Roughness lengths for momentum and heat derived from outdoor urban scale models

Urban climate experimental results from the Comprehensive Outdoor Scale Model (COSMO) were used to estimate roughness lengths for momentum and heat. Two different physical scale models were used to investigate the scale dependence of the roughness lengths; the large scale model included an aligned array of 1.5-m concrete cubes, and the small scale model had a geometrically similar array of 0.15-m concrete cubes. Only turbulent data from the unstable boundary layers were considered. The roughness length for momentum relative to the obstacle height was dependent on wind direction, but the scale dependence was not evident. Estimated values agreed well with a conventional morphometric relationship. The logarithm of the roughness length for heat relative to the obstacle height depended on the scale but was insensitive to wind direction. COSMO data were used successfully to regress a theoretical relationship between kappa B-1, the logarithmic ratio of roughness length for momentum to heat, and Re*, the roughness Reynolds number. Values of kappa B-1 associated with Re* for three different urban sites from previous field experiments were intercompared. A surprising finding was that, even though surface geometry differed from site to site, the regressed function agreed with data from the three urban sites as well as with the COSMO data. Field data showed that kappa B-1 values decreased as the areal fraction of vegetation increased. The observed dependency of the bulk transfer coefficient on atmospheric stability in the COSMO data could be reproduced using the regressed function of Re* and kappa B-1, together with a Monin-Obukhov similarity framework.