Differentiation of nitrous oxide emission factors for agricultural soils Jan Peter Lesschen * , Gerard L. Velthof, Wim de Vries, Johannes Kros Alterra, Wageningen UR, P.O. Box 47, 6700 AA Wageningen, The Netherlands article info Article history: Received 25 March 2011 Accepted 1 April 2011 Keywords: Controlling factors Inference scheme Nitrous oxide Soil Tier 2 approach abstract Nitrous oxide (N 2 O) direct soil emissions from agriculture are often estimated using the default IPCC emission factor (EF) of 1%. However, a large variation in EFs exists due to differences in environment, crops and management. We developed an approach to determine N 2 O EFs that depend on N-input sources and environmental factors. The starting point of the method was a monitoring study in which an EF of 1% was found. The conditions of this experiment were set as the reference from which the effects of 16 sources of N input, three soil types, two land-use types and annual precipitation on the N 2 O EF were estimated. The derived EF inference scheme performed on average better than the default IPCC EF. The use of differentiated EFs, including different regional conditions, allows accounting for the effects of more mitigation measures and offers European countries a possibility to use a Tier 2 approach. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Nitrous oxide (N 2 O) is one of the major greenhouse gasses with a contribution of 8% to the anthropogenic global warming (IPCC, 2007). Fifty to sixty percent of the anthropogenic induced N 2 O emissions comes from agriculture of which the major part is direct emission from agricultural soils (Mosier et al., 1998). The N 2 O soil emissions from applied fertilizer and manures are often estimated using a default emission factor (EF). In the IPCC 2006 guidelines the updated default EF for N inputs from mineral fertilizers, organic amendments and crop residues (EF 1 ) is 1% (IPCC, 2006; De Klein et al., 2006), i.e. the direct fertilizer-derived N 2 O soil emission is equal to 1% of the amount of N applied. This factor is based on a large number of measurements (Bouwman et al., 2002a,b; Stehfest and Bouwman, 2006; Novoa and Tejeda, 2006), which lead to a mean value close to 0.9%. However, given the uncertainties associated with this value, IPCC (2006) considered the round value of 1% appropriate. Nevertheless, a large variation in EFs exists (Stehfest and Bouwman, 2006; Flechard et al., 2007). Based on the Stehfest and Bouwman (2006) data set the control emissions corrected EFs (n ¼ 352) range from 0.0% to 10.8% with an average of 1.1% and standard deviation of 1.7%. This large variation is due to differences in environmental factors (e.g. climate and soil conditions), crop factors (e.g. crop type and crop residues) and management factors (e.g. type of manure and fertilizer, application rate, time of application). In this paper we elaborate a simple and transparent approach to estimate direct N 2 O soil emission using EFs that depend on N-input sources and environmental factors. The approach is mainly based on literature data and expert knowledge and is intended for implementation in large-scale models such as INTEGRATOR (De Vries et al., this issue) and MITERRA-Europe (Velthof et al., 2009), which contain spatially explicit information on fertilizer and manure types and use, soil properties and climate, for use in European wide applications. Although the focus is on Europe we did not confine our literature survey to Europe, but used data from experiments in agricultural systems of temperate regions. The EFs used in our study all refer to values corrected for the control emissions, thus representing the fertilizer induced emissions. This paper starts with the conceptual framework and an over- view of the different factors that control N 2 O emissions. We reviewed literature data to quantify N 2 O EFs, using studies that compared N 2 O emissions for different N inputs or for different environmental factors. Based on these data and expert knowledge, an inference scheme for N 2 O EFs for the different sources of N input and different environmental conditions was established. Next, we incorporated the developed inference scheme in the INTEGRATOR model, applied this for the EU-27, and compared the calculated N 2 O emissions from agricultural soils with the emissions using the default IPCC EF. Finally, we evaluated our N 2 O EF inference scheme based on the Stehfest and Bouwman (2006) data set. 2. Methods 2.1. Conceptual framework The major factors that control N 2 O emission are N input and nitrate content, oxygen content, available C content, temperature and pH (Firestone et al., 1980; Granli and Bøckman, 1994; Sahrawat and Keeney, 1986; Tiedje, 1988). Table 1 gives an overview of the effects of these controlling factors on denitrification and * Corresponding author. E-mail address: janpeter.lesschen@wur.nl (J.P. Lesschen). Contents lists available at ScienceDirect Environmental Pollution journal homepage: www.elsevier.com/locate/envpol 0269-7491/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2011.04.001 Environmental Pollution 159 (2011) 3215e3222