The turbulent motion of atmosphere adjoined to the earth’s surface is caused by the surface friction and buoyancy force in a short time scale which is of the order of few seconds to minutes. This interactive layer is atmospheric boundary layer with several detailed features viz., roughness layer, surface layer for small eddies, mixed layer for large eddies, which may extend vertically more than a kilometer under deep convective condition. The surface layer is primarily characterised through several scaling parameters, viz., roughness length, momentum flux and sensible heat flux as triggering force of turbulence and its maintenance through eddy formation and dissipation. Air dispersion and hence the distribution of atmospheric pollutants from various sources follow flow path and its turbulent characteristics, which is dependent on geometry of the terrain specially for complex terrain, meteorological conditions and source term inventories. In the present study, an virtual environment is designed in a computational fluid dynamics based solver for limited area/local scale air quality modelling by incorporating digital elevation model of intermediate valley zone of Sahyadri Range of Western Ghats of India as site. Finite volume method and mass flux transfer approach is followed by the solver to estimate 3D atmospheric or ground level concentration of atmospheric non-reactive pollutants for unit discharge per second from a hypothetical 100 m tall stack for 72 hours. Flow field is computed for various realistic cases for which initial and boundary conditions are incorporated from larger scale model, viz., Mesoscale Model (MM5) developed by the National Centre for Atmospheric Research, USA and also from measured local data taken from distributed micro-meteorological observatories. Computed flow field is validated with independent observation not utilised in computation giving a validation to the estimated contaminant concentration so. The uncertainty in diffusion is also studied when Lagrangian particle movement is driven fully by flow field and its turbulent phenomena in comparison to random walk model.

This content is only available via PDF.
You do not currently have access to this content.