Managing the Data Flood to Prevent Flooding

RSS-Hydro CEO, Guy Schumann on his company’s work to predict and map flooding around the globe


Where did you get the idea to create RSS-Hydro?

I am a research scientist, but I also wear an academic hat. I’m affiliated with a couple of universities around the world - The University of Bristol in England and University of Colorado Boulder, in the US.

After I finished at Bristol, I had the opportunity to go to NASA JPL, in Southern California. There I worked on NASA’s first dedicated hydrology mission, in collaboration with CNES, the French space agency.

This gave me the idea to start research and development of Earth Observation for flood applications in the private sector in Luxembourg.

So, as a business, RSS-Hydro has three main pillars:

One is research into the use of satellite data from programs like Copernicus to innovate flood mapping. The second pillar is more traditional. We develop flood prediction and flood mapping products and services. Our third business pillar is expert consultancy.

How has the use of satellite data by ESA and NASA for Earth observation to study water, or hydro dynamics evolved?

The US accelerated the use of data during hurricane Katrina in 2005, using their Landsat satellites and others to look at the extent of flooding in New Orleans. And for Europe, it was the Elbe floods in Germany in 2002 and, at around the same time in Luxembourg it was the Alzette floods of 2003 that I worked on during my PhD.

That's also when I saw the first big wave of scientific papers coming out, saying we need to use remote sensing from space to look at floods.

Do you see an increased focus on global warming in your work?

Yes. Because of growing populations, ageing populations, and populations moving to urban centers, there are automatically more assets at risk.

Even if the flood hazards don't change, the flood risk will change because there are more people at risk and there are more assets exposed.

So for insurance, climate change is definitely becoming a  serious concern.

Luxembourg is a landlocked country, so are the majority of your projects international in flavour?

Yeah, we work on a lot of international projects. For example, we had quite a big project with the German Institute for Evaluation of Development Projects.

They evaluate projects supported by German government development funds. So, we had a project in Morocco, where we had to come up with a methodology to derive climate change impacts on the hydrology and local flooding, covering the entire country.

We have other projects in evaluation with ESA too.

How do you go about evaluating flood risk?

As part of an Industrial Fellowship between the University of Luxembourg and RSS-Hydro, Mohammad Zare is working on improving the simulation and prediction of flash floods. The goal is to develop a decision-making model for flood protection in Luxembourg.

As Mohammad points out, “Flooding is the number one natural disaster in terms of insured and uninsured losses, on an annual basis.”

Developing reliable simulations has helped us to predict floods and so reduce damages and loss of life. But, Mohammad is right when he says there’s still a lot of work to do, particularly in the area of flash floods in urban areas.

Guy Schumann interview

What kind of improvements could be made?

One area that could be improved is map resolution, which always uses a 10 by 10 meter pixel. This makes it very hard to observe some local flooding. And below cloud cover, it’s impossible.

But we think that, with machine learning, if the entire image is not cloud covered, there should be ways of reconstructing events. This would be the Holy Grail of flood mapping.

The idea is that we use Sentinel satellite images from the LSA data center, to train machine learning. Then we try to reconstruct the flooding seen below the cloud cover.

Mohammad does a great job of explaining how this would work in practice: “The proposed methodology links flood hazard modelling, remote sensing and machine learning methods. Combining these physical models with data-driven methods will result in a more reliable hybrid model that can be employed for prediction of (flash) floods and event analysis,”

The final approach will be to cross-evaluate Light Detection And Radar (LiDAR) topography with available super-resolution drone data, to assess the ability to incorporate local flood defenses into the models.

How does the LSA data center make your work more efficient?

Well, to be honest, first of all, it’s great because it's free for commercial entities like ourselves.

So, in the past could you contact ESA directly, to get the data?

Yes, but it would be a much lengthier process. And, historically ESA data was always free for science and research and public entities, but not for commercial.

So the Copernicus Sentinel program from the European Commission was the first one actually to push it towards something that the US always had by constitution - That publicly funded data is free for everybody.

Sentinel is a game changer.

So the LSA data center, simplifies things?

Here in Luxembourg you cut the paperwork, for sure. Also, there's much higher connectivity speeds these days. The LSA data center offers the fastest connectivity to download data. So if you don't want to work with a cloud system, but you would like to have the data downloaded for testing and calibration, which is what we do... It's really good to have to LSA data center,

Is it easy to find what you need?

Yes, it's very easy to find, it's very intuitive. You don't need to be an expert to search through the catalogue. The other thing is that the data is streamed from ESA to the LSA data center instantly. So it's almost like a satellite ground station.

So how recent is the latest image that you can get?

From live satellite acquisition, I would say less than an hour. Definitely.

There must be a hell of a lot of data to sift through

Yes. And that's the problem! We often joke about the flood of data, right? Well, here, the data flood can cause real problems.

There can be so much data that people get confused. They don’t know what to do with it, or how to handle that data. And that’s without data from commercial satellite missions!

So, can you say how safe Luxembourg is today?

We are still very lucky with the natural disasters in Luxembourg, but that's not to say that we don't get them.

Only a year ago, we had a tornado and Germany has tornadoes every year. So we are not resting on our laurels! 

Last update