# How accurately can a tsunami be predicted? Have there been any successful evacuations?

With the current instruments in hand, how accurately can we predict tsunamis? Please consider me as a novice with only basic information how tsunamis are formed.

Has there ever been a successful evacuation in history in regards to incoming death waves?

## 1 Answer

The ability to predict depends on the cause of the tsunami. Earthquakes can't be predicted, so a tsunami caused by an earthquake can't be predicted. A meteor can be tracked for a considerable time in space before sticking earth so a tsunami caused by an meteor strike can be predicted.

There is also the possibility of a tsunami early warning system. There are currently sensors in some parts of the ocean which can detect a tsunami before "landfall" and provide an early warning system. Think of it somewhat like tracking a hurricane which generally takes a lot longer to reach land.

In a webpage it was reported that "in 1933 HVO [Hawaiian Volcano Observatory] used seismograms to predict (accurately) the arrival of a tsunami from a distant earthquake, and people in low-lying areas of Hawaii were for the first time successfully warned in advance." The word predict here is somewhat ambiguous. the scientists didn't know the earthquake would happen, but recognized that the earthquake would likely cause a tsunami.

Looking at the map below. Where tectonic plates connect, or where there are underwater volcanoes are prime locations to generate tsunamis.

• National Geographic has an episode in its Megacity series about the warning systems implemented for Mexico City. They have only minutes, but every minute (+ exercises) can save lives. Video here Jan 2 '16 at 15:55
• @JanDoggen The sensors in the see must be placed in the region where there is a high relative probability of an earth quake under the sea, right? Is it not possible to detect such areas geographically? I suppose there must be some statistical data on the likely hood of earthquakes under the sea. Putting all this information together, can we provide an estimate of the relative probability that a region will hit a tsunami as a function of the geometrical coordinates and time? Correct me if I am wrong? Jan 2 '16 at 16:37
• @Sathyam - The idea is to have two different types of sensor networks. One which detects earthquakes, and another which detects large changes in water pressure. The water sensors don't detect "probabilities" they are essentially measuring the size of the incoming tsunami. The idea is to have them far enough out to sea so that an evacuation alarm can be sounded in enough time to be useful.
– MaxW
Jan 2 '16 at 16:58
• @MaxW That is obviously logical. Is there any other variable other than the distance from the shore. If a region has a bad reputation of causing underwater earthquakes, it's plausible to keep a sensor there right. I was thinking in terms of what if the earthquake struck in a region halfway between the sensor and the shore. Jan 2 '16 at 17:30
• @Sathyam - Of course the idea is to make a network of underwater sensors so that the tsunami can be detected as early as possible before hitting land anywhere.
– MaxW
Jan 2 '16 at 17:43