Technologies and applications of interactive flood simulation


>> Hi everybody. Thank you for coming to the Tech talk today. I’m pleased to host the, you know, today’s talk–Tech talk. Let me introduce him, his name is Satoshi Yamaguchi from Hitachi Center Research Laboratory. So he’s going to present to us about stay in interactive Flood Simulation system on the Earth Viewer. Please welcome Satoshi. >> YAMAGUCHI: Hi. Thank you for coming my tech talk. And today I want to–I will talk about the Technology and the Application of Flood Simulation on Earth Viewer. Thank you for this time, it’s great opportunity for me and our team to make a presentation here. So at first, let me talk about the motivation of this study. We start the–we started the study because floods are serious disaster now.

And you know, cyclone nowadays attacked Myanmar, Burma last two weeks ago and about that is for–several thousands of people are killed by the cyclone but a lot of us travel all over the world. According at the Red Cross 93,000 people are killed in this decade. And economic loss is also huge $20 billion lost in Japan only in 2004. And in addition, higher chance of flooding is expected in the future because of climate change. According to IPCC, there is a high confidential–a high confidence in the increase of extreme weather events. It means droughts, heat waves and floods. Increase of temperature–increase in temperature cause increase of humidity and increase of precipitation. But the increase of–the change [INDISTINCT] in temporary and stationary evening. I mean, extreme weather will be increased. So last week, I attend the International Conference on flood defense.

At the conference, flood–climate change is very hot issue, almost all researcher forecasted on climate change. Flood will be more serious problem. What should I do? I think simulation will be–will be a good source to mitigate flood disaster. To mitigate the flood disaster, we need preparation and evacuation. Before the flood, we should prepare for the flood. And in the case of the flood, we should evacuate but to act correctly–accurately, we need precise information on floods. Simulation is a good term to generate such information. And I believe to mitigate–to build an awareness of flood risk–flood risk convenient for general public is needed. That’s why we create our product. The name is DioVISTA and the DioVISTA makes simulation easy to use. DioVISTA makes simulation easy to understand. Too easy too use–we make automatic settings and functions. And to make simulation easy to understand, we provide an interactive–intuitive demographic user interface. I will show you some of our application. This is our product DioVISTA.

You can see the globe and you can see anywhere in the world. But it’s not only a view but also a simulator. Now we are going into a city in Japan. And this is a Fukui city and 2004–in 2004, there’s a huge water hazard happened. You can see red arrows here. This is already filled up point. So from this red section, water has spread into the city. Now I will simulate the situation on this in my laptop computer. I start–I adjust–okay. I start–I click start button then I click start button again. Now this stimulator is calculating the flood floor on the city. Color indicator water left so you can understand how complex water flow is. There’s a database–database on the bank here. So water stops here and there’s a big street and water flows around the street. And there’s a car bus or underpass here.

So water flows through the underpass and go another area like this. Urban flood in urban area is very complex program. So we need a simulator to assess for the risk and to make evacuation decision. Our simulator can simulate a lot of situation on this computer, like this. Now I add another condition. I break this [INDISTINCT] so water flows is ready, like this. So you can understand how easy our simulator simulate complex phenomena and how fast our simulator is. Okay. Any questions so far? Okay. I will go ahead. I want to show this–a list now but if we collect this data to us database, we can generate this kind of list. In this list, you can see owner’s name, address and water depth. And you can understand the number of damaged house in this simulation. So it’s very convenient to assess damage of the flood.

I divided–I validated the flood result using a comparing site investigation and a simulation. A red line indicate the site investigation and the blue line indicates simulation. A simulated flood area covered 91% of the investigated flooded area. And investigated flooded area covered 87% of the simulated flooded area. It means very accurate to use at practical use. Okay. I will show you another application of the flood simulation. Before previous demo, I showed the [INDISTINCT] simulation but in this demo, I will show you the whole process from rainfall to overflowing, okay. For the simulation we use given data like this. This is a transversal cross section of river and this is a cross section of rivers; upstream is here and downstream is here. And lateral is the distance from the origin point of river. Using this given date and we also use rainfall data which simulate the whole process.

Again I will simulate [INDISTINCT] of the flood. Now this computer calculating the catchment area, catchment area means the area in which rainfall goes to this river. Now you can see the change of the river–water level in the river. As time goes on, time is indicated this year. For the processor is calculating the now. And so I moved back the simulation because we crop this here going to back and now. At this time, [INDISTINCT] of a flowing point. You can see right across here, this indicate overflowing point and along this river, flooded area spread. This is a flood simulation done by us. Okay. Now, I will talk about technology behind that simulator. In the simulator, we use three modules, the simulation modules, mapping modules and user interface module. All the modules in the–in each module, we use different coordinate. So it means, in the each module, we use the different X, Y, Z and T. Because we use four dimensional coordinate. And to integrate for the system, we converted coordinate in the Earth time. So simulation module time is converted to mapping modules time and vice versa. And in the module–simulation module, we use four models.

Runoff model, River model, Levee model, Inundation model. So using these models, we simulate from rainfall and flooded area. This here indicate the schematic view of our simulator. A Runoff model is used for mountainous area. River model is in the river. Levee model is–is in using Levee and the Inundation model is used to in cities. In each models, we use different kind of equations like this. In the Runoff model, we used a so-called Kinematic Wave Equations. I think nobody knows this equation. But the setting of a Runoff model is very complex. Because for each–each slope of mountain, we should generate the network of model. I mean, in this figure, you can see the slopes and the lines. These lines indicate the steeps direction–the fixed direction of a slope. So along this line, water goes down and then finally goes to river. This red line indicate river so all the slopes connected to this river. We need to generate this kind of network to work. And in the River model, we use different kind of simulation equation so-called One Dimentional Unsteady Flow Equations like this.

