Assessing the Predictive Accuracy and Uncertainty of the CAN-DNDC Model to Simulate Water Use Efficiency in Alfalfa Crops.

English Abstract:

Improving the sustainability of water use in agricultural crop production requires an understanding of the existing water use efficiency (WUE) of individual crops. WUE in agricultural ecosystems may be expressed as net ecosystem exchange (NEE) / evapotranspiration (ET). When accurately calibrated and validated, the Canadian agricultural ecosystem model, Denitrification-Decomposition (CAN-DNDC) provides an ideal tool for predicting WUE by providing simulations of water and carbon cycle processes, as well as the nitrogen processes, of specific crops at the site or regional level. In support of a project addressing water use in the dairy production sector, Agriculture Canada is developing crop-specific growth curves and improved algorithms of nitrogen/water uptake for Canadian alfalfa. Our objective is to assess the predictive accuracy and uncertainty of the alfalfa model through a comparison of simulated NEE and ET data with in situ eddy covariance (EC) flux measurements. Eddy covariance measurements were obtained during the 2014 growing season from two Eastern Ontario alfalfa fields using a closed-path CO₂/H₂O gas analyzer (CPEC200, Campbell Scientific, Logan UT) and an integrated CO₂/H₂O open-path gas analyzer and 3-D sonic anemometer (IRGASON, Campbell Scientific, Logan UT). To assess the predictive accuracy of the model, a Matlab statistical program is used to estimate the structural uncertainty and a Monte Carlo analysis is performed to assess uncertainties arising from input variables.