Emission Factor of Nitrogen for N2O

The fertilizer-induced emission factor of N2O in the present study was, on average, 0.42% for the water regime of F-D-F and 0.73% for the F-D-F-M. Obviously, these estimated emission factors of N2O in rice paddies are significantly lower than the IPCC (1997) default factor of 1.25% or estimates in upland croplands in this area (Zheng et al. 2004). Yan et al. (2003) estimated that N2O emission factors and background emissions averaged 0.25% and 0.26kgN2O-N ha-1 in the rice growing season, respectively. However, they did not distinguish N2O emissions under different water regimes in rice paddies. In contrast, Akiyama et al. (2005) recently reported that the EFs averaged 0.22% for the continuous flooding rice paddies and 0.37% for the fertilized paddies with mid-season drainage. These estimates rep resent the mean of 16 and 23 emission factors directly measured from field studies in which both nitrogen and no-nitrogen treatments were designed, respectively. As the authors pointed out, seasonal total N2O emissions were not significantly related to nitrogen input during the rice growing season over all water regimes or for continuous flooding (Akiyama et al. 2005). However, based on the data that were exhibited in Table 9.2 by Akiyama et al. (2005), a pronounced relationship between N2O emission and N input during the rice growing season was found in the rice paddies with mid-season drainage. Using Akiyama et al. (2005) identical data (excluding measurements from nitrification inhibitors and controlled released fertilizers), the simulated OLS linear model MSD-Akiyama predicted that N2O emission factor averaged 0.43%, with a background emission of 0.20kgN2O-N ha-1 (Table 9.5). This emission factor simulated by the OLS model is slightly higher than that obtained by Akiyama et al. (2005), which was based on the Maximum Likelihood (ML) estimate.

Indeed, the emission factor of N2O estimated by Akiyama et al. (2005) refers to the value of EF that maximizes the likelihood of observing N2O emission in measurements. Point and interval estimation using the ML model relies heavily on distributional assumptions that a sample of observations (response variables) has a normal distribution (Quinn and Keough 2004). However, the direct EF data in Akiyama et al. (2005) study had a log-normal distribution pattern. In contrast, EF estimated by the OLS model in this study represents the value with the least uncertainty. The OLS point estimates require no distributional assumptions for variables, but instead concentrate on residual distribution (Quinn and Keough 2004). As shown in Fig. 9.3d, the residuals of the model MSD-Akiyama were close to being normally distributed. A power analysis also showed that it was strong enough to model the data (Table 9.4). In order to minimize the uncertainty in estimates of emission factor for N2O, presumably, the OLS model could be more appropriate for N2O data than the ML model used by Akiyama et al. (2005).

9.4.3 Background N2O Emission

In the present study, background emission of N2O in rice paddies was pronounced only under the water regime of F-D-F-M (Table 9.5). Background N2O emissions during the rice growing season were negligible for the continuous flooding paddy fields. Under the water regime of F-D-F, background N2O emission was estimated to be 0.009 kg N2O-N ha-1, which was not significantly different from "0" (Table 9.5). This negligible background emission is partially due to water management that water logging dominated over the rice season except for a short-term episode of mid-season drainage. An intensive N2O emission occurred only in the course of mid-season drainage under the F-D-F. In contrast, background N2O emission was, on average, as high as 0.79 kg N2O-N ha-1 under the F-D-F-M water regime. Based on seven measurements in the continuous flooding and intermittent irrigation rice paddies, Yan et al. (2003) estimated background N2O emission, on average, to be

0.26kgN2O-N ha-1 in the paddy rice season. However, as the authors acknowledged, there was much uncertainty with respect to the background emission estimate in paddy fields. Indeed, the background emission has become one of the most sensitive factors for developing an inventory of agricultural N2O emissions (Bouwman et al. 2002; Yan et al. 2003; Akiyama et al. 2005).

Was this article helpful?

0 0
Renewable Energy Eco Friendly

Renewable Energy Eco Friendly

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable.

Get My Free Ebook


Post a comment