21 April 2022
Wairoa: hundreds of landslides triggered by heavy rainfall in New Zealand
Wairoa: hundreds of landslides triggered by heavy rainfall in New Zealand
The last few weeks have seen high levels of rainfall in the Wairoa area of New Zealand. Most notably, in late march a series of major rain events triggered multiple landslides in this important agricultural area within Hawke’s Bay. RNZ has an excellent report on the situation as per late March, noting that:-
A Wairoa farming leader says the relentless rain still falling in northern Hawke’s Bay is like a recurring nightmare and farmers are beside themselves. Federated Farmers Wairoa branch chair Allan Newton said conditions were now 10 times worse than after last week’s deluge. Consistent rain meant the soil could not soak up any more water and the ground had turned liquid, he said.
The RNZ report includes a couple of images of examples of the landslides, including this one:-
Landslides in the Wairoa area of New Zealand. Image by Wairoa Helicopters via RNZ.
.
These landslides are visible even in the daily imagery collected by Planet. For example, this is an area affected by these failures (the image is in the area of 38.827, 177.618):-
Satellite imagery showing some of the new landslides in the Wairoa area of New Zealand. Image copyright Planet, used with permission. Image dated 10 April 2022.
.
As a comparison, this is the same area on 1 February 2022:-
Satellite imagery showing the area affected by landslides in the Wairoa area of New Zealand. Image copyright Planet, used with permission. Image dated 1 February 2022.
.
Below you should be able to see a slider allowing comparison of the before and after imagery:-
.
As the original image shows, this is an area that is heavily affected by landslides even before this year, but these rainfall events have triggered multiple shallow slips. It is notable that these are located primarily in the deforested grassland areas – there is little evidence of new failures in areas of woodland – and there appears to be a strong control by slope aspect (I see fewer slips on the westward facing slopes).
These shallow landslides affect an area that is much larger than I have shown in the images.
The slips will likely cause short and medium term challenges for the farmers, who have lost valuable grazing land, and the sediment into the channels will also have downstream impacts. The sediment in the river is clear in the photograph at the top of this piece.
.
Acknowledgement
Planet Team (2022). Planet Application Program Interface: In Space for Life on Earth. San Francisco, CA. https://www.planet.com/
20 April 2022
UNOSAT analysis of the Leyte landslides from Tropical Storm Megi
UNOSAT analysis of the Leyte landslides from Tropical Storm Megi
The United Nations Institute for Training and Research (UNITAR) has published an initial satellite image analysis undertaken by the United Nations Satellite Centre (UNOSAT) of the landslides and flooding caused by Tropical Storm Megi in Leyte in the Philippines. This is based on 50 cm resolution Pleiades imagery acquired on 14, 16 and 17 April. The analysis is available to download.
Included in the analysis is imagery of some of the large landslides. One image shows the landslide at Bunga village, using an images dated 14 April 2022:-
UNOSAT analysis of a Pleiades image of the landslide at Bunga village in Leyte.
.
The track length of this landslide is about 600 m, so it is a large failure. This has probably initiated as a shallow failure high on the hillside that entrained a large volume of material from the slope. It is interesting that the landslide appears to have initiated on a forested hillslope. The UNOSAT analysis suggests that 20 structures were destroyed or seriously damaged.
A second example is the Kantagnos village landslide, which I featured yesterday. The UNOSAT Pleiades image has captured the source area of the landslide:-
UNOSAT analysis of a Pleiades image of the landslide at Kantagnos village in Leyte.
.
With a track length of about 3 km, this is a very large failure. The width of the source area is a surprise to me – this is a landslide that needs a much more detailed investigation. I wonder to what degree it would have been possible to anticipate a failure of this scale at this location. The before and after comparison in the imagery is quite sobering:-
Before and after UNOSAT analysis of Worldview and Pleiades imagery of the landslide at Kantagnos village in Leyte.
.
Mapping of existing and potential landslides requires 3D imagery, but dangerous sites are often evident even in 2D images. But in the case of the Kantagnos landslide I would have struggled to identify this as a dangerous site on first inspection of the morphology of the slopes. The elements that might have made me suspicious are the areas that had been deforested for agriculture (and old landslide surface?) and, more importantly, those large, 90 degree bends in the river (has it been displaced by a landslide deposit?).
At Kantagnos, UNOSAT reports that 83 structures were destroyed or heavily damaged. The location has been declared to be uninhabitable. Inquirer.net reports that the latest information is that 59 people were killed at this site, and at least 63 remain missing.
