One of the scariest and most dangerous aspects of major natural disasters like earthquakes and tsunamis is that each event is unique and the full impact and damage can't be predicted ahead of time. While prediction systems are getting better at providing warnings before one occurs, there's still not much indication in the way of how big these events will be.
Researchers at the University of Granada have developed a simulator that not only predicts the occurrence of tsunamis after landslide events but how they will behave. This system can process the simulation within five to 10 minutes, giving authorities and rescue teams valuable information on how to respond before the wave hits.
Since tsunamis vary in the time they take to form, the system gives responders anywhere from 10 minutes to several hours to plan evacuation and rescue operations. The fast processing uses several graphic processing units (GPUs) simultaneously to create the simulations.
To test the system, the team created a simulation of one of the worst landslide-caused tsunamis in history in Lituya Bay, Alaska in 1958. The landslide saw 30 million cubic meters of glacier fall into the bay, which generated a giant wave that flooded areas more than 500 meters above sea level.
These landslide-caused tsunamis are common in mountainous areas by fjords and lakes as well as mountainous islands. Underwater landslides can also trigger the large waves. Although these events tend to have a smaller area of impact than earthquake-caused tsunamis, they actually have a greater amplitude.
The simulation of these types of tsunamis can help scientists to learn more about other larger tsunamis as well because the physics are largely the same and earthquakes and landslides often occur together. The researchers also see this technology being useful for predicting floods from river overflows, coastal storms and hurricanes.