LES study of the energy imbalance problem with Eddy covariance fluxes

The spatial representativeness of heat fluxes on the basis of single-tower measurements, and the mechanism of the so-called energy imbalance problem, are investigated through numerical experiments using large-eddy simulation (LES). LES experiments are done for the daytime atmospheric boundary layer heated over a flat surface, as a best-case scenario completely free of sensor errors and the uncertainties of field conditions. Imbalance is defined as the deviation of the 'turbulent' heat flux at a grid point from the horizontally averaged 'total' heat flux. Both the theoretical and numerical results of the present study suggest the limitation of single-tower measurements and the necessity of horizontally-distributed observation networks. The temporally averaged 'turbulent' flux based on a point measurement systematically underestimates the 'total' flux (negative imbalance). This is attributed to local advection effects caused by the existence of turbulent organized structures (TOS), whose time scale is much longer than that of thermal plumes. The temporal and spatial change of TOS patterns causes low-frequency trends in the velocity and temperature data resulting in large scatter of the flux estimates. The influences of geostrophic wind speed, averaging time, observation height, computational domain size and resolution on tower-measured fluxes are also discussed. Finally, it is suggested that a weak inhomogenity in surface heating may reduce the negative bias of flux estimates.