The total toll from Tropical Storm Megi is now estimated to be 175 people, with a further 110 missing. Recovery operations have been terminated at most locations.
19 April 2022
Planet image of the Kantagnos village landslide in the Philippines
Planet image of the Kantagnos village landslide in the Philippines
My friends at Planet have very kindly acquired a high resolution satellite image of the Kantagnos landslide in the Philippines, which occurred during Tropical Storm Megi / Agaton on 11 April 2022. This image, which shows most of the the landslide, was captured on 17 April 2022 at 22:15 UT. This is the image:-
High resolution satellite image of the Kantagnos village landslide in the Philippines. Image copyright Planet, used with permission.
.
The main source of the landslide is in the cloud to the northeast of the image. As I noted previously, there is another substantial source that is visible. The flow has passed through a comparatively narrow gap and then spread. Note that there is very little evidence of landslide debris, implying a very mobile, fluid flow that has been removed by the river, which of course would have been in flood at the time of the failure.
For comparison, this is a Google Earth image of the same site ( I have included the location for reference), dated 20 April 2019:-
Google Earth image of the site of the Kantagnos village landslide in the Philippines.
.
The latest NDRRMC report lists 50 known fatalities at Kantagnos, with 93 people listed as missing. The total death toll from Tropical Storm Megi is now 175 people, with a further 110 people missing. The vast majority of the deaths occurred in landslides. Remarkably, an 11 year old boy survived the Kantagnos village in a refrigerator, but tragically his parents and a sibling were killed.
The impact of the landslide can be made clear by examining the inhabited area near to the river. This is the site in 2019:-
Google Earth image of the toe of the site of the Kantagnos village landslide in the Philippines.
.
And this is the site from the Planet image:-
High resolution satellite image of the Kantagnos village landslide in the Philippines. Image copyright Planet, used with permission.
.
Hope fully, below you will see a slide comparison of the two images:-
.
It is clear that multiple houses have been destroyed in the landslide, as well as a substantial area of agricultural land.
.
Acknowledgement
Planet Team (2022). Planet Application Program Interface: In Space for Life on Earth. San Francisco, CA. https://www.planet.com/
Many thanks in particular to Rob Simmon of Planet for providing the image. Simply amazing.
14 April 2022
Three very large landslides triggered by Tropical Storm Megi / Agaton
Three very large landslides triggered by Tropical Storm Megi / Agaton
Over the last two days more information has started to emerge about the landslides triggered by Tropical Storm Megi / Agaton, which caused mayhem on the island of Leyte in the Philippines on 11 April 2022. The latest news on casualties is that at least 123 people were killed, the majority in landslides, but more remain missing.
Earlier this week I highlighted a major landslide at the village of Kantagnos in Baybay. The Inquirer has now published an image that provides a detailed view of this site:-
The aftermath of the landslide at Kantagnos village in Baybay City, Leyte, Philippines, triggered by Tropical Storm Megi. Image from Inquirer.net.
.
This is a very large failure in deeply weathered materials. The source appears to be a shallow slide on a steep, forested hillslope, but note that there are at least two other sources, albeit substantially smaller. The landslide has spread considerably to affect a very large swathe of land. The small number of standing trees might suggest that it was a comparatively shallow flow, but to a human or a weak structure it would have been devastating. Note the houses on the margin of the landslide on the right side of the image.
A second major event occurred in the village of Pilar in Abuyog. Here at least 29 people are known to have died, but as many as 177 more are reported to be missing. This total is likely to reduce, but it is clear that the landslide here has been devastating.
I can find few images of this landslide online, but Youtube has a drone video of the site:-
.
I have taken this image from the video, providing a decent impression of the failure:-
The devastating landslide at Pilar, triggered by Tropical Storm Megi. Image from Youtube drone video.
.
This appears to be a deeper seated landslide that has transitioned into a flow. Note the stripping of the vegetation on the slope on the right side of the image, which suggests to me that this was a high runout speed event. I also wonder if this event generated a displacement wave.
There are some videos online of secondary failures at this site, such as this one:-
.
The power of even this smaller landslide is appalling.
A third example occurred at the village of Bunga, also in Baybay. This landslide was also fatal – at least 17 people were killed here. The Jakarta Post has an image of this landslide:-
The landslide at Bunga in Leyte, triggered by Tropical Storm Megi. Image from AFP/Bobbie Alota via the Jakarta Post.
This landslide is broadly similar to the one at Kantagnos, although on a smaller scale.
