Effect of local and upwind stratification on flow and dispersion inside and above a bi-dimensional street canyon

Abstract

The effects of a stably-stratified boundary layer on flow and dispersion in a bi-dimensional street canyon with unity aspect ratio have been investigated experimentally in a wind tunnel in combination with differential wall heating. Laser-Doppler anemometry together with a fast flame ionisation detector and cold-wire anemometry were employed to sample velocities, concentration, temperatures and fluxes.A single-vortex pattern was observed in the isothermal case, preserved also when leeward wall was heated, but with a considerable increment of the vortex speed. Heating the windward wall, instead, was found to generate a counter-rotating vortex, resulting in the reduction of velocity within the canopy. The stable stratification also contributes reducing the speed, but only in the lower half of the canyon. The largest values of turbulent kinetic energy were observed above the canopy, while inside they were concentrated close to the windward wall, even when the leeward one was heated. An incoming stable stratification produced a significant and generalised turbulence reduction in all the cases. Windward heating was found to produce larger temperature increments within the canopy, while in the leeward case heat was immediately vacated above the canopy. A stable approaching flow reduced both the temperature and the heat fluxes.A passive tracer was released from a point source located at ground level at the centre of the street canyon. The resulting plume cross-section pattern was mostly affected by the windward wall heating, which produced an increment of the pollutant concentration on the windward side by breaking the main vortex circulation. The application of an incoming stable stratification created a generalised increment of pollutant within the canopy, with concentrations twice as large. Turbulent pollutant fluxes were found significant only at roof level and close to the source. On the other hand, in the windward wall-heated case the reduction of the mean flux renders the turbulent component relevant in other locations as well.The present work highlights the importance of boundary layer stratification and local heating, both capable of creating significant modifications in the flow and pollutant fields at microscale range.

Dispersion in an array of buildings in stable and convective atmospheric conditions

“…Each measurement point was sampled for 2.5 minutes, following previous experiments in neutral (Castro et al, 2017) and non-neutral (Marucci and Carpentieri, 2019a) conditions. The standard errors for first and second order statistics was evaluated at each measurement point and deemed satisfactory for high-quality experiments (see, in particular, Marucci and Carpentieri, 2019b).…”

“…Surface aerodynamic (friction velocity u * , roughness length z 0 , displacement height d, BL detpth δ) and thermal (scaling temperature θ * = − w θ 0 /u * , thermal roughness length z 0h , thermal displacement height d h ) were estimated as described in details by Marucci and Carpentieri (2019a) and Marucci and Carpentieri (2019b), by fitting the logarithmic profiles and the vertical shear stress profiles. Other values reported in the table are a reference temperature close to the floor (Θ 0 ), the temperature at the boundary-layer height (Θ δ ), a velocity scale valid on the mixed layer of a CBL, defined as (Kaimal and Finnigan, 1994):…”

Stable and convective boundary-layer flows in an urban array

“…The measurement setup was identical to the one used by Marucci and Carpentieri (2019b), to which we refer the reader for more details. Temperatures and two components of velocity were measured simultaneously using, respectively, a fast-response cold-wire probe (CW) and a laser Doppler anemometer (LDA).…”

“…The LDA target acquisition frequency was set to 100 Hz, while temperatures were sampled at 1000 Hz. Given the irregular nature of the LDA measurements and the different frequencies, a resampling and synchronisation of the three signals was necessary for computing heat fluxes (Marucci and Carpentieri, 2019b). The measurement rig also included a fast flame ionisation detector (FFID) to measure tracer concentrations, even though concentration and mass flux results are not reported here (see, e.g.…”

“…The averaging time for each measurement was set to 2.5 minutes, in line with previous work both in neutral (Castro et al, 2017) and non-neutral (Marucci et al, 2018;Marucci and Carpentieri, 2019b) conditions. The standard error for first and second order statistics was evaluated at each measurement point.…”