Predicting tsunamis, what to measure from earthquake data?

  • Thread starter panxi
  • Start date
I have a project to create a program to predict tsunami possibility by gathering data and finding pattern from previous earthquakes. It would be using neural network. But I don't really know what I should consider from the data. Is it the magnitude that cause the tsunami? Or is it the depth level? Any other measurable data that I should consider?

Are there any specific textbook that cover this?

Btw, does anyone know any dataset that have both earthquake and tsunami occurences from the last 100 years?
 

jim mcnamara

Mentor
3,439
1,636
Based on your questions it is hard to know where to start. Part of the problem is that a lot of relevant research is behind a paywall as well.

Suggestions:
It might be better to ask a researcher in the field you want (email that person).
Try: Chau An ---- use the email address (icon next to his name) in this abstract link (useful part of the paper is behind a paywall):


See the video referenced in this post on PF to start thinking abut your project:
 

Astronuc

Staff Emeritus
Science Advisor
18,491
1,612
Magnitude, depth, location (latitude and longitude), fault geometry and type, are key elements/variables.


Earthquakes on land (away from coast) will not cause tsunamis. Earthquakes near the coast or underwater may cause tsunamis if a lot of a water is displaced.

Most tsunamis are generated by shallow, great earthquakes at subductions zones. More than 80% of the world's tsunamis occur in the Pacific along its Ring of Fire subduction zones.
 
Magnitude, depth, location (latitude and longitude), fault geometry and type, are key elements/variables.


Earthquakes on land (away from coast) will not cause tsunamis. Earthquakes near the coast or underwater may cause tsunamis if a lot of a water is displaced.


When you said location is one of the key element, what does it mean? How does the location of earthquake epicenter affect the tsunami occurence? Say, there's an earthquake occurs in Pacific Ocean and Indian Ocean with same magnitude, depth, fault & type. Is the probability of the earthquake in Pacific different with that in Indian Ocean?
 

Astronuc

Staff Emeritus
Science Advisor
18,491
1,612
When you said location is one of the key element, what does it mean? How does the location of earthquake epicenter affect the tsunami occurence? Say, there's an earthquake occurs in Pacific Ocean and Indian Ocean with same magnitude, depth, fault & type. Is the probability of the earthquake in Pacific different with that in Indian Ocean?
It would be difficult to find two earthquakes nearly identical in all parameters considered, but one may start by comparing the characteristics of two earthquakes of similar magnitude, one in the Pacific Ocean and the other in the Indian Ocean.



Both earthquakes generated large tsunamis.

One could compare the tsunami of the lesser magnitude Chilean earthquakes in 2010 (Mag 8.8) and 2015 (Mag 8.3).

and the earthquake near Sumatra in 2012, which did not generate any significant tsunamis.

Consider the type of fault, magnitude and depth.

Also, one could research megathrust earthquakes starting with the most recent and working back in time.
 

davenn

Science Advisor
Gold Member
8,572
5,279
When you said location is one of the key element, what does it mean? How does the location of earthquake epicenter affect the tsunami occurence?
Astro answered that in the text you quoted from him, did you not read it all ? 😉

Earthquakes on land (away from coast) will not cause tsunamis. Earthquakes near the coast or underwater may cause tsunamis if a lot of a water is displaced.

That's the basic importance as to whether a quake will produce a tsunami or not.
After that, ( as has been stated) it is the type of rupture.
Thrust earthquakes will produce tsunamis. The size of the tsunami is then related to the size of the rupture,
that is, the length and offset on the fault.

Strike-slip ruptures underwater are not likely to produce tsunamis as there is no or little vertical offset.

The problem is, faults are rarely pure thrust or strike-slip. Most times even the best thrust fault event will have some strike-slip motion and even the best strike-slip fault event is likely to have a little thrust movement.
And there is every variation imagineable between those two extremes.

