Author(s): Daiki Omori; Akira Kurihara; Koji Tanaka; Yutaka Oyagi

Linked Author(s): Daiki Omori

Keywords: Flood discharge; Rainfall duration; Climate change; Extreme value statistics

Abstract: Unregulated peak discharge, which is the standard for river management in Japan, has been determined by statistical processing of future flood forecasts. However, it is difficult to deal directly with the changing discharge statistically because it includes various factors, such as landform characteristics, climatic conditions, and artificial systems. Therefore, rainfall, which contributes the most to runoff in a watershed, has been statistically processed, and flood discharge has been predicted by flow simulation using this rainfall. On the other hand, this method does not take into account factors other than rainfall in predicting the future river flow, and therefore has not been able to extend discussions on the response of river flow to rainfall. In addition, it is necessary to take into account climate change due to global warming when predicting future flood discharges. In this study, the correlation between flood discharge and future rainfall is researched and evaluated by simulating the river flow condition with consideration of climate change using the large ensemble climate prediction data (d4PDF). The Yamato River system, a first-class river that originates in Nara Prefecture and flows through the Osaka Plain in Japan, was selected for the study. For the simulation, we used a distributed runoff model that has been modified to take into account intake and wastewater by the artificial system and water supply from other basins (the Kinokawa River basin). We used 5,400 years of the +4K climate simulation data (60 years x 90 ensembles) and the Historical climate simulation data (30 years x 50 ensembles) from d4PDF as input values to simulate the long-term flow conditions of the Yamato River. To evaluate the extreme value statistics for flood discharge, the maximum temporal discharge values for each 5-day interval were compiled from the flow simulation results, and the top five annual data were extracted from them. In addition, the basin-averaged total rainfall for the 18-hour rainfall duration in the Yamato River basin was extracted from the time of the maximum flow. The results that we analyzed the correlation between the 18-hour total rainfall and the flood discharge showed that as the rainfall increased, the discharge also increased in a correlated property with a wide range. Also, outliers, which significantly affect the range of correlation, were found for both rainfall and discharge.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221072

Year: 2022