This paper reports the first effort to include carbon, water, and heat exchange in a Large Eddy Simulation (LES) model for 3D canopy flows with dynamic response of leaf temperature and stomatal aperture. The LES model simulates eddy motion from 3D, transient integration of a filtered form of the Navier–Stokes equations. Carbon exchange between the vegetation and air is predicted in space and time following biophysical considerations, which act to maximize carbon assimilation while minimizing water loss. The vegetation’s stomatal conductance is inferred from these same considerations and used to regulate both transpiration and carbon assimilation rates. Variations in transpiration and radiation distribution propagate to foliage temperature and ultimately heat exchange through a local, transient vegetation energy balance. The wind field is affected by the foliage patterns and by the temperature profile’s control on vertical mixing. These temperature and mixing patterns control the concentration profiles that, in turn, affect water and CO2 exchange processes. By comparing a simulation of horizontally heterogeneous canopy behavior to simulations of several homogeneous canopies with different leaf area index (LAI) values we evaluate the relative importance of local and regional LAI values on the local microenvironment variables and fluxes from the forest canopy. We focus on a pine forest with ample soil moisture as a case study. We demonstrate from these simulations that primitive state variables (e.g. concentrations and velocity) exhibit noticeable non-local controls. However, these features are offset in their effects on land surface fluxes, such that the local fluxes scale well with local LAI values. Furthermore, the resulting relationships between LAI and fluxes are quasi-linear (for the forest morphology studied here) allowing for robust relationships between forest averaged LAI and forest averaged fluxes. The offsetting nature of the non-local effects is described in the context of the dual regulation of stomatal conductance by the rates of carbon assimilation and water loss as opposed to independent regulating effects of the various state variables. Hence, non-local variations in state variables naturally induce offsetting variations in stomatal conductance thereby buffering the water use efficiency of the plant from environmental excursions associated with the turbulent microclimate.