Surface Thermal Heterogeneities and the Atmospheric Boundary Layer: The Relevance of Dispersive Fluxes

Abstract

While the increasing availability of computational power is enabling finer grid resolutions in numerical-weather-prediction models, representing land–atmosphere exchange processes remains challenging. This partially results from the fact that land-surface heterogeneity exists at all spatial scales, and its variability does not necessarily ‘average’ out with decreasing size. The work presented here uses large-eddy simulations and the concept of dispersive fluxes to quantify the effects of a surface that is thermally inhomogeneous (with scales that are approximately 10% of the height of the atmospheric boundary layer), but uniformly rough. These near-canonical cases describe inhomogeneous scalar transport over a broad range of unstable atmospheric flows. Results illustrate the existence of a regime where the mean flow is mostly driven by the surface thermal heterogeneities. In this regime, the contribution of the dispersive fluxes can account for more than 40% of the total sensible heat flux at 100 m above the ground and about 5–10% near the surface. This result is independent of the spatial distribution of the thermal heterogeneities and weakly dependent on the averaging time used to define the dispersive fluxes. Additionally, an alternative regime exists where the effects of the surface thermal heterogeneities are quickly blended and the dispersive fluxes match those obtained over an equivalent homogeneous surface. Results further illustrate the existence of a new cospectral scaling for the dispersive sensible heat fluxes that differs from the traditional turbulence cospectral scaling. We believe that these results might elucidate pathways for developing new parametrizations for the non-canonical atmospheric surface layer.