Assessing the effects of manure application rate and timing on nitrous oxide emissions from managed grasslands under contrasting climate in Canada

It is uncertain whether process-based models are currently capable of simulating the complex soil, plant, climate, manure management interactions that influence soil nitrous oxide (N2O) emissions from perennial cropping systems. The objectives of this study were (1) to calibrate and evaluate the DeNitrification DeComposition (DNDC) model using multi-year datasets of measured nitrous oxide (N2O) fluxes, soil moisture, soil inorganic nitrogen, biomass and soil temperature from managed grasslands applied with manure slurry in contrasting climates of Canada, and (2) to simulate the impact of different manure management practices on N2O emissions including slurry application i) rates (for both single vs. split); and ii) timing (e.g., early vs. late spring). DNDC showed fair to excellent performance in simulating biomass (4.7% <= normalized root mean square error (NRMSE) <= 29.8%; -9.5% <= normalized average relative error (NARE) <= 16.1%) and good performance in simulating soil temperature (13.2% <= NRMSE < 18.1%; -0. 7% <= NARE <= 10.8%) across all treatments and sites. However, the model only showed acceptable performances in estimating soil water and inorganic N contents which was partially attributed to the limitation of a cascade water sub-model and inaccuracies in simulating root development/uptake. Although, the DNDC model only demonstrated fair performance in simulating daily N2O fluxes, it generally captured the impact of the timing and rate of slurry application and soil texture (loam vs. sandy loam) on total N2O emissions. The DNDC model simulated N2O emissions from spring better than split manure application (fall and spring) at the Manitoba site partially due to the overestimation of available substrates for microbial denitrification from fall application during the wet spring periods. Although DNDC performed adequately for simulating most of the manure management impacts considered in this study we recommend improvements in the simulation of soil freeze-thaw cycles, manure decomposition dynamics, soil water storage, rainfall canopy interception, and microbial denitrification and nitrification activities in grasslands. Crown Copyright (C) 2019 Published by Elsevier B.V.