Energy and water exchanges modulated by soil–plant nitrogen cycling in a temperate Pacific Northwest conifer forest

Nitrogen controls on the seasonal and inter-annual variations in evapotranspiration in a temperate Pacific Northwest Douglas-fir forest in British Columbia, Canada, were studied using a coupled carbon and nitrogen (C&N) model developed by incorporating plant–soil nitrogen cycling algorithms in the Carbon-Canadian Land Surface Scheme (default C-CLASS). In the coupled C&N-CLASS, the maximum carboxylation rate of Rubisco (Vcmax) is determined nonlinearly from the modeled leaf Rubisco-nitrogen, rather than being prescribed. Hence, variations in canopy assimilation, stomatal conductance and evapotranspiration are sensitive to leaf nitrogen status through Rubisco activity. Model output from both default and coupled versions was compared with eddy covariance water vapor flux measurements made over a 5-year period (1998–2002) in this forest. Incorporation of leaf Rubisco-N modulated photosynthesis algorithms, which interact with canopy conductance and other soil–plant N cycling processes, improved canopy transpiration simulations (about 10% higher than from the default model) and thus latent heat flux (half-hourly and monthly), particularly during the main growing seasons. Values of soil moisture simulated using the coupled model were in better agreement with observations than those using the default model. Compared to the observed evapotranspiration values of 432, 435, 415, 382 and 388mm for 1998–2002, respectively, coupled C&N model simulated annual values were 380, 398, 382, 403 and 386mm, while the default model's were 357, 375, 363, 386 and 374mm for respective years. The coupled C&N land surface model framework will be a useful tool for evaluating impacts of nitrogen cycle on evapotranspiration and energy fluxes from terrestrial ecosystems and its feedbacks on Earth's climate system.