In natural soil deposits there is generally no need to assess liquefaction below 15-18 meters if you have shallow foundations. There are some reasons for this:

- if you have 11-12-13 meters of non-liquefied layers the settlement of the liquefiable deep layers would not result in subsidence close to the surface

- the thickness of the holocene sediments (that are usually succeptuble to liquefaction) is rarely above this range). 

- natural deposits at greater depth are usually densified (that is why neogene sediments are not liquafiable)

However in man-made fills, and embankments such as dams, dikes and levees liquefaction is possible if the sandy soils are not well compacted. Also in most upstream tailings dams where the rate of filling is up to 3 meters/year you may have low density sands in greater depths such as 30 meters. If you have water you may have liquefaction or at least strain softening and cyclic mobility

For me the criteria stated by Prof. Jonathan D. Bray (UCB) give very good envelope of possible liquefaction, see this video: https://www.youtube.com/watch?v=mO7LAZgtdoI

I second Stoyan Andreev's response above.  What I would add is that while Holocene deposits represent the vast majority of the natural deposits that are susceptible to liquefaction, there are rare occurrences of Pleistocene deposits that are still loose enough to liquefy.  As was stated above, I'm not aware of any examples of Neogene or older deposits that are susceptible to liquefaction.

Thanks a lot Dr. Andreev and Mabey!

If in a location, there is a water table in ground surface and sandy soil is located in the depth of 15 meters or more, isn't there any liquefaction potential for this location? and if liquefaction is happened during the earthquake, do you think the movement in the ground surface can be neglected?

The evaluation of liquefaction depth can be done graphically. First, the variation of equivalent cyclic shear stress (earthquake loading, ), with depth is plotted . The number of equivalent cycles, , corresponding to the earthquake magnitude must be determined if liquefaction resistance is to be characterized using laboratory test results. The variation of the cyclic shear stress required to cause liquefaction (liquefaction resistance) with depth is then plotted on the same graph (the values of must correspond to the same earthquake magnitude, or same number of equivalent cycles, as . Liquefaction can be expected at depths where the loading

exceeds the resistance. Below this depth, the resistance available is more than the cyclic shear stress caused by the earthquake and hence soil will not liquefy. See the attached figure. Also, read my online lecture notes from the following link for more details.

http://nptel.ac.in/courses/105108074/31

I think the liquefaction potential depends on the magnitude of excess pore pressure for sandy soils, no matter how deep it is. Normally, shallow location is easily to liquefy.

The nature of geologic processes is that one will seldom have a water table at the surface, no liquefiable sand/silty material in the upper 15 meters, and a loosely packed, liquefiable sand/silt below 15 meters.  If there is sandy/silty material in the upper 15 meters, then it will almost always (never say never) be more liquefiable, and more of a hazard, than any material beneath 15 meters.

It is important to remember what the processes of liquefaction is.  What happens when an earthquake shakes the ground (or some other source of cyclic stressing like heavy truck traffic or an explosion) is that the sand/silt grains are moving relative to one another in such a way that the vertical loads being carried by the grains of soil are instead transferred to the water filling the pore spaces.  That is the mechanism for the pore pressures rising.  The reason that the depth of a deposit is important is that the greater the depth, the greater the loads being carried by the sand grains.  This has two effects, both of which serve to inhibit liquefaction.  The first is that the loading, over time, tends to cause the grains to take on a denser packing arrangement which means that the grains have to move more to be moved into a less dense arrangement.  The second factor is that the higher load/greater stress on the grains means that the soil needs to be shaken harder to get the grains to move at all relative to one another.

Yet another factor is that the the strength of shaking actually diminishes with depth, as one move away from the free surface of the ground.  Thus, the cyclic stress that soil experiences is less at greater depth.

The damaging effects from liquefaction are due to both permanent ground deformation and differential ground movements (horizontal or settlement).  If a hypothetically liquefiable layer is overlain by 14 meters of rigid clay material, then it is very difficult to produce deformation that will cause damage.  The pore pressures may rise, but the water will likely be trapped in the layer, limiting the settlement that can occur.  Likewise the material above the liquefied layer will then be decoupled from much of the shaking coming from beneath.  Significant lateral spreading would require that there be an free face of 8+ meters height somewhere nearby.

I should mention that my comments are made in the context of earthquake induced liquefaction.  If one is instead invoking liquefaction due exclusively to excess pore-water pressures (i.e. a broken water main) then much that has been written doesn't apply.  We have been focused (rightly or wrongly) on liquefaction induced by cyclic stresses, because the 15 meter rule of thumb is a product of the work on this particular type of liquefaction.

I hope this helps to clarify why the 15 meter rule of thumb exists and works.  I will add that this rule of thumb is also supported by empirical observation of liquefaction occurrence due to earthquakes.  We just haven't seen settlement, ground deformation, or structural damage due to liquefaction in layers greater than 15 meters deep.