I have been trying to track down the locations of these three landslides. At the moment, I believe that the following at correct:
Kantagnos in Baybay: 10.677, 124.853
Pilar in Abuyog: 10.714, 125.048
Bunga in Baybay: 10.773, 124.785
The impact of these landslides, and there are reports of many other events as well, is very serious. Tropical cuclones often cause landslides in the Philippines, but not in April.
12 April 2022
Landslides from Tropical Depression Agaton (Megi) in the Philippines
Landslides from Tropical Depression Agaton (Megi) in the Philippines
On Sunday 10 April 2022 Tropical Depression Agaton (known as Tropical Depression / Storm Megi internationally) tracked across Leyte in the Philippines, bringing torrential rainfall. Reports indicate that there have been multiple landslides causing widespread damage. The most damaging landslides appear to have occurred in the Baybay City region, where at least 22 people have been killed in multiple slides. Further people are missing.
Another very significant landslide has occurred at Kantagnos village in Baybay City. The Philipppines News Agency has this image of the downstream impact of this landslide, which is reported to have killed at least three people:-
The aftermath of the landslide at Kantagnos village in Baybay City, Leyte, Philippines, Tropical Depression Agaton. Image by Pobreng Biyahero via Philippine News Agency.
.
Al Jazeera has this image of what is I think the same site, showing a very large landslide scar:-
The aftermath of a landslide, possibly at Kantagnos village in Baybay City, Leyte, Philippines, triggered by Tropical Depression Agaton. Image from Al Jazeera.
.
The location of this landslide appears to be 10.677, 124.853. This is a perspective view on Google Earth:-
Google Earth perspective view of the likely site of the landslide at Kantagnos village in Baybay City, Leyte, Philippines, Tropical Depression Agaton.
.
This appears to be a very large failure and flow. There are images online of other large events, and one Youtube video of a substantial failure, but at the moment I am struggling to determine whether these landslides are from Tropical Depression Agaton.
Hopefully the true picture will become clearer in the coming days. Leyte has a long track record of damaging landslides, usually triggered by tropical cyclones. However, April seems to me to be an unusual time of year for a rainfall event of this magnitude.
11 April 2022
Bedruthan Steps: another “face” in a landslide scar
Bedruthan Steps: another “face” in a landslide scar
Over the years I have collected a set of stories in which a “face” apparently appears in a landslide scar (and other examples of landslide weirdness). For example, in 2015 an image of Jesus apparently appeared in the scar of a landslide in the San Francisco area of Putumayo in Colombia.
Devon Live has a really good example, based upon a rockfall that occurred in a sea cliff at Bedruthan Steps in Cornwall:-
The image of a face following a rockfall at Bedruthan Steps in Devon. Image by Louise Coe via Devon Live.
.
In reality of course this is the fascinating phenomenon know as face pareidolia, in which the human brain is programmed to identify faces rapidly. At times this mechanism identifies faces in inanimate objects – the theory is that the disadvantages of doing so are outweighed by the advantages of rapid facial recognition. Interestingly, recent research has indicated that we are biased towards seeing male rather than female faces in such situations.
Most rockfalls are defined by pre-existing joints or other discontinuities in the rock mass, meaning that the scar contains strong linear features. In many cases, the discontinuities are not random, meaning that patterns emerge. It is unsurprising that these occasionally show similar features to faces.
In the case of Bedruthan Steps, the discontinuities have defined two inverted wedges that form the angular “eyes” of the face. There is a less well-defined wedge with a similar form that has created the “nose”. These joints can be seen in other locations across the rock mass.
7 April 2022
The long term history of fatal landslides in Colombia
The long term history of fatal landslides in Colombia
Colombia is one of the most landslide-prone countries in the world. In the last few days heavy rainfall has triggered flooding and landslides, including at least one death in Marulanda Municipality in Caldas. Understanding the patterns of landslides in time and space is an important element of managing the risk in such countries.
In a paper just published in the journal Landslides (Garcia-Delgado, Petley et al. 2022), an international team has constructed a long term dataset of fatal landslides in Colombia, extending right back into historical times. This comprises 2.351 individual fatal landslides causing 37,959 fatalities. The earliest event was in October 1743, but the vast majority have been recorded since 1912, when the data becomes more robust.
