New real-time flood projection system developed by joint Waseda and University of Tokyo team to be tested in Tokyo

  • Artificial Intelligence

New real-time flood projection system developed by joint Waseda and University of Tokyo team to be tested in Tokyo

A team led by Waseda University and University of Tokyo researchers has developed a flood projection computer system that they claim is able to provide more accurate real-time estimates than existing ones.

The new system will initially be launched for Tokyo’s 23 wards on a trial basis, and the team hopes full-fledged prediction service will start by the beginning of the Tokyo Olympics next year.

 

The new system would help ease flood damage, the team said.

Risks of heavy rain on an unprecedented scale have grown recently due to global climate change, the team said. The system, called S-uiPS (Sekine’s urban inundation Prediction System), predicts floods based on detailed urban infrastructure data, including how the sewage system and streets are structured and connected to rivers. Additionally, it uses rainfall monitoring and projection data from the transport ministry and the Meteorological Agency to predict floods in real time, the team said.

The new projection system, which shows constantly updated flood map on the screen, can help municipalities and people make critical decisions, the team said.

The map shows the result of a simulation with five different colors indicating the projected level of flooding.

“I can only think of two places where flooding might kill people: underground spaces and underpasses,” said Masato Sekine, professor at Waseda University and an expert in hydraulics and river engineering. Sekine developed the core components of the program.

When it rains heavily, underground spaces have a high risk of being inundated, and drivers can become stuck in flooded railroad or highway underpasses. Thus, judging when the appropriate time to seal underground spaces with watertight doors and to block such underpasses is important, Sekine said.

Because S-uiPS maps out a detailed prediction for 30 minutes into the future, it will provide crucial information for the decision-making process, said Sekine.

It takes 10 minutes for the program to reflect the necessary calculations, so people will see a simulation of the situation 20 minutes ahead of time.

There have been other flood prediction programs, but those programs lack comprehensive inputs regarding how rainwater flow would be affected by urban infrastructure and rivers. They also often apply simplified calculation models.

But in order to achieve greater accuracy, the S-uiPS does not simplify calculations, said Sekine, adding that researchers from the University of Tokyo have helped realize the fast processing needed for the real-time projections.

Sekine said a massive flood disaster in 2000 that hit the Tokai region, which includes the city of Nagoya, prompted him to pursue the project.

“If the same kind of disaster happens in Tokyo, it will be problematic. … I started this because I wanted to somehow contribute to” lessening the risks, said Sekine.

The trial of S-uiPS in Tokyo will kick off by the end of June.

If local urban infrastructure data is available, the new method can be applied to other cities, the team said.

Publish news on
Unpublish news on
Publication Date
Tue, 05/21/2019 - 00:00