Vol. 45 Supp. SCIENCE IN CHINA (Series C) October 2002 The consequences of uncertainties in land use, climate and vegetation responses on the terrestrial carbon Rik Leemans, Bas Eickhout, Bart Strengers, Lex Bouwman & Michiel Schaeffer Institute of Public Health and the Environment, P.O. Box 1, 3720 BABilthoven, The Netherlands Correspondence should be addressed to R. Leemans (email: rik.leemans@rivm.nl) Received May 14, 2002 Abstract The IPCC SRES narratives were implemented in IMAGE 2.2 to evaluate the future con- dition of the climate system (including the biosphere). A series of scenario experiments was used to assess possible ranges in emissions and concentrations of greenhouse gases, climate change and impacts. These experiments focussed on the role of the terrestrial carbon cycle. The experi- ments show that the SRES narratives dominate human emissions and not natural processes. In contrary, atmospheric CO 2 concentration strongly differs between the experiments. Atmospheric CO 2 concentrations range for A1B from 714 to 1009 ppmv CO 2 in 2100. The spread of this range is comparable with the full SRES range as implemented in IMAGE 2.2 (515ü895 μmol/mol CO 2 ). The most important negative and positive feedback processes in IMAGE 2.2 on the build-up of CO 2 concentrations are CO 2 fertilisation and soil respiration respectively. Indirect effects of these processes further change land-use patterns, deforestation rates and alter the natural C fluxes. The cumulative effects of these changes have a pronounced influence on the final CO 2 concentrations. Our scenario experiments highlight the importance of a proper parameterisation of feedback proc- esses, C-cycle and land use in determining the future states of the climate system. Keywords: carbon cycle, feedback processes, integrated assessment modeling, land use, SRES scenarios. The Intergovernmental Panel on Climate Change (IPCC) has assessed the consequences of climate change [1, 2] using six Standardised Reference Emission Scenarios (SRES) [3] . These sce- narios were based on qualitative, comprehensive narratives defined by two dimensions. The first dimension differentiates between ‘material consumption’ (A) and ‘sustainability and equity’ (B); the second between ‘globalisation’ (1) and ‘regionalisation’ (2). These dimensions specify typical characteristics and trends for each of the four quadrants (A1, B1, A2 and B2). Different combina- tions of energy carriers and energy efficiency further elaborated upon the A1 scenario. A1F uses only fossil fuels, A1B a balanced mix of fossil and renewables, and A1T is based on fuel-efficient technologies and renewables. The SRES CO 2 emissions range from 5 to 30 Pg C in 2100. These emissions lead to atmospheric CO 2 concentrations ranging from 540 to 970 μmol/mol [1] . When other GHGs and aerosols are added, the range in global temperature increase is 2.0 to 4.5 o C. This range is a result of straightforward use of emissions scenarios in simple models of at- mosphere, carbon (C) cycle and climate. Some uncertainties were considered. The uncertainty related to the climate sensitivity (i.e. the global mean warming at a doubled GHG-forcing) was