Defining the roughness sublayer and its turbulence statistics.

The roughness sublayer in a turbulent open-channel flow over a very rough wall is investigated experimentally both within the canopy and above using particle image velocimetry by gaining complete optical access with new methodologies without disturbing the flow. This enabled reliable estimates of the double-averaged mean and turbulence profiles to be obtained by minimizing and quantifying the usual errors introduced by limited temporal and spatial sampling. It is shown, for example, that poor spatial sampling can lead to erroneous vertical profiles in the roughness sublayer. Then, in order to better define and determine the roughness sublayer height, a methodology based on the measured spatial dispersion is proposed which takes into account temporal sampling errors. The results reveal values well below the usual more ad hoc estimates for all statistics. Finally, the double-averaged mean and turbulence statistics in the roughness sublayer are discussed. [ABSTRACT FROM AUTHOR]