Modelling the impacts of inhibitors and fertilizer placement on maize yield and ammonia, nitrous oxide and nitrate leaching losses in southwestern Ontario, Canada

Application of urease and nitrification inhibitors is an effective strategy for decreasing nitrogen (N) losses without negatively impacting yields, but the long-term impacts of inhibitors and N placement on N dynamics remains unclear. The objectives of this study were to 1) use a 3-year maize experiment in southwestern Ontario to evaluate the Denitrification-DeComposition (DNDC) model fitted with default and enhanced urea hydrolysis mechanisms, 2) determine the impacts of climate variability (30-year) on inhibitor performance on maize yield and N losses under a range of N treatments, and 3) estimate the optimal N application rates for maintaining high yield and minimizing N losses. The experiment treatments included a factorial arrangement of two N sources (urea, urea ammonium nitrate [UAN]), three N placement methods (broadcast urea [B-Urea], broadcast incorporated urea [BInc-Urea], injected UAN [Inj-UAN]), and three inhibitors (no inhibitor, urease inhibitor [UI], urease plus nitrification inhibitor [UI&NI]). Although both the default and improved model provided "fair" to "good" simulations of maize yield, N uptake, soil temperature, soil moisture, soil inorganic N content and nitrous oxide (N2O) emissions across treatments, more accurate simulations were obtained using the improved model. The default DNDC model indicated "poor" simulations of ammonia (NH3) volatilization according to the Nash-Sutcliffe efficiency (NSE <0) and index of agreement (d ≤ 0.38), whereas the improved DNDC produced "good" simulations (NSE = 0.22–0.39, d = 0.65–0.84) because of improved estimates of urea diffusion, hydrolysis and inhibitor efficiency. The 30-year simulation based upon 95% of the maximum yield resulted in annual NH3 losses of 35.4 kg N ha−1 for B-Urea and these losses were reduced by 40.1% with BInc-Urea and by 92.7% with Inj-UAN. When UI was added to either B-Urea or BInc-Urea, NH3 losses were reduced by 61.8–66.4%, whereas when UI&NI was added, losses were reduced by 41.0–49.4%. The Inj-UAN reduced annual N2O losses by 18.8% compared to B-Urea whereas two inhibitor treatments of B-Urea + UI&NI and BInc-Urea + UI&NI reduced N2O losses by 11.5–11.8% compared to B-Urea and BInc-Urea. Simulated annual N losses from Inj-UAN + UI&NI were reduced relative to Inj-UAN + UI because of decreased NO3− leaching. Simulated long-term average percentage of annual total N losses at the optimal N rate were 36.8, 30.8 and 22.0% from B-Urea, BInc-Urea and Inj-UAN, respectively, which reduced to 25.9–28.1, 24.9–25.5 and 21.2–21.9%, when inhibitors were added. Our study suggests that the DNDC model can be used to investigate the impacts of climate variability on inhibitor performance and determine the optimum range of N application rates for different inhibitor treatments.