Interactions between vegetation, atmospheric turbulence and clouds under a wide range of background wind conditions

The effects of plant responses to cumulus (Cu) cloud shading are studied from free convective to shear-driven boundary-layer conditions. By using a large-eddy simulation (LES) coupled to a plant physiology embedded land-surface submodel, we study the vegetation-cloud feedbacks for a wide range (44) of atmospheric and plant stomatal conditions. The stomatal relaxation time is prescribed as an instantaneous, symmetrical (10, 15 and 20min) and asymmetrical (5min closing, 10min opening) response, and the background wind ranges from 0 to 20ms-1. We show that in free convective, non-shading (i.e. transparent) cloud conditions the near-surface updraft region is marked by an enhanced CO2 assimilation rate (A n; 7%) and increased latent (LE; 9%) and sensible heat (H; 19%) fluxes. When we introduce Cu shading, we find an enhancement in plant transpiration and CO2 assimilation rates under optically thin clouds due to an increase in diffuse radiation. However, these effects vanish when a background wind is present and the Cu are advected. Optically thick clouds reduce the assimilation rate and surface fluxes under all simulated wind conditions.With increasing background wind, the shaded surface area is enlarged due to Cu tilting. The consequent decrease in surface fluxes by a reduction in incoming radiation, is partly offset due to an enhancement in the surface exchange and turbulent mixing as a result of stronger wind speeds. Different and non-linear processes control the H and LE response to shading. H is mainly radiation driven, whereas plant responses dampen the shading effects on LE. As a result, the regional averaged (48km2) reduction in H and LE are found to be 18% and 5%, respectively, compared to non-shading cloud conditions. Surprisingly, a nearly uniform regional net radiation reduction of 11% is found, with only a deviation between all 35 Cu shading cases of 0.5% (i.e. 1.2Wm-2) at the moment of maximum cloud cover. By comparing four representative simulations that are equal in net available energy, but differ in interactive and prescribed surface energy fluxes, we find a relative reduction in cloud cover between 5 and 10% during the maximum cloud cover period when the dynamic surface heterogeneity is neglected. We conclude that the local and spatial dynamic surface heterogeneity influences Cu development, while the Cu-vegetation coupling becomes progressively weaker with increasing stomatal relaxation time and background wind