On the segregation of chemical species in a clear boundary layer over heterogeneous land surfaces

Using a Large-Eddy Simulation model, we have systematically studied the inability of boundary layer turbulence to efficiently mix reactive species. This creates regions where the species are accumulated in a correlated or anti-correlated way, thereby modifying the mean reactivity. We quantify this modification by the intensity of segregation, I-S, and analyse the driving mechanisms: heterogeneity of the surface moisture and heat fluxes, various background wind patterns and non-uniform isoprene emissions. The heterogeneous surface conditions are characterized by cool and wet forested patches with high isoprene emissions, alternated with warm and dry patches that represents pasture with relatively low isoprene emissions. For typical conditions in the Amazon rain forest, applying homogeneous surface forcings and in the absence of free tropospheric NOx, the isoprene-OH reaction rate is altered by less than 10%. This is substantially smaller than the previously assumed I-S of 50% in recent large-scale model analyses of tropical rain forest chemistry. Spatial heterogeneous surface emissions enhance the segregation of species, leading to alterations of the chemical reaction rates up to 20%. The intensities of segregation are enhanced when the background wind direction is parallel to the borders between the patches and reduced in the case of a perpendicular wind direction. The effects of segregation on trace gas concentrations vary per species. For the highly reactive OH, the differences in concentration averaged over the boundary layer are less than 2% compared to homogeneous surface conditions, while the isoprene concentration is increased by as much as 12% due to the reduced chemical reaction rates. These processes take place at the sub-grid scale of chemistry transport models and therefore need to be parameterized.