Land-use change effects on soil emissions of N 2 O in the tropics: a 3-continent comparative analysis Introduction Nitrous oxide (N 2 O) predominantly emitted by soils is a greenhouse gas with a global warming potential 300 time larger than carbon dioxide over a 100 year time horizon. Tropical forests are the largest natural source of N 2 O of all biomes. Changes to this flux associated with land-use change could be a source of positive feedback to the climate system. While we understand general trends and have been able to explain how land-use change affects temporal and spatial variability of this flux, our ability to extrapolate from a study in one region to other regions and predict the magnitude of the changes in fluxes remains weak. The objective of this study is to assess the deterministic factors affecting N 2 O emission changes across a dataset of tropical studies including different land-use change trajectories in varying soil, climate and nitrogen (N) availability conditions. Methods We combined four datasets of N 2 Ofluxes and associated air and soil temperatures, soil respiration rates, water-filled pore space (WFPS), ammonium (NH 4 + ) and nitrate contents (NO 3 - ), and net nitrification and mineralization rates. The land-use changes considered and a brief description of the sites are presented in Table 1. Most of the land-uses were unfertilized except the recently cleared land and 3-year-old coffee plantation on a landslide of South Sumatra that received 25 and 50 kg ha -1 of urea, respectively, 5 months and several days previous to measurements. The 8-year-old oil palm plantation in Jambi, Sumatra also received one dose of 71 kg urea ha -1 during the year. We tested the normality of the distribution of soil N 2 O emissions, verified whether there was a general seasonal or land-use effect on the emissions, compared the fertilized treatments to the non-fertilized ones and proceeded to simple and multiple regression analysis. Results Soil emissions of N 2 O weren’t normally distributed nor were their logarithmic transformation. Across the dataset, emissions were significantly (P = 0.003) smaller during the dry (3.6 ± 0.3 g N-N 2 O ha -1 d -1 ) than during the wet season (6.7 ± 1.0 g N- N 2 O ha -1 d -1 ), as were the corresponding average WFPS values (54.8 ± 1.8% and 69.1 ± 1.1% during the dry and wet seasons, respectively). Indeed we found a significant exponential relationship between soil WFPS and N 2 O emissions (Figure 1). A multiple regression indicated that N 2 O emissions were primarily related to soil respiration and air temperature (R 2 = 0.1, P < 0.001, N = 162) as follow: Ln(N 2 O) = 0.01*CO 2 - 0.1*AirTemp + 3.22 With the soil respiration ranging from 8 to 489 kg C-CO2 ha -1 d -1 and the air temperature from 17 to 39°C. Another analysis performed with a smaller dataset (N = 75) gave better results (R 2 = 0.36, P < 0.05) and showed that N 2 O emissions were associated with changes in both WFPS and nitrate content: Ln(N 2 O) = 0.03*WFPS + 0.06*NO 3 -1.38 With soil NO 3 contents in the range [0-32] mg N kg -1 d.w. Finally, N-fertilized system emitted significantly (P = 0.02) more N 2 O than non fertilized systems. Conclusion This study demonstrated that the changes in N 2 O emissions following different tropical forest conversions in varying soil, climate and N availability conditions are determined by and can be expressed as a function of the changes in factors known to control soil-atmosphere trace gas exchange. The inclusion of additional datasets, especially those comprising mineral N availability and turnover, would help improving our predictive capacity on the magnitude of flux changes. Hergoualc’h K. 1 , Verchot L.V. 1 , Aini F. 1 , Brienza Júnior S. 2 , Cattânio J.H. 2 , Costa de Oliveira V. 3 , Davidson E. 3 , Hairiah K. 4 , Neufeldt H. 4 , Thiongo M. 4 , van Noordwijk, M. 4 1 CIFOR, 2 EMBRAPA, 3 WHRC, 4 ICRAF For further information, please contact k.hergoualch@cgiar.org Table 1: Land-use changes studied and site description. References Aini, F. K., Hergoualc’h, K., Verchot, L. V., Smith, J. 2011 CH 4 and N 2 O flux changes from forest conversion to rubber and oil palm plantation in Jambi, Sumatra, Indonesia. Presented orally at the 6 th International Symposium on non-CO2 Greenhouse Gases (NCGG-6), Amsterdam, the Netherlands, 2- 4 November 2011. Verchot, L.V., Davidson, E.A., Cattânio, J.H., Ackerman, I.L., Erickson, H.E., Keller, M. 1999 Land use change and biogeochemical controls of nitrogen oxide emissions from soils in eastern Amazonia. Global Biochem Cy13, 31-46. Verchot, L., Hutabarat, L., Hairiah, K., van Noordwijk, M. 2006 Nitrogen availability and soil N 2 O emissions following conversion of forests to coffee in southern Sumatra. Global Biogeochem Cy 20, doi:10.1029/2005GB002469. Country Soil type Annual rainfall (mm) Altitude (m) Land-uses Brazil (Eastern Amazonia) Latosol 1850 < 200 Primary forest Secondary forest Degraded pasture Active pasture Indonesia (Southern Sumatra) Inceptisol 2570 1000 Forest Wet forest Recently cleared land Coffee 1, 3, 7,10 yr Coffee on landslide 3 yr Rorak* coffee 10 yr Multistrata coffee 18 yr Grass fallow Indonesia (Jambi, Sumatra) Oxisol/ Inceptisol 2500 100 Primary forest Disturbed forest Rubber 1, 20 yr Oil palm 8 yr Kenya (Kakamega, Western Kenya) Ultisol 1900 1600 Forest/Maize 1920s Forest/Maize 1940s Forest/Pasture 1986 (paired plots**) * Rorak is a local sediment-trapping structure built to reduce erosion losses and maintain soil fertility. ** 1920s, 1960s and 1986 indicate the date of forest conversion. -6 -4 -2 0 2 4 6 5 25 45 65 85 105 125 Ln (N 2 O) (g N-N 2 O ha -1 d -1 ) WFPS (%) y = 0.02x - 0.115 R² = 0.1 N = 314 Figure 1: Relationship between daily soil N 2 O emissions and water-filled pore space (WFPS). View publication stats View publication stats