Interaction between Synthetic N Fertilization and Rainfall on Nitrous Oxide Emissions in Spring Wheat Fields Measured using Flux Towers and Estimated using DNDC and STICS Models

English Abstract:

Synthetic N fertilization of cultivated fields contributes to nitrous oxide (N₂O) emissions. These emissions are enhanced when soil water saturation is in the range of ~60-90%, when incomplete denitrification is dominant. The timing and intensity of rainfall influence the occurrence and magnitude of N₂O fluxes. This study presents daily N₂O emissions measured during four growing season (2001, 2003, 2005, 2011) of spring wheat in eastern Canada. Twin flux gradient towers were used to measure N₂O emissions for contrasting N fertilizer application rates and forms in order to compare the influence of management practices on these emissions and the response to climate. The N rates were as follows: 41 and 68 kg N ha^-1 of ammonium nitrate (AN) in 2001; 45 and 78 kg N ha^-1 of urea in 2011, 62 kg N ha^-1 of urea in 2003 and AN and urea applied at a rate of 77 kg N ha^-1 in 2005. Although N₂O emissions tended to increase with AN application, no significant difference was observed between the N forms, therefore the N rate was studied regardless of the form. Because the synthetic N fertilizers were mostly applied early May in spring wheat, the magnitude of the N₂O emissions were linearly correlated to the cumulative rainfall in May (R²=0.74). The contributions of the N₂O emissions in May to the seasonal ones were: 54% in 2001 (for 93 mm of rain in May), 65% in 2003 (for 117 mm of rain in May), 28% in 2005 (for 45 mm of rain in May), 57% in 2011 (for 89mm of rain in May). Although the rainfalls were similar in May and over the season for 2001 and 2011 (281 mm vs 303 mm, respectively), the limited and recommended N rates did not systematically generate contrasting N₂O emissions. In 2011, there was no significant difference following the two N application rates, while in 2001 the emissions decrease by 23% for the limited N rate. The difference in rainfall distribution and intensity between the two years would explain the results. In 2005, which had the lowest rainfall in May and the largest seasonal rainfall (343 mm), the N₂O emissions associated with the recommended N rate were half those observed either in 2001 or 2011. The prediction of N₂O emissions obtained using DNDC and STICS models for 2001-2005 will also be discussed.