A research team of the University of Tokyo developed a system to estimate expected rainfall more precisely. By analyzing big data, the system can estimate rainfall over the next 15 hours inside Japan. The researchers reckon that it can prevent a flood through appropriate discharge of water from the dam in the case of typhoon and localized torrential rain. They plan to start the substantiative experiment within the year and put the system into practical use in 1-2 years.
The system uses meteorological data from around the world that include images from radars and meteorological satellites. The volume of data it collects automatically is equivalent to more than 250,000 DVDs. It can estimate worldwide weather and rainfall in points around the world precisely. In addition, it can estimate the change of flow volume of rivers using such information as rainfall, water capacity, and geography for the control of multiple dams in the watershed. The researchers tried the new system by reproducing the flood caused by a river in the Kanto district in 2002 and found that the flood could have been prevented if discharge had started earlier because the new system could have been able to decrease the water volume of the river at peak hour by 10%.
Integration and merger of global observation and predicted data