Russian Scientists Develop Algorithms to Check Tsunami Sensors
Special instruments on the floor of the Pacific Ocean measure water pressure and movements of the seafloor. When an underwater earthquake or landslide occurs, it can trigger a tsunami. The sensors detect changes and transmit signals to shore. Their readings help determine whether a tsunami will form, when it will arrive, and how high the waves will be.
But the harsh underwater environment can damage the instruments, causing them to transmit distorted readings. A faulty sensor might trigger a false alarm or fail to detect a real threat. Researchers at the Department of Marine and Inland Water Physics at the Faculty of Physics of Lomonosov Moscow State University have proposed a way to identify failures without having to retrieve the sensors. Instead, they use the data itself to detect problems: by analyzing the readings transmitted by the stations, the system can determine when an instrument is malfunctioning.
The Results Are in the Data
The researchers developed four different ways to check suspicious readings. The first compares them with tidal patterns. Ocean levels rise and fall on a predictable schedule as tides come in and go out. The system subtracts the expected tidal changes from the sensor’s actual readings. For a properly functioning instrument, the remaining signal should be small and explainable. The second method compares readings from neighboring sensors. A real tsunami or earthquake will usually be detected by several stations at once. Another algorithm attempts to reconstruct the final quarter of a time series using the first three quarters as input. Faulty instruments generally produce much larger errors in this forecast. At the final stage, a machine-learning algorithm combines the results of all four checks and identifies the stations that need attention from specialists. Each method can produce false results on its own, but combining them makes errors much less common. The researchers report an accuracy of about 90% in detecting faulty instruments.
The technology has already been tested on Japan’s S-net offshore observation network along the country’s eastern coast. The researchers fed two weeks of data from 150 stations into the algorithm. It identified 15 faulty instruments. One station actually failed during the observation period, and the algorithm detected the failure as it happened.
Thinking Several Moves Ahead
Tsunamis are among the most destructive natural hazards on Earth. Over the past century, they have claimed more than 260,000 lives. Researchers around the world continue to develop technologies that can provide warning of these disasters and help save lives.
Yuri Korolev, a researcher at the Institute of Marine Geology and Geophysics of the Far Eastern Branch of the Russian Academy of Sciences, has studied the potential for short-term tsunami forecasting in the Kuril Islands. He believes that measurements of ocean levels at one location, such as a deep-ocean monitoring station, can be used to calculate the shape of a wave at another location, including near the shore. Korolev’s rapid forecasting method requires reliable, real-time data from deep-ocean stations. The Moscow State University system, which monitors the condition of measuring equipment, and Korolev’s method, which uses those measurements to forecast tsunamis, therefore complement each other. In the future, the two technologies could form the basis of an upgraded Russian tsunami warning system.
Earlier, researchers at the Institute of Automation and Electrometry and the Institute of Computational Mathematics and Mathematical Geophysics, both part of the Siberian Branch of the Russian Academy of Sciences, developed a device that can be connected to an ordinary computer to simulate tsunami wave propagation in just minutes. The system is particularly relevant to Sakhalin, Kamchatka, and Japan, where a tsunami can reach the coast 25 to 30 minutes after an earthquake.
Meanwhile, Sirius University of Science and Technology has announced the development and testing of a multisensor geodynamic monitoring system. Researchers deployed a network comprising 10 seismic monitoring stations and 23 high-precision satellite geodesy stations. The equipment detects even the slightest movements of the Earth's crust and shifts in the surface. Data from the network are sent to a central analytics center for automated processing. The system is designed to assess geological hazards along the Black Sea coast, including the risks of earthquakes, landslides, and tsunamis.
The Power of Digital Technology
Russia has a national tsunami warning system that primarily covers the country's Far East. Reliable measurements are especially important for Kamchatka, the Kuril Islands, Sakhalin, and Primorye. Digital technologies could make tsunami warnings more reliable and, in turn, improve safety for people living in coastal communities. By quickly identifying faulty sensors, the system would allow specialists to exclude questionable readings from their calculations and make decisions sooner. That matters when organizing evacuations, coordinating emergency response teams, and protecting infrastructure.
The developers could eventually work with Roshydromet, Russia’s Federal Service for Hydrometeorology and Environmental Monitoring, and the Russian Academy of Sciences’ Geophysical Survey. Similar technologies could be used to monitor not only oceanographic sensors but also other distributed measurement networks, including seismic and meteorological systems and industrial monitoring equipment. One thing is clear: Russian science is moving beyond the development of individual tools for observing natural hazards toward intelligent systems that can assess the quality of incoming data on their own.
