Evaluating management effects on nitrous oxide emissions from grasslands using the process-based DeNitrification–DeComposition (DNDC) model

Abstract

The development of agricultural mitigation strategies to reduce greenhouse gas (GHG) emissions is urgent in the context of climate change – land use interactions. In this study the DNDC biogeochemical model was used to study nitrous oxide (N2O) emissions from grazed grasslands in southern Ireland. The objectives of this study were: (1) to evaluate the DNDC model using a two year (2008–2009) data set of chamber measured N2O fluxes at eight grassland sites and (2) to investigate the impact of different management scenarios on N2O emissions including changes in i) inorganic nitrogen (N) fertilizer application rates ii) slurry application rates; and iii) animal density (livestock unit per hectare LU ha−1). The comparison of modeled daily DNDC fluxes (using a combination of measured and default soil parameters) and measured fluxes resulted in an r (coefficient correlation) = 0.48. To improve the model performance, the fluxes for 2008 were used in a calibration exercise during which the soil properties were optimized to obtain the best fit of N2O fluxes. This resulted in an improved model performance, with an r = 0.62. In a validation exercise using 2009 data, we used the model parameters set (e.g. soils) from the calibration exercise and this resulted in a model performance with an r = 0.57. The annual N2O fluxes (measured and modeled) were appreciably higher than those estimated using the IPCC emissions factor of 1.25%. In scenario analysis, the modeled N2O fluxes only increased/decreased on average ±6% and ±7% following a 50% increase/decrease of inorganic N and slurry N applications respectively. These modeled scenario % changes are much lower than the IPCC emission factor % changes of a 50% increase in N2O emissions for a 50% increase in nitrogen applied. An absolute change scenario (±50 kg) in inorganic N and slurry N resulted in greater change in N2O fluxes (±9% inorganic N and ±17% slurry N) as compared to the relative change scenario (above). Furthermore, DNDC N2O flux estimates were not sensitive to changes in animal density (LU ha−1). The latter is a scenario limitation in the current model version. This study suggests that the calibration of soil parameters for Irish conditions is necessary for optimum simulation with DNDC and highlights the potential of management strategies for reducing N2O emissions from grazed grasslands. It further highlights the difference between DNDC and IPCC estimates that require further research.