- Email: [email protected]
The formation and impact of landslide dams – State of the art
Journal Pre-proof The formation and impact of landslide dams – State of the art
Xuanmei Fan, Anja Dufresne, Srikrishnan Siva Subramanian, Alexander Strom, Reginald Hermanns, Carlo Tacconi Stefanelli, Kenneth Hewitt, Ali P. Yunus, Stuart Dunning, Lucia Capra, Marten Geertsema, Brendan Miller, Nicola Casagli, John D. Jansen, Qiang Xu PII:
S0012-8252(19)30476-3
DOI:
https://doi.org/10.1016/j.earscirev.2020.103116
Reference:
EARTH 103116
To appear in:
Earth-Science Reviews
Received date:
15 July 2019
Revised date:
5 February 2020
Accepted date:
5 February 2020
Please cite this article as: X. Fan, A. Dufresne, S.S. Subramanian, et al., The formation and impact of landslide dams – State of the art, Earth-Science Reviews(2020), https://doi.org/ 10.1016/j.earscirev.2020.103116
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier.
Journal Pre-proof
The formation and impact of landslide dams – state of the art Xuanmei Fan1, Anja Dufresne2* , Srikrishnan Siva Subramanian1* , Alexander Strom3 , Reginald Hermanns4,5, Carlo Tacconi Stefanelli 6, Kenneth Hewitt7, Ali P. Yunus1, Stuart Dunning8, Lucia Capra9, Marten Geertsema10, Brendan Miller10 , Nicola Casagli 6, John D. Jansen1,11, Qiang Xu1 1
State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of
Technology, 610059, Chengdu, Sichuan, People’s Republic of China 2
Engineering Geology and Hydrogeology, RWTH-Aachen University, Lochnerstr. 4-20, 52066 Aachen,
Germany Geodynamics Research Centre, Volkolamskoe Shosse 2, 125080 Moscow, Russia
4
Geohazards and Earth Observation, Geological Survey of Norway, Trondheim, N-7491, Norway
5
Department of Geoscience and Petroleum, Norwegian University of Science and Technology, Norway
6
Department of Earth Sciences, University of Firenze, Via La Pira, 4, Florence 50121, Italy
7
Department of Geography and Environmental Studies, Wilfried Laurier University, 75 University Avenue
pr
oo
f
3
8
e-
West, Waterloo, Ontario N2L 3C5, Canada
School of Geography, Politics and Sociology, Newcastle University, Newcastle upon Tyne, NE1 7RU,
9
Pr
UK
Centro de Geosciencias, Universidad Nacional Autónoma de México, Mexico City
10
Ministry of Forests, Lands, Natural Resource Operations, and Rural Development, Prince George,
al
British Columbia, Canada 11
ur n
GFÚ Institute of Geophysics, Czech Academy of Sciences, Prague, Czechia.
*Corresponding author: Dr Anja Dufresne, email: [email protected], Dr.
Jo
Srikrishnan Siva Subramanian, email: [email protected]
Highlights:
Compilation of a new worldwide database of 410 landslide dams of > 1 million m3 in volume and design of a template for future data collection. A preliminary landslide dam stability classification matrix combining topographic setting and the sedimentological features. Comprehensive review and evaluation of landslide dam stability criteria and geomorphic indices. Landscape impacts of landslide dams discussed on different time scales.
1
Journal Pre-proof
Abstract The blocking of river courses by mass movements is very common in mountainous areas with deep and narrow valleys. Landslide dams may pose serious threats to people and their livelihoods downstream in the case of abrupt dam failure. Since the publication of benchmark reviews of Costa and Schuster (1988) and Korup (2002), there is a growing number of studies focusing on the formation, stability, and short-term impacts of landslide dams. This review combines the insights of all these studies, builds
oo
f
on current concepts of landslide dams, and suggests ways to unify terminologies and classifications. We furthermore present a new worldwide database compiled from
pr
literature data. It contains 410 landslide dams >1 million m3 in volume that were formed
e-
since 1900 since these have the most complete data entries. These data show that
Pr
dam longevity is, among other factors, correlated with the type of landslide forming the dam. Those formed by rock/debris avalanches and rockslides have longest lifespans.
