Surface Temperature Heterogeneity Effects on Dispersion in the Stable Boundary Layer

Previous studies investigated the effects of surface-temperature heterogeneity (HET) in the stable boundary layer (SBL) on the turbulence structure and transport properties using large eddy simulations (LESs) (Stoll and Porte-Agel, 2009; Mironov and Sullivan, 2010). The effects were demonstrated by comparison with properties for a surface-temperature homogeneous (HOM) SBL. The homogeneous case was studied earlier for the GABLS-1 weakly stable boundary layer as part of an LES-model intercomparison (Beare et al., 2006). The LESs for the two SBLs were run for the same domain size [(400 m)^3], number of grid points (200), geostrophic wind speed (8 m/s), and initial stratification. In the HET SBL, the temperature heterogeneity was simulated using spanwise homogeneous surface strips having a sinusoidal temperature variation with distance in the mean-wind direction and the same mean surface temperature as in the HOM case. The HET SBL was more turbulent and well-mixed with respect to temperature than the HOM SBL, and an outstanding feature was the temperature variance, which peaked at the surface and SBL top with maximum values far greater than in the HOM case.

In this paper, we investigate the dispersion characteristics of the HET and HOM SBLs using the LES fields to drive a Lagrangian particle dispersion model. Simulations were conducted for several source heights but the main focus was on a surface release in order to explain the large surface temperature variance found by Mironov and Sullivan. Calculations were made for an array of sources at the leading edge of the hot and cold strips and in the middle of each strip to obtain "phase-averaged" concentrations; the strip length was 200 m. In addition, the mean concentration obtained for a uniform distribution of surface sources was in excellent agreement with the mean from the HOM SBL. However, the phase-averaged concentrations showed an oscillatory distance dependence (about the mean) due to the horizontally-varying surface temperature, which generated gravity waves. Results for the "leading-edge" sources of the hot and cold strips were in good agreement with one another, but the wavelength of the concentration distribution was about half that of the surface temperature field. We believe that this was caused by the wave-induced heat flux which has a wavelength one-half that of the surface temperature, due to the wave-induced temperature and vertical velocity being 90 degrees out of phase. For sources at the midpoint of each strip, the concentration was 180 degrees out of phase with the leading-edge sources. The net result for a uniform distribution of surface sources is that the phase-averaged concentration variances would be superposed and produce a total variance greater than that due to the SBL turbulence alone. We believe that this explains the large temperature variances found at the surface by Mironov and Sullivan.