For a surface covered by vegetation, the scalar roughness for sensible heat zoh (or equivalently, any bulk heat transfer coefficient or surface resistance) depends directly on the way in which the effective or bulk surface temperature &thgr;s is measured. One undesirable feature of any such parameter like zoh is that when &thgr;s is measured radiometrically under conditions of solar heating of the surface, it tends to depend strongly on solar elevation. An analysis of experimental surface heat flux and related data herein confirms that this dependency can be weakened by making the radiometric surface temperature measurements obliquely, that is, by properly weighting the temperature of the upper layers of the canopy; these are the more direct heat exchangers with the turbulent air than the lower layers. It is also shown, however, that such oblique radiometric measurements produce lower values of &thgr;s, which in turn, require larger values of zoh, in the bulk heat transfer equation. Unfortunately, lower &thgr;s and larger zoh result in larger relative error in heat flux calculations. Therefore, under certain conditions, it may be preferable to design a compromise strategy and to accept some dependency of zoh on solar elevation, in order to maintain a sufficiently low error level in the calculated heat flux. In any event, this error level can be reduced considerably by making use of air temperature measurements at higher levels above the ground. ¿ American Geophysical Union 1996