The temporal pattern of the landslides in time is really interesting – this is illustrated in the graph below:-
Numbers of landslides and landslide fatalities in Colombia through time. Graph from Garcia-Delgado et al. (2022). The figure caption is: Temporal distribution of deadly landslides, their fatalities and cumulative total of fatalities from 1912 to 2020. The numbers in circles correspond to the top 10 deadliest events as follows: (1) Túquerres (9 January 1936; n = 170 fatalities); (2) Ibagué (29 June 1959; n = 250); (3) Chiriguaná (10 November 1970; n = 123); (4) Quebrada Blanca–Guayabetal (28 June 1974; n = 400); (5) El Playón (25 November 1979; n = 300); (6) Armero–Chinchiná (13 November 1985; n = 23,084); (7) Villatina–Medellín (28 September 1987; n = 562); (8) Andes (26 April 1993; n = 120); (9) Páez (6 June 1994; n = 1100); and (10) Mocoa (31 March 2017; n = 333). To check the location of these events please refer to Fig. 2. The upper-left inset shows annual fatalities normalised by the number of events.
.
This is an incredibly rich and interesting figure. Note the huge jump in the cumulative fatalities associated with the lahars caused by the Armero–Chinchiná lahars, which killed over 23,000 people). Also notable is the increase in fatalities through the period from 1912 to the 1970s. Since then the number of fatal landslides has remained high but the strong positive growth has slowed. The two decades with the highest numbers of fatalities were 2001-2010 and 2011-2020. The inset graph normalises fatalities with the population (which has grown over the study period), suggesting that the per capita rate of loss has remained about the same.
The paper looks at this data in a great deal more detail, including the role of mining and road construction in landslide trends, and also the impact of the El Nino / La Nina cycle. I will return to these themes in subsequent posts. Putting together a dataset like this is an epic job – Helbert Garcia-Delgado of the Department of Earth and Environmental Sciences at Syracuse University Servicio Geológico Colombiano, who led the research, has been brilliant.
.
Reference
Garcia-Delgado, Petley, D.N., Bermúdez, M.A. and Sepúlveda, S. 2022. Fatal landslides in Colombia (from historical times to 2020) and their socio-economic impacts. Doi: https://doi.org/10.1007/s10346-022-01870-2.
6 April 2022
A tailings dam failure in Wenquan Township Jiaokou County, Shanxi Province, China
A tailings dam failure in Wenquan Township, Shanxi Province, China
Yesterday my friends at World Mine Tailings Failures alerted me to a report from China of a tailings facility failure in Wenquan Township, Jiaokou County, Shanxi Province, China. The report provides limited detail, the salient points of which are as follows:-
On the afternoon of March 27, a landslide occurred in a material yard of Shanxi Daoer Aluminum Co., Ltd.
According to reports, at around 17:00 on March 27, a tailings pond of Daoer Company, located in Wenquan Township, Jiaokou County, broke a dam. Part of the factory area was buried; a large pit was formed in the tailings pond, the dam body was flushed with a gap of about 70 to 80 meters, and the mud accumulation blocked the local road...According to the report of the People’s Government of Jiaokou County, after preliminary inspection, a landslide occurred in a material storage yard of Daoer Company in the county, causing 7.5 mu of arbor forest land to be buried, more than 200 meters of seasonal ditches and rural roads were blocked, and part of the surrounding walls of adjacent enterprises were washed away. No casualties were found.
I have been able to track this event down on Planet imagery. The area around Wenquan Township is decimated by mining, but the tailings failure stands out quite clearly:-
Satellite image of the 27 March 2022 tailings dam failure at Wenquan Township in Shanxi Province China. Image courtesy of Planet, used with permission.
.
The location of the tailings dam failure is 37.1059, 111.4301. The mobilised tailings appear to have moved about 700 m from the crest of the scar to the toe of the waste, or about 500 m from the edge of the failed retaining structure to the toe of the dam.
.
I have created a Juxtapose slider to compare a Planet image from before the failure (collected on 7 March 2022) with the one after (collected on 28 March 20222 – the day after the collapse):-
.
There are inevitable resolution limitations in this genuinely amazing daily Planet imagery, but it confirms aspects of the report in the news report. In the runout zone the tailings have clearly partially inundated an industrial facility – presumably the concentrator – and a road is blocked. The remainder of the tailings are mostly covering vegetated slopes.
The most recent (5 April 2022) imagery suggests that the road has been cleared and the tailings have been covered in a grey material. I am unsure as to what this might be.
The Wenquan Township event is not a large tailings failure, and the runout is comparatively modest. Nonetheless, it is of concern given the number of such facilities in China.
.
Acknowledgement
Planet Team (2022). Planet Application Program Interface: In Space for Life on Earth. San Francisco, CA. https://www.planet.com/
5 April 2022
Mount Bolu: an interesting case study of a tunnel portal landslide in Turkey
Mount Bolu: an interesting case study of a tunnel portal landslide in Turkey
On 2 April 2022 heavy rainfall triggered an economically significant landslide in at Mount Bolu in Turkey. The landslide occurred after a spell of rainfall that in places has fallen onto snow. This is a recipe for landslides in upland areas.
