Nutrient Cycling in Agroecosystems 60: 159–175, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands. 159 Comparing a process-based agro-ecosystem model to the IPCC methodology for developing a national inventory of N 2 O emissions from arable lands in China Changsheng Li 1 , Yahui Zhuang 2 , Meiqio Cao 2 , Patrick Crill 1 , Zhaohua Dai 2 , Steve Frolking 1,∗ , Berrien Moore III 1 , William Salas 1 , Wenzhi Song 2 & Xiaoke Wang 2 1 Institute for the Study of Earth, Oceans, and Space, Morse Hall, University of New Hampshire, Durham, NH 03824, USA; 2 Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, PO Box 2871, Beijing 100085, China ( ∗ Corresponding author; e-mail: steve.frolking@unh.edu) Key words: assessment, China, cropland, IPCC, modeling, Nitrous oxide, N 2 O Abstract Nations are now obligated to assess their greenhouse gas emissions under the protocols of Article 4 of the United Nations Framework Convention on Climate Change. The IPCC has developed ‘spreadsheet-format’ methodologies for countries to estimate national greenhouse gas emissions by economic sector. Each activity has a magnitude and emission rate and their product is summed over all included activities to generate a national total (IPCC, 1997). For N 2 O emissions from cropland soils, field studies have shown that there are important factors that influence N 2 O emissions at specific field sites that are not considered in the IPCC methodology. We used DNDC, a process- oriented agroecosystem model, to develop an unofficial national inventory of direct N 2 O emissions from cropland in China. We assembled county-scale data on soil properties, daily weather, crop areas, N-fertilizer use, livestock populations (for manure inputs to cropland), and agricultural management for the 2500 counties in mainland China. Total 1990 cropland area was 0.95 million km 2 . Total N-fertilizer use in China in 1990 was 16.6 Tg N. The average fertilization rate was 175 kg N ha -1 cropland. One-year simulations with DNDC were run for each crop type in each county to generate estimates of direct N 2 O emissions from soils. National totals were the sum of results for all crop simulations across all counties. Baseline simulations estimated that total N 2 O emission from arable land in China in 1990 was 0.31 Tg N 2 O-N yr -1 . We also ran simulations with zero N-fertilizer input; the difference between the zero-fertilizer and the baseline run is an estimate of fertilizer-induced N 2 O emissions. The fertilizer- induced emission was 0.13 Tg N 2 O-N yr -1 , about 0.8% of total N-fertilizer use (lower than the mean but within the IPCC range of 1.25±1.0%). We compared these results to our estimates of county-scale IPCC methodology emissions. Total emissions were similar but geographical patterns were quite different. Introduction Article 4 of the United Nations Framework Conven- tion on Climate Change states that ‘All Parties ... shall ... develop, periodically update, [and] publish ... national inventories of anthropogenic emissions by sources and removals by sinks of all greenhouse gases not controlled by the Montreal Protocol, using comparable methodologies ...’ In response to this man- date, the Intergovernmental Panel on Climate Change (IPCC) has coordinated groups of experts to develop and periodically update national inventory methodolo- gies for various greenhouse gases. These methodolo- gies are generally developed separately by greenhouse gas and economic sector, e.g., N 2 O emissions from agriculture. In order to meet the requirement of being applicable by all nations the methodologies must be developed such that nations will have or can generate the data necessary to apply the methodology to de- velop a national inventory. This has led to assessment methodologies that are in a ‘spreadsheet’ format. In these methodologies, each activity (e.g., synthetic ni- trogen fertilizer use) has a magnitude (e.g., tonnes of N applied per year) and an emission rate (e.g., 1.25%