In the eastern Libyan city of Derna, heavy rains cause mass flooding, displacing thousands. In Greece, record rainfall impacts the islands of Crete and Evia. Rio Grande do Sul in Brazil sees sweeping waters touching the lives of over 100,000 people. Japan reels from pummeling rainfall that results in heavy flooding and mudslides. Here at home, more of the same… with the east coast currently bracing to take more thanks to Hurricane Lee.
This is life now… we know that thanks to sophisticated instruments that belong to NOAA (the National Oceanic and Atmospheric Administration). This scientific and regulatory agency within the United States Department of Commerce helps us monitor our climate. They analyzed data from 98 different tide gauges along the U.S. coast, and here’s what they found: thanks to sea level rise, it no longer takes severe weather to cause disruptive flooding along the coast.
Historically, floods are tricky for humans. We haven’t really had very sophisticated ways to predict or measure floods. One tried-and-true method is called flood gauging, and it’s basically a stick in the ground. You mark the stick at a given water level, then mark it again when the water level moves… it’s not that much different from measuring the grandkid’s growth progress on the doorframe over the years.
We’ve also relied on observing animal behavior, since we can count on animals to instinctively do certain things like fly to higher ground or abandon their burrows when floods are on the way. Now… we clearly have more advanced flood prediction and measuring capabilities today, but we’re still lagging on progress thanks to the very nature of floods. Namely, they are extremely complex, which makes them difficult to predict.
So we’ve had these huge, complex creatures (floods) that we don’t quite understand, and very little way to gain any sense of control… until recently. Thanks to our up-and-coming super computing powers and our artificial intelligence boom, we have new goodies to help us do the hard work of saving folks from floods. For instance, in Nova Scotia, where LiDAR helps predict flooding days in advance.
Unfortunately, we’re seeing America’s aging flood control infrastructure struggle in real-time. That won’t do… not with flooding being the costliest type of natural disaster in the U.S. It’s already responsible for about 90% of the damage we see from natural disasters each year. In 2021, The American Society of Civil Engineers gave the nation’s dams, levees, and stormwater infrastructure a grade of D. If America had a mom, we’d be grounded…
If there’s one thing the next wave of flood prediction tools will have in common, it’s big data. Right now, big data is being fed into AI by researchers at Texas A&M, where they are learning to use what’s called predictive risk analysis to hone the next generation of flood predicting tools. Predictive risk analysis pairs historical data (what has already happened) with statistical models (probabilities) to let us know if something is likely to happen in the future or not.
To imagine this in the wild, we simply need to venture to where risk is at play… insurance. The technology is buried in practicality, but the example holds. To understand it, we need to know about spatial finance, which uses AI to analyze geospatial data in risk assessment and management. NVIDIA offers AI tools for use powering spatial finance applications, and insurance company Lemonade uses spatial finance to price policies.
Speaking of… NVIDIA powers Luxembourg-based RSS-Hydro’s FloodSENS 3D animations. FloodSENS is a machine learning app that maps flood impact from satellite images… it all sounds very advanced. Could we one day be able to fully simulate weather patterns across the globe? Maybe… a small group of researchers plan to build a digital twin of earth to do just that. Come back next week, we’ll have more from the green tech space.