5 July 2021
The Atami mudslides in Japan
Posted by Dave Petley
The Atami mudslides in Japan
Over the weekend there was considerable attention paid to the remarkable videos of the mudflows in Atami, Japan. This one in particular gained traction:-
.
But there are several others, which are helpfully combined into a compilation here:
.
I particularly draw your attention to the excerpt that begins at 8 minutes 16 seconds into the second video (the same sequence appears in the first video too).
The main, extremely violent, footage is a particularly nasty example of a channelised flow, common in steep terrain that is subject to extreme rainfall. Long term readers of this blog will remember the classic examples from Lantau in Hong Kong in 2008 and South Korea in 2011 for example, but there are many, many other examples.
The Google Earth image below shows the general location of the Atami disaster. I have placed the marker on what is, I think, the location of the prominent red building in the main video:-
.
Some important insight into this event is provided in a tweet by Yoshi Kariya:
#熱海 #伊豆山 #土石流 発生の谷。斜面崩壊の最前線である谷頭が火山斜面を刻んでいくつも伸び、谷底では堆積した過去の土石流・地すべり堆積物が棚田状に造成された様子がわかる。流域上部の尾根に造成地があったようだ[地理院2017年写真]。 pic.twitter.com/aRCbV1pQNJ
— 苅谷愛彦 Yoshi Kariya (@yoshi_kariya) July 3, 2021
.
A translation of the text is:
Valley of occurrence. It can be seen that the valley head, which is the forefront of slope failure, extends several times along the volcanic slope, and the past debris flow and landslide deposits that have accumulated at the bottom of the valley are formed in the shape of a terraced rice field. It seems that there was a land on the ridge in the upper part of the basin [Geographical Survey Institute 2017 photo].
Shizuoka Prefecture has released a drone video that shows the landslide at the crown of the mudslide:
.
This suggests that the Atami mudflows started as a conventional landslide high up in the catchment. This is a typical response to very heavy rainfall. The ground will probably have undergone static liquefaction to form the highly mobile flow. Once channelised, it will have entrained water from the channel and deposits within the channel, especially those left from previous landslides, to create this terrifying mudflow.
The destructive power of these events is much greater than that of a river flood as the density of earth materials is much higher.
There are two interesting elements to this landslide. The first is that the initial failure appears to be quite deep-seated. The second is the presence of the road right at the headscarp. In other cases, most notably the Sarno landslides in Italy, roads have played a key role in the destabilisation of the slope. Given the deep-seated nature of the landslide, that may not be the case here, but it is an interesting juxtaposition.
Japan has high quality landslide management programmes and amongst the best engineering prowess for managing these hazards. However, the combination of the geological setting, the climate and the high population levels means that entirely preventing these events is impossible.
Sadly, but unsurprisingly perhaps, the number of people reported missing at Atami has jumped overnight. Latest reports suggest that there are three known fatalities but as many as 80 people may be missing. Hopefully this will reduce in the next few days as people away from home are traced, but the losses may well be high.
Looks like the event was on the SE flank of Mount Hakone (Hakoneyama), a massive caldera neighboring Mount Fuji. This area has tens of meters thick pyroclastics and ashfalls, which destabilize quite nicely with heavy rains (think Pinatubo in the Philippines). The mudflow can also be described as a lahar. Google Earth also describes Atami having numerous natural hot springs. This means it is connected to the hyrdothermal system of Hakone, which further destabilizes the pyroclastic and ashfall deposits, turning them into clays over time, a bad thing on a slope. Cheers –
Could it be that the road at the upper limit of the detachment actually marks the upper limit of its propagation up the slope? The slide may have nucleated lower down.
In the drone video, on the left side of the failure, there seems to have been some work in progress installing pipes. A pile of pipes is visible behind the remaining road and on the near side of the road the remnants of a trench are visible. Wonder if the trench filling with water might have triggered the initial failure
The initial and supply water of the landslide may affect the final disaster form:
The landslide in Shuicheng County, China, is similar to this case. Its failure on the side of the road and was located at a high level. Its initial water content was low and finally transformed into a debris avalanche.
The Wangcang landslide in China has medium initial water content, but the supplied water is much higher, which is finally transformed into a debris flow.
In this video, the soil showed a flowing state after failure at the back scarp, which may be related to the high-porosity volcanic ash with a strong water-holding capacity at the bottom.
Guo, J., et al. “The effect of topography on landslide kinematics: a case study of the Jichang town landslide in Guizhou, China.” Landslides 17.4(2020):959-973.
Guo, J., et al. “Discussions on the transformation conditions of Wangcang landslide-induced debris flow.” Landslides 18.5(2021):1833-1843.
That is clearly the materials used to fill the upper valley during previous earthworks (houses and solar panels) that failed. BAd matrieal, lack of drainage, lack of vegetatlisation… pretty sure the rain falling on that upper road made its way violently through that slope…
is there an estimation of the volume of this landslide?
I note a broken culvert spewing water at 4:18 in the second video. During the El Nino rains of 1996/97 in NW California, there were a couple of instances I recall where under-sized drainage pipes appeared to have been running 100% full (and under pressure because of standing water above their inlets – no breather pipes, no pipe inlet mitering). Surrounding cross-gully fill appeared to have deformed enough to pull pipe joints apart allowing water to be injected into the fill.
This may have been a similar case.
I am so sorry for all involved. Praying no one has lost lives. Please know we all here on Pacific Northwest Coast USA are thinking of you. What do you need we citizens might help with?
I will be thinking of you all now. All love.
Have seen terrible mudslides. So much damage.I am so sorry for all involved. Praying no one has lost lives. Please know we all here on Pacific Northwest Coast USA are thinking of you. What do you need we citizens might help with?
I will be thinking of you all now. All love.
I thought I was sending this to Atami!