Here's a fault I visited recently in the NE of the South Island of New Zealand. This is one of a number of faults that were both on-land ( mostly) and some were coastal that the combined rupturing produced a Mw7.8 event in November 2016.
This fault is partially on-land and runs offshore. It's mapped length is around 3km
The greatest offset observed is 2m vertical and 6m horizontal. It uplifted some 10km of coastline that once used to be submerged at high tide but now is well above high water level.

my photos
looking along a long section of the rupture
View is NW, uplifted side is seaward side. Around 1 - 1.5 metres beside me
IMG_9649sm.jpg


further NW along (~100m) the fault and at the largest vertical offset, ~ 2m
Sadly the dumping of rock from the various landslips has almost covered the fault
IMG_9652sm.jpg


Google Earth aerial view
Kaikoura faulting1.jpg


Kaikoura faulting2.JPG


This uplift produced a small tsunami along the coast of around 1m



Dave
 
650
164
There is also the possibility of an earthquake-destabilised slope-failure causing a tsunami, which seems to be what happened recently in Indonesia.

If you count the 'slosh' from massive falls into lakes as a tsunami, you have the double-jeopardy that one quake may create a rock/ice dam forming a lake, then an after-shock drops a further zillion tons into that. Slosh swamps shore-line, may catastrophically over-top and erode dam, empty lake into valley below. Potential 'domino' effects...
{ Brrr... }
 

davenn

Science Advisor
Gold Member
8,572
5,279
There is also the possibility of an earthquake-destabilised slope-failure causing a tsunami, which seems to be what happened recently in Indonesia.
really ? which one was that ? ...
there was a very recent tsunami in Indonesia related to slope failure of the
Anak Krakatau volcano in the Sunda Strait but it wasn't earthquake initiated.

The only case of that in recent years was ~ 23 years ago on the north coast on New Guinea, PNG
where a ~M7.1 produced a 10-12m tsunami when it really shouldn't have. Further investigation found
that the quake has caused a significant undersea landslide on the shelf.
 

davenn

Science Advisor
Gold Member
8,572
5,279
Last edited:
650
164
With respect, video footage suggests the tsunami arose at a right-angle to the fault line rather than usual parallel. Also, it seemed to start after significant landslides on the flanks of the bay hit the water.
Such is but correlation, though, not causation...
YMMV.
--
"substantial land movement"
Agreed, any 'made ground' would be susceptible to subsidence and liquefaction, and soft ground would be especially vulnerable, per Mexico City's experience.
 

davenn

Science Advisor
Gold Member
8,572
5,279
With respect, video footage suggests the tsunami arose at a right-angle to the fault line rather than usual parallel. Also, it seemed to start after significant landslides on the flanks of the bay hit the water.
Such is but correlation, though, not causation...
YMMV.
don't know what video you were referring to ... 2 of the 3 videos in that link don't work

that doesn't negate the possibility of it being caused by strike slip motion given the right seafloor topography

Also, it seemed to start after significant landslides on the flanks of the bay hit the water.
what landslides ? again if there was actual video of that, in one of those 2 missing videos, it's no longer available
for me to view and make an assessment
 

Tom.G

Science Advisor
2,511
1,345
over 800 000 finds:

this from the United States Geological Service (USGS.GOV):

On July 9, 1958, a magnitude 7.9 earthquake on the Fairweather Fault triggered a rock avalanche at the head of Lituya Bay, Alaska. The landslide generated a wave that ran up 524 meters (1,719 feet) on the opposite shore and sent a 30-meter-high wave through Lituya Bay, sinking two fishing boats and killing two people.
 

davenn

Science Advisor
Gold Member
8,572
5,279
over 800 000 finds:
tsunami caused by landslide - Google Search
this from the United States Geological Service (USGS.GOV):
On July 9, 1958, a magnitude 7.9 earthquake on the Fairweather Fault triggered a rock avalanche at the head of Lituya Bay, Alaska. The landslide generated a wave that ran up 524 meters (1,719 feet) on the opposite shore and sent a 30-meter-high wave through Lituya Bay, sinking two fishing boats and killing two people.

Ohh dear

Tom,
I am not arguing against that... I know full well quakes can caused landslide induced tsunamis
I have been at this game for 40 years. I even stated a quake, landslide induced tsunami in an earlier post
:wink:


Dave
 

Want to reply to this thread?

"Predicting tsunamis, what to measure from earthquake data?" You must log in or register to reply here.

Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving

Hot Threads

Top