We have developed a method of using eddy covariance latent heat flux data to scale evapotranspiration (ET) from the leaf-level at the half-hourly timescale to the canopy-level at the monthly time-scale to develop a canopy conductance algorithm. This algorithm is incorporated into a global biogeochemistry model (Terrestrial Ecosystems Model, TEM) to determine how ET influences carbon dynamics and ozone uptake in temperate forests across the conterminous U.S. Relative to older versions of TEM, the new model reduces our estimates of mid-summer ET, thereby increasing water use efficiency, and increasing Net Primary Production (NPP), especially in regions predicted to become water stressed. During the 1995-1999 period for the temperate forests of the conterminous U.S. (representing 3,146,199 km²), ET in the new version is 838 mmyr^-1 less than the older version and the NPP increases from 1850 to 2133 TgCm^-2yr^-1. During this same period, the ozone uptake resulting from the new ET estimates contributes to a 6.3% decrease in NPP. Eddy covariance water flux data is useful in constraining carbon dynamics through limitations to water use efficiency, while the resulting canopy conductance determines the amount of ozone or other pollutant damage.