Abstract
Although global integrated assessment models have been used for the long-term climate mitigation analysis, they are rarely evaluated their predictability. This article made an experiment of hindcasting by using AIM/ CGE (Asia-Pacific Integrated Model/ Computable General Equilibrium) which is one of the IAMs and evaluated the model results against to the historical statistics. The historical simulation is made under exogenous GDP, population and energy technological progress assumptions as well as future simulation and the energy consumption is compared with statistics information. There are three main findings. First, global aggregated information tend to show high reproducibility and for example, total energy primary energy is about 10% discrepancy to statistics. However, those of disaggregated regional results are lower than that and especially low income countries' results tend to be low. Second, as years go past, the reproducibility of primary energy supply and final energy gets lower. Third, reproducibility in terms of energy sources, oil and natural gas, and in terms of sectors, transport sector showed lower reproducibility than others and there are rooms to improve the model in these components.
