LANDSLIDE TSUNAMIS, Summaries of Mechanics

Download LANDSLIDE TSUNAMIS and more Summaries Mechanics in PDF only on Docsity!

LECTURE 7

LANDSLIDE TSUNAMIS

• An important aspect of an Earthquake Rupture is that the walls of the fault remain cohesive continuous media out- side of the dislocation surface. In particular, the continu- ity of the structure is preserved near the ends (tips) of the fault.

Contrast this with the case of a Slump or Landslide.

[Mathematically, this is expressed through different boundary conditions for the analytical representations of the source].

THE PNG PUZZLE

1. LOCAL RUN-UP AMPLITUDE TOO LARGE

RELATIVE TO EARTHQUAKE SIZE

Local run-up amplitude is consistently 10 m, with a

peak at 15 m.

It cannot be reconciled with the size of the earth-

quake, and in particular with its fault length, with-

out leading to strains in excess of the strength of

crustal rocks.

[ Synolakis et al., 2002]

THE PNG PUZZLE (continued)

2. THE LARGE LOCAL RUN-UP AMPLITUDES ARE

CONCENTRATED ALONG TOO SHORT A SECTION

OF COAST (at most 30 km).

- Contrast with the run-up distribution for the 1992

Nicaragua tsunami

The aspect ratio of the run-up distribution cannot be pre-

dicted by dislocation models based on continuum

mechanics — they would require a strain release greater

than the yield strain of rock.

THE PNG PUZZLE (continued)

4.

THE TSUNAMI IS

ABOUT 10 minutes

LATE !!

Comprehensive interviews by

Davies

[1998] indicate that:

• In some areas (Malol), the tsunami

did not arrive until after the "second

felt shock"

(main aftershock at 09:09 GMT);

• In other areas (Arop, Warapu), the tsunami arrived before the population

had a chance to feel the main aftershock.

This essentially rules out the mainshock as a plausible source of thetsunami, and requires that its source take place^ Some time between the mainshock (08:49) and the main aftershock (09:09).

WAKE ISLAND

HYDROACOUSTIC RECORD -- 17 JULY 1998

ANOMALOUS EVENT(Duration; High Frequencies)

Interpreted as

UNDERWATER SLUMP

08:

MAINSHOCK

DOUBLET AFTERSHOCK

09:09 –09:10 MAIN

09:

09:25 GMT

09:53:11 GMT

Time after 09:15 GMT (hundreds of seconds)

• In short, the event at 09:02 is

TOO WEAK FOR ITS DURATION

or

TOO LONG FOR ITS AMPLITUDE

→→ In other words, it

VIOLATES SCALING LAWS

which suggests that it must represent a different physical phenomenon.

IT IS THERE !!!

TSUNAMI SIMULATIONS — SLUMP SOURCE

[ Heinrich et al., 2000]

t = 90 s

t = 360 s

Vertical exaggeration: 750

PERSPECTIVE on LANDSLIDE TSUNAMIS

• As compared to earthquakes,

Landslides move SMALLER AMOUNTS of material over MUCH LARGER DISTANCES.

• Therefore, their tsunamis have

MUCH LARGER AMPLITUDES

MUCH SHORTER WAVELENGTHS

→ Hence, they will be MORE EFFICIENTLY DISPERSED during propagation.

• They may also become intrinsically unstable and BREAK (like surf) rather than propagate.

As a result, LANDSLIDE tsunamis are DEVASTATING locally, but pose

LITTLE HAZARD in the FAR FIELD.

Estimated Volume of Rock Slide: 4 million m

3

1999 FATU−HIVA TSUNAMI:

The SOURCE

MODELING the 1999 FATU-HIVA LANDSLIDE

Note the refraction of the wave around the promon-

tory and into Omoa Bay (Plates D, E, F).

Note also the arrival of two principal waves.

[ Okal et al., 2002]

LABORATORY MODELING of LITUYA BAY LANDSLIDE & TSUNAMI [ Fritz et al., 2001]

Conclusion: Exceptional run-up well reproduced in laboratory experiment.

Importance of large air cavity developing during impact of landslide.

Maximum splash on opposite hill:

524 meters

VOLCANIC LANDSLIDES at La Sciara, STROMBOLI

(Italy) — 30 DECEMBER 2002

Run-up reached 10 m in nearby village Miraculously, no victims [ La Rocca et al.,

2004]