Keywords: atmospheric dispersion models, complex terrain, mesoscale modelling, model evaluation, numerical simulations, tracer experiments
The "Tracer Potential" method of evaluating atmospheric models: main algorithms
In this paper, we describe the basic concepts and main algorithms of a method which uses tracer measurements to compare and evaluate wind fields as predicted by different atmospheric models or obtained from measurements. In particular, we focus on details of the trajectory calculation and the estimation of the relative and absolute successes of the wind fields. A cost function "Tracer Potential" is defined which takes into account the separation between each segment of the simulated or measurement-derived trajectory and each of the tracer receptors, as well as the magnitude of the tracer concentration. The objective is to identify and quantitatively describe the wind fields which are best able to reproduce the main transport of tracers. The method has been applied to data from a field programme that included tracer measurements. The programme was conducted in the Colorado River valley area of the southwest USA in 1992, and the results have been reported elsewhere.