Linking near surface turbulence under stable conditions to regional flow stability, and geometric and surface cover properties of the landscape

Averaged surface heat and momentum fluxes obtained from a network of surface flux stations in a heterogeneous landscape are found as function of a regional Richardson number (Ri_br). The results indicate that the need for extra mixing above a critical bulk Richardson number (Ri_cr = 1/4) in numerical weather prediction models (NWP) to avoid unrealistic cooling at surface is a consequence of spatial averaging. Network-averaged flux persists when the regional flow stability is well above critical (Ri_br >> Ri_cr). We found that there are preferential places for the occurrence of strong turbulence under same stability conditions of the mesoscale flow. Further, we identify two most prominent local landscape characteristics near observation stations that signal the likely occurrence of intermittent turbulence, terrain concavity and site sheltering. Site sheltering was determined using Fujita's ‘transmission factor' and surface concavity by determining the Laplacian ( ∂²/∂x²+∂²/∂y² ) of a quadratic surface ( s = s(x,y) ) fitted to the local topography. Each site characteristic exerts systematic influences on nocturnal heat and momentum fluxes, with sheltering being more influential under windy conditions ( > 5m s^-1 ) and concavity under weak windy conditions.