Evaluation of a Gaussian and a Lagrangian model against a roadside data set, with emphasis on low wind speed conditions

The evaluation of the high percentiles of concentration distributions is re quired bq most national air quality guidelines, as well as the EU directive s. However, it is problematic to compute such high percentiles in stable, l ow wind speed or calm conditions. This study utilizes the results of a prev ious measurement campaign near a major road at Elimaki in southern Finland in 1995, a campaign specifically designed for model evaluation purposes. In this study, numerical simulations were performed with a Gaussian finite li ne source dispersion model CAR-FMI and a Lagrangian dispersion model GRAL, and model predictions wen compared with the field measurements. In comparis on with corresponding results presented previously in the literature, the a greement of measured and predicted data sets was good fur both models consi dered, as measured using various statistical parameters. For instance, cons idering all NO, data (N = 557), the so-called index of agreement values var ied from 0.76 to 0.87 and fr om 0.81 to 1.00 for the CAR-FMI and GRAL model s, respectively. The CAR-FMI model tends to slightly overestimate the NO, c oncentrations (fractional bias FB = +14%), while the GRAL model has a tende ncy to underestimate NOx concentrations (FB = -16%). The GRAL model provide s special treatment to account for enhanced horizontal dispersion in low wi nd speed conditions; while such adjustments have not been inducted in the C AR-FMI model. This type of Lagrangian model therefore predicts lower concen trations, in conditions of low wind speeds and stable stratification, in co mparison with a standard Lagrangian model. In low wind speed conditions the meandering of the flow can be quite significant, leading to enhanced horiz ontal dispersion. We also analyzed the difference between the model predict ions and measured data in terms of the wind speed and direction. The perfor mance of the CAR-FMI model deteriorated as the wind direction approached a direction parallel to the road, and for the lowest wind speeds. However, th e performance of the GRAL model varied less with wind speed and direction; the model simulated better the cases of low wind speed and those with the w ind nearly parallel to the road. (C) 2001 Elsevier Science Ltd. AII rights reserved.