This document presents a summary of ADMS model results compared against three well known field data sets: Prairie Grass. Kincaid and Indianapolis. Results are also presented for two U.S. models: ISCST and AERMOD.
ISC ST is the United States Environmental Protection Agency model. It is a so-called 'old generation' model, similar to the UK R-91 model, based on the 1970's description of boundary layer physics. Such models use several discrete categories to describe the stability of the atmospheric boundary layer. Gaussian distributions are assumed for the concentration in the crosswind vertical and horizontal directions in all stabilities.
ADMS has been developed in the UK and is widely used internationally by industry, consultants and regulatory bodies. AERMOD is the U.S. regulatory model. ADMS and AERMOD are both described as 'new generation1 models. These models describe the state of the atmospheric boundary using two parameters: boundary layer depth and Monin-Obukhov length. The vertical concentration distribution is Gaussian in neutral and stable atmospheres but is a skewed in Gaussian in convective conditions. As with the 'old generation1 models a Gaussian distribution is assumed in the crosswind horizontal direction for all stabilities.
ADMS has been extensively validated during its development against field data sets and wind tunnel data sets. Studies have covered a range of meteorological conditions (Carruthers et ah, 1994 ; Carruthers et ah, 1996 ; Carruthers et ah, 1998 ). including a comparison against LIDAR data undertaken for the UK Environment Agency with the emphasis on convective conditions. Advanced model features have also been tested, particularly the buildings module (Carruthers. McKeown et ah, 1999 ).
The three data sets used here are drawn from those that are openly available and have been generally accepted as containing enough measurements and of sufficient quality for meaningful validation. A series of workshops have been held over the last few years on 'Harmonisation of Atmospheric Dispersion Modelling for Regulatory Purposes'. At these 'harmonisation workshops' a Model Validation Kit (MVK) has been developed to evaluate models. It includes the Kincaid and Indianapolis data sets and the BOOT statistical package (Olesen, 1994 ; Olesen, 1998 ; Hanna etal, 1997 ). The BOOT statistical package is described further in Section 2.1. The Prairie Grass data set is derived from the files that have been used for the ASTM model evaluation (D6589) [I], with additional input data from United States Environmental Protection Agency Website .