연구보고서
- 저자
- 음형일 박사
- 작성일
- 2017.07.04
- 조회
- 410
- 요약
- 목차
The global water cycle, driven by global warming, may have impacts on natural and human systems, resulting in hazardous consequences due to changes in intensity and frequency of extreme events. Climate-induced changes have caused temporal
and spatial alterations of hydrologic regimes, which may lead to necessary modifying and rebuilding water-resource planning and management in order to adapt to climate change. Therefore, it is a prerequisite step to understand the alteration of hydrologic
regimes under climate change for sustainable water-resource management in a system. Assessment of climate change impacts on water resource systems has suffered from the large uncertainty in hydrologic projections. Identifying the sources of uncertainty and quantifying the contribution of each component in a model chain are crucial not only to prevent from erroneous analysis, but also to decrease the uncertainty in future hydrologic projections.
Global climate models (GCMs) provide the fundamental information used to assess potential impacts of future climate change. However, the mismatch in spatial resolution between GCMs and the requirements of regional applications has impeded the use of GCM projections for impact studies at a regional scale. In addition to Quantile Mapping (QM), a traditional bias correction method, this study applied state-of-the-art statistical downscaling methods that preserve long-term temporal trends: Bias-Correction/Spatial Disaggregation with Detrended Quantile Mapping (BCSD/DQM); Quantile Delta Mapping (BCSD/QDM); Bias-Correction/Contructed
Analog with Detrended Quantile Mapping (BCCA/DQM) and BCCA with Quantile Delta Mapping (BCCA/QDM). These methods downscale 26 CMIP5 GCM climate projections under Representative Concentration Pathway (RCP) 4.5 and 8.5 for daily precipitation,
minimum temperature, and maximum temperature over South Korea. Therefore, six statistical downscaling methods were employed to generate downscaled climate projections; BCSD/QM, BCSD/DQM, BCSD/QDM, BCCA/QM, BCCA/DQM, and BCCA/QDM. This study also presented a standardized evaluation framework for ranking 26 GCMs using the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). Employing the downscaled climate projections into multiple hydrologic models that have different structures, this study set up a model chain that consists of two RCPs, twenty-six GCMs, six downscaling methods (DS), and three hydrologic models (HM). For the model chain, this study quantified the contributions of components to the uncertainty in hydrologic alteration indicators for multi-purpose reservoirs in major Korean river basins, using a variance-based approach.
The results showed that the statistical downscaling methods significantly improved the performance in extreme climate indices. QDM preserved GCM-driven long-term trends much better than QM for climate indices that show a strong upward (or downward) trend. Climate change projections showed an increase in both extreme precipitation events and consecutive-dry-day index, which implies a more frequent occurrence of flood and drought events in the future. This study found that the spatial resolution of GCM is closely related to the skill of GCM for temperature in terms of the standardized evaluation framework, whereas there is no relationship of ranking with resolution for precipitation. This may be a result of low skill of GCM in simulating local precipitation, mainly due to coarse resolution. Regarding hydrologic projections, future yearly and seasonal flows increase when compared with those of the reference period, with the exception of winter flow, which showed the increasing frequencies in high and low terciles and decreasing frequency in the middle tercile. This implies that polarized flows (i.e. extreme high or low flows) may occur in the future during the winter season, which result in decreasing extreme minimum-flow related hydrologic indicators. Three hydrologic models showed higher variability in simulating the extreme minimum-flow related hydrologic indicators, due to different skills of hydrologic models in simulating low flows. As a result of the contributions of components in the model chain to the uncertainty in hydrologic indicators, the uncertainty in extreme minimum-flow related hydrologic indicators is dominated mainly by the choice of GCMs and HMs (>90%), where the relationship between rainfall and runoff is similarly high during the winter season (i.e. high runoff coefficient). For extreme high-flow related hydrologic indicators, the contribution of RCP’s choice became up to 18% for 7 and 90 day maximum flows. For other hydrologic indicators, the hydrological projections are associated with high uncertainty, mainly due to the choice of GCM, with 78% ~ 96% of total variance for the hydrologic indicators.
The Paris 21st Conference of the Parties in December 2015 set the agreement of a global action plan to mitigate climate change impacts by limiting global warming to below 2°C. After the Paris agreement, RCP2.6 has been the focus for scientists to investigate the impacts of climate change. Dynamical downscaling methods (regional climate models) may provide different types of climate projections considering physics and non-stationarity in climate variables. However, this study employed only statistical downscaling methods, which may project a limited range of climate variability. Because this study applied lumped hydrologic models despite
the different structures, physical-based fully distributed hydrologic models should be included in a model chain such as Soil and Water Assessment Tool (SWAT) and Variable Infiltration Capacity (VIC). Therefore, more various range of models at each component in a model chain may lead to a more comprehensive assessment of climate change impacts on hydrologic indicators and quantify the contributions of all components to the uncertainty in hydrologic projections.