Jump to Main Content
A multi model evaluation of long-term effects of crop management and cropping systems on nitrogen dynamics in the Canadian semi-arid prairie
- Dutta, B., Grant, B.B., Campbell, C.A., Lemke, R.L., Desjardins, R.L., Smith, W.N.
- Agricultural systems 2017 v.151 pp. 136-147
- Triticum aestivum, aboveground biomass, cropping systems, data collection, environmental impact, environmental quality, gas emissions, grain yield, leaching, long term effects, model validation, nitrogen, nitrogen content, nitrogen cycle, nutrient use efficiency, regression analysis, silt loam soils, simulation models, spring wheat
- Process-based biogeochemical models such as the DeNitrification-DeComposition (DNDC) and DayCent models can provide reliable estimations of components of the nitrogen (N) cycle but have rarely been evaluated for a more complete N balance. This is important in order to assess the long-term effects of management practices on soil and environmental quality. Using published data collected from a long term study in the Canadian semi-arid prairie, the Canadian DNDC version (DNDC v.CAN) and DayCent models were evaluated for their ability to simulate the long term nitrogen dynamics and budgets as well as nitrogen use efficiencies (NUEs) in a loam/silt loam soil for three distinct spring wheat (Triticum aestivum L.) cropping systems. Both DNDC v.CAN and DayCent models predicted the spring wheat grain yields, above-ground plant biomass and nitrogen uptake well. The predicted NUEs in DNDC v.CAN, calculated using two approaches with respect to grain yield and grain N concentration, indicated good correlations to the observed values with r≥0.70 and low biases and average relative errors. The N balances were also simulated well in the two models, however DayCent showed a higher estimate of the deficit between N inputs and outputs, termed ‘Unaccounted N’, in all three systems compared to DNDC v.CAN. For both model simulations and the observed data, N outputs in the form of grain N uptake and N losses (nitrogen leaching, N gas emissions) were greater than N inputs except in the ContW (NP) system. In general, a multiple linear regression for estimations of NUEs with respect to N balance and N inputs across all three cropping systems showed that, DNDC v.CAN correlated better with the observed data compared to DayCent. Thus, based on model performance in this study, DNDC v.CAN as a process-based model offers promise as a tool for analyzing different cropping systems with varying N rates in terms of N dynamics and subsequent environmental impacts and benefits.