A cautionary note on the use of Monin-Obukhov similarity theory in very high-resolution large-eddy simulations