There is no simple answer to the question posed. The factors affecting liquefaction are numerous, and only when studied together can an estimate be made. It also depends on the application in hand, i.e., for a simple single story residential town house liquefaction below 20meter would not be of great concern.

However, as thoroughly discussed by peers above, the general rule of thumb is somewhere between 15-20meters below ground level provided that the ground is of deposits and water exists.

Note: There has been recorded events of liquefaction at depths of 60 meters below ground level and down to as deep as 300 meters below ground level.

Of course liquefation could occur at any depth. It depend on the tensional state, the depth beeing the origin of the main stress, sigma=depth x densitiy. In presence of dynimical load, like earthquakes, or quick changes in pore pressure (rapid drop of water level behind e dam, loose sands can liquefy at any depth, I presume.

Dear Rouhollah Basirat,

it is an very interesting question. At first you have to differentiate in natural and geotechnical sand bodies ( for example slopes and dump slopes in mining and s.o.) The liquefaction potential is depend how deep is the sand layer saturated. If the natural sand layer (loose rock) is more as 15 m saturated, then it is possible that exist a potential for land slide, otherwise not.

For example:

A abandoned dump slope of lignite mine of Nachterstedt/Germany slipped down, on 18.07.2009. Three people died, see picture. The potential of sliding was deeper as 15 m.

I hope I could help you.

Greetings

Michael Lersow

Thanks a lot everyone for answering this question and help me!

As told by other experts, liquefaction at larger depths may not occur due to overburden pressure.

@Michael: Can you give some details on this particular failure and if there is some published post-failure reconnaiscence and assessment?

Thanks Dr. Cavus!

Can you introduce some example with its references?

Thanks a lot everyone for answering this question!

For a result based on different ideas about liquefaction zone, there is no assurance about the depth of it. It's related to some parameters such as the strength of material, SPT, thickness of sand layer and water table level.

Please write your comments about the mentioned result. If you had any example of liquefaction potential zone with the depth of more than 15 meters, please share it with me!

Best Regards

Even though your question is 1 year old, I thought of adding more into Rouhollah Basirat's question for future followers: ͕

The maximum depth at which liquefaction can occur is probably the same as the maximum depth at which sands and silts can still remain unconsolidated and maintain a sufficient porosity and hydraulic conductivity. Whatever those depths, earthquakes of magnitude 8.0 or greater are capable of producing stresses between the hypocentral and epicentral zones sufficient to overcome overburden pressures at depths of 300 meters or more. If this is true, then a simple test well to determine the depth of unconsolidated sediments would suffice to determine the depth to which liquefaction Is possible In any given area threatened by great earthquakes.

From:

https://scholarsmine.mst.edu/cgi/viewcontent.cgi?article=3136&context=icrageesd

Hi. As Boriana said, Though it's been 2 years, I believe this question deserves a valid reference.

According to Geotechnical Design Manual of Washington State Department of Transportation, The maximum depth to consider liquefaction is limited to 80 ft (almost 27 meters). Although, if loose materials are present in the deeper depths, analysis should be performed to determine the probability of liquefaction.

However, the majority of the codes in Japan, being used by various organization, stipulate that the analysis of liquefaction be made for sandy deposits throughout depths shallower than 20m.

The depth limit of soil liquefaction, ideally, could exist. The question is how to identify the deep liquefied soil underneath if there is not evidence like sand ejection. The most used liquefaction evaluation methods were built on testing data on "liquefied" layers. The depth limit in code like Chinese seismic design code is supposed to be 20m because no deep liquefaction data below 20m were available when establishing the liquefaction evaluation method. However, the code methods would be updated with more and more deep liquefaction data. Please have a look at:

Yang Yang, Longwei Chen, Rui Sun, et al. A depth-consistent SPT-based empirical equation for evaluating sand liquefaction. Engineering Geology, 2017, 221: 41-49.

David Stewart and Ray Knox. What is the maximum depth liquefaction can occur? In proceedings: Third international conference on recent advances in geotechnical earthquake engineering and soil dynamics. 1995, St. Louis, Missouri.

Liquefaction could occur at any depth, if the hydrodynamic conditions (also caused by earthquakes) are such that the pore pressure exceeds the vertical load. It always happens, for example, in deep drillings if the fluid column is not dense enough. But it could also happen in layers with confined overpressures, in the presence of variable lateral forces.

At Kushiro Port in Japan, liquified layer lies between 6m to 26m depth. The main player is the excess pore water pressure.

Thabet, M., Nemoto, H. & Nakagawa, K. Variation process in stiffness inferred by nonlinear inversion during mainshocks at Kushiro Port vertical array site. Earth Planet Sp 60, 581–589 (2008). https://doi.org/10.1186/BF03353121

You can check this paper link- Article Assessment Of Liquefaction Potential From Sirajganj To Kurig...

DOI:10.9790/1684-1701033143

There is a possibility to occur liquefaction if the criteria match with liquefiable soil even though the depth is greater than 20 to 30 meters. @Rouhollah Basirat