And from river to city, there’s a levee or a dike or some artificial structures; to simulate this, we use Overflowing Formula like this. And in the city, we use also different kind of equations like this so called Shallow Water Equations. Using this equation, we calculate water depth and water speed for each grid. You can see the grid, so as time goes on, each grid, water depth and the water speed on each grid is calculated like this using ground elevation and ground roughness. But it’s not enough for simulate in other area because in other area, there’s a lot of artificial structures like railways, wide street, culvert etc. I think I show this before so–so I should simulate all the–this kind of artificial structures using other equations, so it’s so difficult simulations. So what we did is automatic–automation over setting of whole process of a simulation use.

The main technology is Dynamic domain definition method or dynamic DDM, this technology, the use for inundation model and the runoff model. Usually, we always need a grid for simulating flood flows. So, we needed to generate grid before simulation like this green indicate the grid. Difficulty of defining–so range indicates domain. Difficulty of defining domain is–flow must not reach boundary because in this situation, we could not simulate anymore because both are the boundary of domain and we cannot simulate this situation because [INDISTINCT] point this is out of domain. So, we cannot simulate this and then we can’t find the root of the dry grid in this area. I mean there’s no–there’s a lot of grid without water. It’s so interesting to calculate. So, what we did is the domain is expanding the shrink domain during the simulation. Like this, we use both the simulation module and the mapping module. At first, we input levee share condition and the mapping module generate domain around the levee share of point. And then, calculate water flows when the grid assume water reach the boundary of the grid–the domain.

So, mapping module generate on the domain. Like this again, I can simulate flood flows that soon reached the boundary so, mapping module add the domain like this. If out of the domain, [INDISTINCT] happened out of the domain, we can generate domain like this. It’s a very sophisticated system. Calculating simulation is very much, much faster than conventional method and the Grid method need care about the–define of the current calculation domain because calculation domain is automatically defined by the–this dynamic DDM. Another application of the dynamic DDM, is setting of runoff model. Runoff–define the runoff domain–runoff model is also the same difficulty. We need to define the calculation domain before analyzing the runoff model. Like this, if I calculate the [INDISTINCT] if I define domain, we can generate the network or runoff model but catchment area must not exist out of the domain and the–you can see I wrote over orange grid here, this is non-catchment area, this should not exist in the domain for faster calculation.

So, I use runoff–dynamic DDM for defining the runoff model. At first, we input levee conditions and then dynamic DDM spread the calculation domain like this so, to a fast range calculation domain automatically generate it. We do not care about the calculation domain. Then I will talk about allocations of our product. Case 1: Application by Local government. This is–the–in this case, grantee is Ohta City, Gunma, Japan. They want risk awareness among public because in the–in the city, there’s a very steep slope so if flood happens, flood flows, speed is very quick so city should be evacuate very quickly. So the desired information is easy to understand the contents that shows how quickly water–flood water spread, so what they did is to provide three dimensional computer graphics movies on the website. I show the website, if you access this site, you can watch the movies and they use a movie to at public meeting about flood mi–mitigation because this content is a very easy to understand. Application two–application by insurance–insurance companies, at this case, client is Phoenix Risk Research Institute Subsidiary of Nissay Dowa General Insurance, they want assess possible water depth, available time and damage of property of their client’s property.

So, they are advised–and they want to advise property owners on countermeasures. So, desired information is quantitative scenarios because they want to provide quantitative advice. Activities they did is stimulate several flood scenarios using this product. This post I make–I made this post–presentation at United Kingdom last year, it’s a poster of the–I make presentation. And the last case is application at technical seminar, this product is used in International Center for Water Hazard and the Risk Management or ICHARM, it’s Japanese government local organization and that to build risk awareness for International communication and help us teach how to map flood disaster–flood hazards so ICHARM want a software that is powerful but can be learned in short time so what ICHARM did is gave a lecture with a product and participants on the lecture came from over 20 countries from China, Philippines, Vietnam, Cambodia, Indonesia, et cetera.

So using this product and then make a hazard map and discuss about the appropriate information. Okay, I write a conclusion briefly, and floods are serious disasters and floods are going to be a serious disaster in the future because of the climate change we need to mitigate disasters and to mitigate disasters we need information, simulation and the mapping technology can contribute a lot to such information, thank you so much for your attention. >> Thank you, Satoshi for the presentation so you know, if anybody in the audience have a question, please don’t hesitate to ask. >> You said, levee [INDISTINCT] manually, is it possible for your system to predict where and when the [INDISTINCT] >> YAMAGUCHI: At present, we could not–we cannot do that because every [INDISTINCT] is very difficult to pre–predict. That’s a–I conduct the research for that so I hope in the future we can predict that automatically. >> Any other question? >> I have a one question, okay? So, it might be idea of the application of this system, I wonder if maybe, you know, in the government or construction company, maybe doesn’t [INDISTINCT] they have some to built kind a large [INDISTINCT] like a building or wide road in some city which might have some risk over the flood.

So, could you–can we use this system to, you know, assess the, you know, possible risk like constructing those kind of wide, you know, large building or, you know, something we can compare current and, you know, the future after we construct that building. >> YAMAGUCHI: I think it can be done; simulation is useful for assessing a possible [INDISTINCT] change for a [INDISTINCT] design so I think simulations–these kinds of simulations could be a very useful to–to do that. >> Is there any other question from the audience? Okay, thank you Satoshi again, you know, for the presentation. >> YAMAGUCHI: Thank you..