The aim of this study was to evaluate the extent to which water use efficiency (WUE) at leaf scale can be used to assess WUE at canopy scale, leaf WUE being assumed to be a constant function of vapor pressure deficit and to thus not be dependent upon other environmental factors or varying leaf properties. Leaf WUE and its variability and dependencies were assessed using leaf gas-exchange measurements obtained during two growing seasons, 1999 and 2000, at the Soroe beech forest study site on Zealand in Denmark. It was found that the VPD-normalized leaf WUE, WUEnormleaf, although dependent on incoming PAR below 500 μmol m−2 s−1 is independent, both of the canopy levels and of variations in the environmental parameters. The average WUEnormleaf for PAR above 500 μmol m−2 s−1 was found to be 5.5 μmol CO2 (mmol H2O)−1 hPa and, for the full range, 2.3 μmol CO2 (mmol H2O)−1 hPa. These results showed that WUE can be up-scaled from leaf to canopy on the basis of WUEnormleaf and the PAR distribution within the canopy. The up-scaling conducted was based on this WUEnormleaf – PAR relationship, the light distribution being assessed using the MAESTRA model, parameterized in accordance with measurements obtained for the Soroe forest. The up-scaled WUE was then compared with WUE as estimated from turbulent flux data measured above the forest with the eddy-covariance technique. The modeled daily canopy WUE obtained for daytime fluxes (6:00 AM–6:00 PM) was found to be in agreement with corresponding canopy WUE estimates based on the turbulent fluxes observed and to be dependent on VDP and light intensity alone, its thus being independent of other environmental factors. Accordingly, canopy WUE can be estimated on the basis of the up-scaled WUE relationships, provided incident PAR and VPD within the canopy are known.