Variability of Effective Roughness Height in Model Urban Canopies

We present the results of roughness and displacement height based on PIV velocity measurements in a water tunnel experiment of flow over idealized models of urban canopies. Experiments were conducted with large roughness elements of regular arrays of buildings of uniform height with aspect ratios H/w_b=1 and H/w_b=3 where H and w_b are the building height and width respectively; values of the plan area density and frontal area density for the above canopies were λ_p= λ_f =0.19 and λ_p= λ_f =0.19 or the H/w_b=1 and H/w_b=3 canopies respectively. Laterally averaged values of non-dimensional displacement height d/H increase monotonically from 0 to 1 with plan area density λ_p of the urban canopy. In contrast, laterally averaged non-dimensional roughness height z₀/H increases to a maximum value as λ_f approaches a value of 0.2 and then decreases to zero. Values of d/H and z₀/H vary spatially with large differences in values behind building wakes, and in the canyons between buildings. We present data for non-dimensional effective roughness heights (z₀+d)/H as a function of aspect ratio H/w_b. This also reveals three categories of values: street canyon, building wake, and laterally averaged values. The measurements taken at the centerline of canyons form a lower bound on the effective roughness height whereas measurements behind building wakes form the upper bound on effective roughness height values. Laterally averaged values of non-dimensional friction velocity u_*/U_H) varied inversely with the aspect ratio (U_H is the mean velocity at the building height).