We investigated interannual variability of canopy transpiration per unit ground area (E (C)) and per unit leaf area (E (L)) across seven tree species in northern Wisconsin over two years. These species have previously been shown to be sufficient to upscale stand-level transpiration to the landscape level during one growing season. Our objective was to test whether a simple plant hydraulic model could capture interannual variation in transpiration. Three species, wetland balsam fir (Abies balsamea (L.) Mill), basswood (Tilia Americana L.) and speckled alder (Alnus rugosa (DuRoi) Spreng), had no change in E (C) or E (L) between 2000 and 2001. Red pine (Pinus resinosa Ait) had a 57 and 19% increase in E (C) and E (L), respectively, and sugar maple (Acer saccharum Marsh) had an 83 and 41% increase in E (C) and E (L), respectively, from 2000 to 2001. Quaking aspen (Populus tremuloides Michx) had a 50 and 21% decrease in E (C) and E (L), respectively, from 2000 to 2001 in response to complete defoliation by forest tent caterpillar (Malascoma distria Hüber) and subsequent lower total leaf area index of the reflushed foliage. White cedar (Thuja occidentalis L.) had a 20% decrease in both E (C) and E (L) caused by lowered surface water in wetlands in 2001 because of lower precipitation and wetland flow management. Upland A. balsamea increased E (L) and E (C) by 55 and 53%, respectively, as a result of release from light competition of the defoliated, overstory P. tremuloides. We hypothesized that regardless of different drivers of interannual variability in E (C) and E (L), minimum leaf water potential would be regulated at the same value. Minimum midday water potentials were consistent over the two years within each of the seven species despite large changes in transpiration between years. This regulation was independently verified by the exponential saturation between daily E (C) and vapor pressure deficit (D) and the tradeoff between a reference canopy stomatal conductance (G (S)) and the sensitivity of G (S) to D, indicating that trees with high G (S) must decrease G (S) in response to atmospheric drought faster than trees with low G (S). Our results show that models of forest canopy transpiration can be simplified by incorporating G (S) regulation of minimum leaf water potential for isohydric species.