Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories Indirect N 2 O Emissions from Agriculture 381 INDIRECT N 2 O EMISSIONS FROM AGRICULTURE

Agricultural perturbations to the global nitrogen cycle, directly and indirectly, lead to enhanced biogenic production of nitrous oxide (N2O). Direct pathways include microbial nitrification and denitrification of fertiliser and manure nitrogen that remains in agricultural soils or animal waste management systems. Indirect pathways involve nitrogen that is removed from agricultural soils and animal waste management systems via volatilisation, leaching, runoff, or harvest of crop biomass. Like their direct counterparts, the long-term fate of agricultural nitrogen also eventually provides substrate for microbial nitrification and denitrification, with associated N2O production. According to the Revised 1996 National Greenhouse Gas Inventories (IPCC Guidelines,) and Mosier et al., 1998), indirect N2O emissions account for one third of the total global agricultural N2O source and approximately two thirds of the uncertainty in the total source. Indirect emissions include small contributions from human sewage and volatilised nitrogen, but are derived predominantly from nitrogen lost through leaching and surface runoff. Because of their importance in defining both the magnitude and the uncertainty of the agricultural N2O source, leaching/runoff emissions deserve high priority in any consideration of “good practice” in the IPCC methodology. Three major areas of uncertainty in estimating leaching-related N2O emissions are identified in this report. Firstly, the entire amount of fertiliser and manure, including that assumed to be volatilised and burned as fuel is subject to a default leaching fraction of 0.3. Secondly, under current practices, this default leaching fraction is commonly used by all countries, despite large variations within individual watersheds and agricultural systems. Finally, the N2O emission factor assigned to leached nitrogen is estimated from a 3-step derivation which tracks the leached fraction through groundwater, rivers and estuaries, and broadly assumes some microbial N2O production at each step on the basis of limited information. The groundwater step, which currently accounts for 60 percent of leaching-related N2O emissions, is particularly problematic. In re-evaluating leaching/runoffrelated N2O emissions, all of these areas of uncertainty should be addressed. The volatilisation N2O source in theory is meant to represent the fraction of fertiliser and manure that volatilises to NOx and NH3 soon after application/deposition and subsequently redeposited on nearby soil, providing substrate for nitrification and denitrification. In practice, the current methodology formulation is fairly insensitive to assumptions about the volatilised fraction of fertiliser and manure. This is true because the direct and volatised N2O emission factors are essentially the same (0.0125 and 0.01, respectively). Thus a re-evaluation of the default volatilisation fractions probably is not a first priority in refining the methodology. One question with respect to the volatilisation source might be, whether the distinction between 0.0125 and 0.01 is meaningful and is meant to be so. Unlike leached and volatilised nitrogen, which are intricately linked to the direct and animal emissions methodologies, sewage nitrogen is estimated from independent activity data. The independence of the sewage nitrogen calculation is both a strength and a weakness of the current methodology. While cross-checks based on independent inputs are useful for validation purposes, one of the original goals of the methodology was to track agricultural nitrogen from “cradle to grave” in an internally consistent manner. Future refinement of the methodology might involve consistently linking harvested crop biomass to the long-term fate of sewage nitrogen, food waste, animal fodder and manure. Quality assurance (QA) and quality control (QC) activities should encourage the collection of more countryspecific information to estimate the various fates of agricultural nitrogen and its associated N2O emission factors. Because these data are sparse and likely to remain so in the near future, QA/QC efforts should also consider independent methods for validating methodology estimates. Two independent methods include the approach of Seitzinger and Kroeze (1998) for estimating nitrogen loading to rivers, and semi-empirical gas transfer calculations, in which N2O flux from estuaries is estimated based on observed surface supersaturation of N2O and modelled gas-transfer coefficients. When compared to the methodology estimates, the former would be useful for validating the estimated amount of leached nitrogen and the latter would be useful for validating N2O