A Model for Local to Global to Local Scale Atmospheric Dispersion

Dispersion is inherently a multiscale phenomenon. Most sources are constrained in their spatial and temporal extent, but the pollutant may travel long distances before it re-interacts with the local terrain to deposit on the surface. This range of scales makes the simulation of the processes involved difficult. The usual method of surmounting this difficulty is to use grid nesting, though this requires prior knowledge as to where to put the nests. This paper discusses an approach that has been used for a number of years now – the application of adaptive unstructured grids – that has been extended to support local to global to local scale dispersion. The Operational Multiscale Environment model with Grid Adaptivity (OMEGA) consists of an atmospheric forecast and simulation system designed around an adaptive unstructured grid with an embedded Lagrangian Atmospheric Dispersion Model (ADM). Originally designed for regional forecasts and simulations, OMEGA / ADM now supports full global problems with both static and dynamic adaptive resolution providing appropriate resolution to model the local processes.