Abstract
Numerical modelling of wind flow over complex dune topography is an ambitious prospect. There is an increasing need to understand wind flow over complex topography for land planning purposes to enable prediction of sediment transport at a particular site. New surveying techniques permit the rapid development of digital terrain models, however a stumbling block is the ability of Computational Fluid Dynamics (CFD) to emulate the wind flow over such a landscape. To overcome these difficulties, it is important to establish the parameters within which such simulations can operate. This paper details an initial two-dimensional numerical model developed in order to test various modelling assumptions against experimental field wind data. Mason Bay, Stewart Island, New Zealand was chosen as an undisturbed but accessible experimental site with a prevalent on-shore wind perpendicular to a simple foredune and a complex down-wind parabolic dune system. A complex topographical two-dimensional model with vegetation represented as a roughness was compared against field data along a transect dissecting a dune system.
This paper establishes that:
Replicating the roughness patterns at the surface is important
The inlet profile should be duplicated with care
Modelling only a portion of the domain can have an effect on the flow patterns due to outflow effects
There is a modelling decision to be made between the complexity of the topography and the sophistication of the turbulence model and degree to which vegetation and sand transportation are modelled.
The long-term aim is to instil confidence in numerical techniques so that such technology can be used for predictive purposes. (C) 2009 Elsevier Ltd. All rights reserved.