Agreed, and another contributing factor to underestimating conduction loss here may be the radiation loss from the window to the clear sky. That is, clear (as opposed to overcast) skies have a high absorptance to emittance ratio and thus your windows will experience net radiation loss to the sky; as a function of your emittance-weighted view factor to your surroundings.
@Tiantian, I think the easiest way to account for these surface convection, and radiation exchanges without tracking all energy interactions through your glass, is simply to substitute the surface temperatures of your glazing assembly in your Fourier’s Law calculation, in place of the interior room and exterior wetbulb temperatures.
I attempted a similar experiment a while ago here: Eplus Conduction Rate and Uvalue of the wall check, albeit for opaque constructions during night time conditions. Since glazing should have negligible thermal mass, this same strategy should work fine for glass during the daytime.
S