Assessing nitrogen controls on carbon, water and energy exchanges in major plant functional types across North America using a carbon and nitrogen coupled ecosystem model

A carbon and nitrogen coupled dynamic vegetation ecosystem model (CLASS-CTEMN+) was used to assess the effects of nitrogen controls on simulated carbon, water and energy exchanges in a range of vegetation ecosystems. Standardized meteorological forcing data and eddy covariance flux measurements of carbon, water and energy from 32 FLUXNET sites covering eight major plant functional types (PFTs) across North America were used in the analysis. Two versions of the model, a carbon and nitrogen (C-N) coupled version and carbon (C) only version, were employed. Simulated diurnal, daily, seasonal and annual values of gross ecosystem productivity (GEE), ecosystem respiration (Re), net ecosystem productivity (NEP), net radiation (Rn), sensible heat flux (H), latent heat flux (LE) and vegetation biomass and soil carbon stocks were compared with available measured values to evaluate the model's performance in each PFT. The C-N version of the model simulated annual mean NEP for all sites was 211 g C m(-2) yr(-1), compared to 174 g C m(-2) yr(-1) from observation and 253 g C m(-2) yr(-1) by the C version, respectively. Overall, the inclusion of a nitrogen cycle with the carbon cycle in the model resulted in better accuracy scores (e.g., reduced RMSE by 0.31, 0.10 and 0.09 g C m(-2) yr(-1) in the C-N version relative to the C version, for GEP, Re and NEP, respectively) when compared with observations. Simulated Rn, H and LE were usually improved with the inclusion of a nitrogen cycle, but the changes were not always statistically significant at the site level as they were with carbon exchanges. Results indicated the simulated N limitation effects varied among PFTs, but were strongest for boreal forests during the early-growing season. Evaluation of N deposition impacts on GEP, NEP and biomass pools showed considerable variability between and within forest types due to non-linearity of N effects and spatial heterogeneity of C and N cycle interactions. Inclusion of the N cycle in the model will help in its application at regional and global scales to evaluate N availability impacts on the C cycle in terrestrial ecosystems and to determine N cycle feedbacks on Earth's climate. (C) 2015 Elsevier B.V. All rights reserved.