Mesoscale transport of energy and matter between the surface and the atmosphere often occurs in the form of non-propagating turbulent organised structures or thermally- induced circulations. Spatially resolving measurements are required to capture such fluxes and, thus far, airborne measurements are the only means to accomplish this. In contrast, tower-based eddy-covariance measurements are conducted at one point and therefore inherently cannot capture the total atmospheric exchange, which is recognised as a major contributor to the energy balance closure problem. As long as there are mean vertical thermal and humidity gradients in the atmospheric boundary layer, with a higher potential temperature and specific humidity in the surface layer than in the outer layer, such organised structures will lead to a systematic underestimation of turbulent energy fluxes from eddy-towers. Firstly, we address the question of how deep such meso-γ scale motions penetrate into the surface layer. We present indications from Doppler- LiDAR, airborne and tower-based measurements, which show that mesoscale motions can indeed be found quite close to the surface, but the mesoscale effect vanishes when measurements are actually conducted within the roughness sublayer and when shear stress is sufficiently large to break up mesoscale contributions into smaller eddies. This is illustrated by observations from Germany and Israel. Secondly, we investigate whether the common practice of adjusting the measured eddy tower fluxes for energy balance closure by conserving the Bowen ratio is supported by experimental evidence. Mesoscale and smallscale turbulent fluxes from four different flight campaigns are presented, which were carried out on board of the Canadian Twin Otter (National Research Council of Canada) and the German Polar 5 (Alfred-Wegener Institute) research aircraft over different landscapes in Canada and Alaska.