Modelling soil compaction impacts on nitrous oxide emissions in arable fields

Abstract

Nitrous oxide (N2O) emissions comprise the major share of agriculture's contribution to greenhouse gases; however, our understanding of what is actually happening in the field remains incomplete, especially concerning the multiple interactions between agricultural practices and N2O emissions. Soil compaction induces major changes in the soil structure and the key variables controlling N2O emissions. Our objective was to analyse the ability of a process-based model (Nitrous Oxide Emissions (NOE)) to simulate the impact of soil compaction on N2O emission kinetics obtained from field experiments. We used automatic chambers to continuously monitor N2O and CO2 emissions on uncompacted and compacted areas in sugar beet fields during 2 years. Soil compaction led to smaller CO2 emissions and larger N2O emissions by inducing anoxic conditions favourable for denitrification. Cumulative N2O emissions during the crop cycles were 944 and 977 g N ha?1 in uncompacted plots and 1448 and 1382 g N ha?1 in compacted plots in 2007 and 2008, respectively. The NOE model (H´enault et al., 2005) simulated 106 and 138 g N2O-N ha?1 in uncompacted plots and 1550 and 650 g N2O-N ha?1 in compacted plots in 2007 and 2008, respectively, markedly under-estimating the nitrification rates and associated N2O emissions. We modified the model on the basis of published results in order to better simulate nitrification and account for varying N2O fractions of total end-products in response to varying soil water and nitrate contents. The modified model (NOE2) better predicted nitrification rates and N2O emissions following fertilizer addition. Using a fine vertical separation of soil layers of configurable, but constant, thickness (1 cm) also improved the simulations. NOE2 predicted 428 and 416 g N-N2O ha?1 in uncompacted plots and 1559 and 1032 g N- N2O ha?1 in compacted plots in 2007 and 2008, respectively. These results show that a simple process-based model can be used to predict successfully the post-fertilizer addition kinetics of N2O emissions and the impact of soil compaction on these emissions. However, large emissions later on during the cropping cycle were not captured by the model, emphasizing the need for further research. (Résumé d'auteur

Reducing nitrous oxide emissions by changing N fertiliser use from calcium ammonium nitrate (CAN) to urea based formulations

“…Long-term undisturbed grasslands will have a greater proportion of fungi to bacteria that could result in higher N 2 O losses due to the fact that fungi lack the N 2 O reductase enzyme to reduce N 2 O to N 2 (Shoun et al, 1992, Laughlin et al, 2003. In addition soils with high clay content are more prone to soil compaction which further induces anoxic conditions favourable for denitrification by affecting oxygen availability and generating larger and longer lasting emissions (Bessou et al, 2010).…”

Using a crop model to account for the effects of local factors on the LCA of sugar beet ethanol in Picardy region, France

“…In the current version of CERES-EGC, nitrogen-fertilisers are distributed over the first 30 cm and nitrification and denitrification calculated over the first 20 cm with average temperature and water content for two sub-layers across these 20 cm. The discretization of ammonium distribution and soil water and temperature within fine soil layers could improve the accuracy of model estimation (Bessou et al 2010b). The modelling of N dynamics in soil also deserves further improvements by differentiating between fertilizer incorporation depths and potentially improved sub-routine for N-release from cover crop residues.…”

“…Calibrated parameters were then used for the simulations of sugar beet cropping system with CERES-EGC (Leviel et al 2000;2003). Despite our efforts, CERES-EGC still underestimated emissions in [2008][2009] due to several combined factors (Bessou et al 2010b). Other sources also showed a marginal error of roughly 20% with CERES-EGC (Hénault et al 2005;Gabrielle et al 2011).…”

“…This calibration reduced the Root-Mean Square Error on daily and annual fluxes by 15% and 63% respectively when compared to the uncalibrated simulations ( Fig. 2), based on experimental data collected in our study site over sugar-beet (Bessou et al, 2010b). Calibrated parameters were then used for the simulations of sugar beet cropping system with CERES-EGC (Leviel et al 2000;2003).…”

“…It also predicts substance losses outside the ecosystem such as nitrate leaching, ammonium volatilization (Génermont and Cellier 1997), and nitrous oxide emissions through nitrification and denitrification which are predicted with the NOE sub-model (Hénault et al 2005). In order to improve its accuracy, we implemented the NOE2 submodel (Bessou et al 2010b) in CERES-EGC and proceeded to a Bayesian calibration of the 22 parameters of NOE2. This calibration makes it possible to reduce significantly the spread in parameter distribution and the subsequent uncertainty on N 2 O emissions by updating these distributions against a priori probability distributions of parameter values gathered from other sites (Lehuger et al 2009).…”