Intercomparison of FLEXPART and CALPUFF dispersion models. An application over a small tropical island

Abstract

A typical practice in air quality modeling assessment is the intercomparison between different dispersion models results and air quality measurements at different atmospheric conditions. In this study, a comparison between the results of two Lagrangian dispersion models, the Lagrangian Particle Dispersion Model FLEXPART and the Lagrangian Puff Model CALPUFF (regulatory model), coupled to the same meteorological fields produced by the Weather Research and Forecasting (WRF) model, was done. As a case study, atmospheric dispersion of anthropogenic nitrogen oxides (NO_x) emissions (considered as a passive tracer) was considered, during a typical case of severe pollution over the densely populated area of the Guadeloupe archipelago (West French Indies), including complex terrain and, of course, coastal influence. Even though Lagrangian models usually provide better results of plume dispersion under strong winds, in this case study weak trade winds are dominant, in order to check both models under non-ideal atmospheric conditions. As a result, compared to NO_x ground level concentration (glc) observations, FLEXPART shows better agreement than CALPUFF. However, as a regulatory model, CALPUFF overestimates both glc observations and FLEXPART maximum NO_x glc results, with higher values when a higher horizontal resolution is applied. Also, differences between models results arise in the spatial distribution of NO_x over a 1 × 1 km² horizontal resolution grid domain, showing quite homogenous isopleths with smooth contours for CALPUFF vs. fragmented isopleths with irregular contours for FLEXPART.