On a Sunday in 2022, Mikhail Nosov was out of town with a bad internet connection, and his phone kept buzzing. A colleague was sending text after text: a tsunami somewhere had strange properties. It did not fade as it moved away from its source. It got stronger.
Nosov is an oceanologist who studies tsunamis, and he could not check anything until he got back to Moscow. What he found there was the eruption of Hunga Tonga–Hunga Ha'apai, and it had broken a rule that tsunami forecasters lean on every day.
That rule is simple. You know where the wave started, you know how fast waves run across the ocean, so you know when it arrives. A warning system built on the Tonga volcano as the starting point would have missed the wave by several hours. The wave got there first.
The reason is a second wave, made of air. When the volcano blew, a shock front of air raced out along the surface of the ocean at about 312 metres per second over the Pacific, roughly 1,100 kilometres an hour. A tsunami in the deep ocean trenches moves at nearly that same speed.
Think of pushing a child on a swing. Push at random and you get nothing much. Push in step with the swing and every shove adds to the last. The air wave and the water wave were in step. As the shock front crossed the ocean, it kept shoving the water at the pace the water likes to move, and in some regions it built gravity waves very efficiently.
So the tsunami did not travel only from the volcano. It was being made all along the way, ahead of the water that had left the crater. The source, in Nosov's words, turned out to be distributed, and it ran ahead of the tsunami by hours.
A study in Nature in 2022, led by R. Omira, put numbers on this. It compared the observed tsunami with a hypothetical one set off at a single point. The real one travelled 1.5 to 2.5 times faster, and it crossed the Pacific, the Atlantic and the Indian oceans in under 20 hours.
Then came the part Nosov calls paradoxical at first glance. His group worked with data from the DART stations in the open Pacific and the DONET stations near Japan, and with models. Pressure sensors on the ocean floor recorded the passing pressure wave, and so did barographs in the air above.
That is, it turned out that the pressure oscillations recorded on the ocean floor have a larger amplitude than the pressure oscillations recorded in the atmosphere.
Оказалось, что колебания давления, которые регистрируются на дне океана, они имеют большую амплитуду, чем колебания давления, которые регистрируются в атмосфере.
— Mikhail Nosov, Osnova, 75:46
About twice as large. A water layer several kilometres thick should shield the bottom from whatever happens up top. Near resonance, it did the opposite. This holds in the measurements and in the model alike.
Compare it with Krakatau in 1883. The setting was similar: a volcano blew, and the crater was left to fill with water. But the instruments were pen barographs scratching pressure onto rotating drums and tide gauges drawing the sea level. They still caught waves as far away as the Atlantic.
In 2022 the whole planet was listening. Satellites watched it. Barographs on land and sensors on the seabed recorded it. Plenty of people had a barometer in their phone, and the air wave went around the Earth more than once. An ordinary household gadget could pick it up. Above the noise, instruments registered it four times; for Krakatau, Nosov says, the figure is reportedly seven. In satellite pictures you could see the shock front spreading in a ring, with the ash cloud rising at its centre.
What would you have to change in a warning system after that? Nosov does not claim a fix. His point is plainer: a volcano can be a moving source, one that travels with the wave. Counting from the crater alone, the clocks would have been wrong by hours.
