Tuesday, 9 May 2000: 3:50 PM
	
	
	
	
	
		Surface roughness knowledge is needed for most boundary-layer analysis and modelling, but for 
applications it is seldom available from local measurements.  To estimate roughness visually or from 
maps, Davenport (J.Am.Soc.Civ.Eng. 1960) classified all the then available well-exposed profile data for 
a wide range of terrai.  Wieringa (Bull.Am.Met.Soc. 1980, J. Wind Eng.Ind.Aer. 1992) validated 
Davenport's eight roughness classes for open and moderately rough terrain and for forests, and extended 
its range to smooth terrain and open water.  The classification is widely used, e.g. by WMO.
     Recently, more good experimental roughness data have become available for cities (Grimmond and 
Oke, Bound.Layer Met. 1998, J.Appl.Met. 1999), as well as for very heterogeneous landscapes from 
tethered balloon observations in Britain and in the Sahel.  This made it possible to validate the 
high-roughness classes more fully.  Some shifts in roughness class descriptions prove to be necessary in 
order to account for differences in turbulence generation between bluff buildings and porous vegetation.  
A slightly reformulated Davenport roughness classification is presented, which is now a fully-checked 
tool for estimating effective aerodynamic roughness across the full range of real world terrain.
	
			
			
 - Indicates paper has been withdrawn from meeting
 - Indicates an Award Winner