Assessing alfalfa production under historical and future climate in eastern Canada: DNDC model development and application

The DeNitrification-DeComposition (DNDC) model was developed to better simulate alfalfa (Medicago sativa L.) production with winterkill effects in eastern Canada. The pre-development DNDC produced "fair" simulations of alfalfa yield and biomass (index of agreement (d) > 0.7, Nash-Sutcliffe efficiency (NSE) > 0) but normalized root mean square error (nRMSE) was > 30% whereas the improved model indicated "good" to "excellent" performance (nRMSE = 3.9-29%; d = 0.93-0.99; NSE = 0.76-0.99). Under future RCP4.5 and 8.5 scenarios, average annual yields increased by 60.3% and 81.8% respectively due to earlier planting/spring regrowth with additional cuttings, increased C assimilation and reduced water stress under higher CO2 concentrations. For locations at Ottawa and Quebec City there could be an increased incidence of winterkill under future climate due to reduced snow cover which leads to colder soil/crown temperatures as well as reduced fall hardening. This study indicated that the use of winter hardy cultivars could mitigate winterkill effects and increase production.