To simulate the airflow through a wind farm across a wide range of atmospheric conditions, microscale models (e.g., large-eddy simulation, LES, models) have to be coupled with mesoscale models, because microscale models lack the atmospheric physical processes to represent time-varying local forcing. Here we couple mesoscale model outputs to a LES solver by applying mesoscale momentum- and temperature-budget components from the Weather Research and Forecasting model to the governing equations of the Simulator fOr Wind Farm Applications model. We test whether averaging the budget components affects the LES results with regard to quantities of interest to wind energy. Our study focuses on flat terrain during a quiescent diurnal cycle. The simulation results are compared with observations from a 200-m tall meteorological tower and a wind-profiling radar, by analyzing time series, profiles, rotor-averaged quantities, and spectra. While results show that averaging reduces the spatio-temporal variability of the mesoscale momentum-budget components, when coupled with the LES model, the mesoscale bias (in comparison with observations of wind speed and direction, and potential temperature) is not reduced. In contrast, the LES technique can correct for shear and veer. In both cases, however, averaging the budget components shows no significant impact on the mean flow quantities in the microscale and is not necessary when coupling mesocale budget components to the LES model.