The Mount Bolu Tunnel landslide is interesting though. It has blocked four lanes of the D-100 road, a critical east-west connection across the north of Turkey. This is an image of the landslide:-
The aftermath of the landslide at Mount Bolu in Turkey. Image by AA via the Daily Sabah.
.
This is an alternative view that shows the landslide source more clearly, as well as some of the debris blocking the carriageway:-
The landslide at Mount Bolu in Turkey. Image from https://www.turkiyegazetesi.com.tr/.
.
The site of the tunnel is at 40.7578, 31.4504. It is clearly visible on Google Earth. The tunnel was completed in 2007.
There are some very interesting images online of the site of this landslide prior to the 2 April 2022 landslide. This image for example shows the tunnel portals and the slope above:-
Archive image of the site of the Mount Bolu landslide, showing the pre-existing landslide scarps. Image from https://structurae.net/.
.
The archive of Google Earth imagery does not seem to indicate that further works have been undertaken to stabilise this slope, although of course satellite/aerial imagery has limitations. But on first inspection this is an ugly slope to be located above a tunnel portal on such an important highway, especially in a region that is seismically active.
The debris has taken 44 hours to clear, but the road has now reopened. The images suggest that further works will be needed on the slopes at this site, both in the scar of the 2 April 2022 landslide and on the slope to the left of it (as seen in the images), which appears to be another potential landslide site.
Northern Turkey has been suffering from a wave of landslides in recent days:-
Rapid snow melts have triggered more than 400 landslides in northern Turkey (in Bolu, Bartin, Samsun, Ordu, and Trabzon) in the last week. This foehn effect looks like will continue along the next week. Continue the landslide watch. Video credit @ihacomtr pic.twitter.com/oUfexjgOJH
— Tolga Gorum (@TolgaGorum) April 3, 2022
4 April 2022
Praia de Itaguaçu: one of a series of destructive landslides in Brazil
Praia de Itaguaçu: one of a series of destructive landslides in Brazil
On Saturday 2 April 2022 a series of landslides were triggered by heavy rainfall in the state of Rio de Janeiro in Brazil. In total 16 people are known to have died, the majority of whom lost their lives in landslides. The picture is a little confused at the moment, but reports include a mudslide in the town of Paraty, which killed two children and injured four people; another landslide in Paraty that killed a mother and seven of her children; and a landslide in the Monsuaba neighborhood of Angra dos Reis, which killed eight people, including four children. Three more people are missing in Monsuba. You will see that the total fatalities in the events described above is greater than the reported death toll, so I await greater clarity.
One major landslide occurred at Praia de Itaguaçu, a popular beach on Ilha Grande in Angra do Reis (note that there are several beaches with this name in Brazil). There are reports that four people may have been killed. There is a video taken from a boat showing the aftermath of this large landslide that has been widely tweeted, such as below:-
Impressionante | Um deslizamento de terra atingiu a Praia de Itaguaçu em Ilha Grande, Angra dos Reis. A faixa de areia do local desapareceu após queda da barreira. Ainda não há informações de feridos. #sbtrio #sbt pic.twitter.com/xrSdpSIgOL
— SBT Rio (@sbtrio) April 2, 2022
.
This includes the following view of the landslide:-
The aftermath of the 2 April 2022 landslide at Praia de Itaguaçu in Brazil. Image from a video that has been widely tweeted.
.
Once again we await better imagery, which is likely to become available today, although further rain is forecast in this area.
Back in 2015, Sergio Sepulveda (now of Simon Fraser University) and I analysed data on landslides in Central and South America (Sepulveda and Petley 2015 – available open access). We noted that there are two annual peaks of landslides in South America, one in January and one in April. As such, these landslides fit the annual pattern, but the current La Nina conditions may be playing a role in intensifying the rainfall that has triggered this most recent series of disasters.
.
Reference
Sepulveda, S.A. and Petley, D.N. 2015. Regional trends and controlling factors of fatal landslides in Latin America and the Caribbean. Natural Hazards and Earth System Sciences, 15, 1821-1833, doi: 10.5194/nhess-15-1821-2015.
Dave Petley is the Pro-Vice-Chancellor (Research and Innovation) at the University of Sheffield in the United Kingdom. His blog provides a commentary on landslide events occurring worldwide, including the landslides themselves, latest research, and conferences and meetings.