-line source geometry was tilted to represent the descent path of the aircraft at a 3 degree inclination. This tilted source was then lowered uniformly by 300 m (descent 300) to represent the effect of transport from descending vortices, resulting in a part of this source being placed at ground level near the airport. Results: Comparison of predicted versus observed ClCs using ANCOVA indicate that either of the ground-level area source models performed better (adjusted R~2 was 0.57 and 0.58 for the trapezoid and the airport area source models, respectively) than the volume line source model (adjusted R~2 = 0.39). In two variable ANCOVA models (line and area source predictions), only the area source term was significant. Conclusions: Two different ground level area source models performed moderately well in predicting background-adjusted, cross-wind integrated concentrations obtained from a mobile monitoring platform at distances up to 9 km downwind of LAX. A descending volume line source model had less skill.