Author(s): Jae-Kyoung Lee; Young-Oh Kim
Linked Author(s): Young-Oh Kim
Keywords: Climate change; Uncertainty propagation; Maximum entropy
Abstract: The projection of future water resources in the previous decades from the climate change impact assessment should have large uncertainties essentially. In addition, it is necessary to study the uncertainty of climate change because the projections from emission scenarios or GCMs as the beginning of impact assessment are very diverse. However, until now, the process for quantifying climate change uncertainties in previous studies is similar to the sensitivity analysis instead of the uncertainty analysis. In other words, all most studies estimated the variation range of output scenarios according to the variation of input scenarios. Therefore, this study quantifies the total uncertainty in the whole process of impact assessment that consists of a serial combination of several models (GCM, downscaling, hydrological model, and others) and the uncertainty and the ratio in each step (or model) using the maximum entropy. Furthermore, this study proposes the framework of uncertainty propagation in climate change. The methodology of this study was as follows: The first was that this study proposed the uncertainty propagation as well as the method for estimating the uncertainty in each step and classified the total uncertainties to emission scenario uncertainties and model uncertainties using the concept of uncertainty propagation from the IPCC 3rd Assessment Report. The second was that a process of climate change assessment including 2 emission scenarios, 4 GCMs, 2 downscaling techniques, and 2 hydrological models was conducted to estimate uncertainties in the impact assessment. The existing method using sensitivity analysis was also considered to compare the results from the uncertainty propagation method proposed in this study. In results, the existing method proposed that the GCMs were recognized as the largest contributor to total uncertainty (89. 34 % ) and this result was identical to existing results. The uncertainty propagation method proposed the uncertainty increased gradually through the process of climate change impact assessment because it was propagated step-by-step. Especially, emission scenarios in all uncertainty sources significantly contributed to the total uncertainties (larger than 50 % ). The GCMs were also main contributor in model uncertainties (26. 50 % of overall uncertainties). Accordingly, the future water resources will be projected very differently, depending on which emission scenarios and GCMs are selected. Therefore, it has been proven that the uncertainty propagation method can explain the entire cascade of uncertainties in the process of climate change impact assessment and quantify the uncertainties in each step.
Year: 2013