al
However, the influence of landslide type or material on dam longevity decreases with
ur n
time after dam formation. To ensure consistency in the next database generation, we suggest guidelines for data collection to provide a solid basis for evaluating dam
Jo
stability and governing factors. A preliminary classification matrix for landslide dam stability that combines topographic setting and the internal structure of the dam body is another outcome of our review. Furthermore, an evaluation of the various geomorphic stability indices proposed in the literature regarding their suitability and limitations in assessing dam formation and stability shows that they predict the probability of dam formation reasonably well, but that their application to longevity estimates requires further assessment. The geomorphic impacts of landslide dams in the short-, mediumand long-term are summarized and illustrated with key examples. Finally, for a better understanding of the factors controlling dam stability, we recommend to (1) include dam 2
Journal Pre-proof composition and sedimentary structures in future case studies, (2) maintain and update the worldwide database for sound statistical analyses, (3) refine landslide dam stability indices and test them for different landslide types, and (4) study hazard cascades related to multiple dams in one watershed. For long-term landscape evolution studies, we suggest to (5) quantify terrestrial sediment flux related to landslide dams, (6) detect ancient landslide dams in river profiles, and (7) further exploit the sediment archives in former impoundment areas.
oo
f
Keywords: worldwide database; classification; stability; geomorphological impacts;
pr
landscape evolution
e-
1 Introduction
Natural damming of rivers by mass movements and their consequences pose a major
Pr
hazard in mountai nous areas with deep, narrow valleys (Capra, 2006; Costa and Schuster, 1988; Ermini and Casagli, 2003; Evans et al., 2011a; Fan et al., 2019a;
al
Hermanns, 2013; Hewitt, 1988; Korup, 2002; Korup and Tweed, 2007; Swanson et al.,
ur n
1986). Though natural river-damming events are induced not only by landslides (there are moraine dams and dams formed by fault ruptures), those formed by large landslides
Jo
are most hazardous (Mason, 1929; Shroder Jr, 1998), and this review is focused only on these. Large landslide dams often result in more complex threats to people than the landslides themselves, both in space and time, by inundating upstream and potentially flooding downstream areas (Davies and Korup, 2010; Groeber, 1916; Penna et al., 2013), making dam stability assessment crucial. The stability of landslide dams largely depends on the composition and geotechnical properties of the dam material (Casagli and Ermini, 1999; Korup, 2002; Schuster, 1986; Walder and O'Connor, 1997). However, until the landslide dam is cut—artificially or naturally—its internal geological structure and material properties are unknown in most cases, though some can be 3
Journal Pre-proof extrapolated from analogue or empirical case studies. Despite clear identification of the factors controlling landslide dam formation and stability, the actual mechanism of formation or failure in relation to the material composition, dam geometry, the impounded catchment area, lake volume, and water inflow rates still remain unclear and warrant a thorough review. Hazards and risks related to these dams have long been recognized and are documented in many historical accounts of catastrophic floods from natural dam failures
oo
f
(Dai et al., 2005; Delaney and Evans, 2015; Fan et al., 2012a; 2012b; 2012c; 2013; Groeber, 1916; Lee and Dai, 2011; O'Connor and Costa, 2004; Peng and Zhang,
pr
2012). Changes in climatic and meteorological conditions often play a part in
e-
‘unexpected’ dam failures. Examples include the Ram Creek landslide dam (New Zealand), which remained stable for 13 years before failure after heavy rain (Harrison et
Pr
al., 2015) or the Rio Barrancas rockslide dam in Argentina, which failed 425 years after
al
its formation due to increased inflow of glacial meltwater during the hottest austral summer on record and an antecedent, unusually wet winter (Evans et al., 2011a). The
ur n
many unknowns in assessing both immediate and long-term stability complicate the management of potential hazards from landslide dams.
Jo
Benchmark reviews on landslide dams by Costa and Schuster (1988) and Korup (2002) addressed the complex interactions between landslide morphometric characteristics and magnitude, geomorphological parameters of the valley, and hydrological parameters of the river at a given location. Since then, a large body of literature on specific aspects of landslide dams has emerged. The most widely recognized works, which are relevant for the preparation of our review, are summarized in Table 1 , including a book edited by Evans et al. (2011a) that presents a review focussed on natural and artificial rockslide dams mostly with volumes in excess of 106 m3 .
4
Journal Pre-proof Efforts were also made in creating both worldwide and regional landslide dam inventories, but there are still some difficulties. For example, inventories encompass dams formed by all types of landslides rather than detailed breakdowns of the subtypes. The other challenge is that the wide range of terminology and landslide dam classification schemes without a global standard makes cross-comparison of methods and data difficult. To understand dam formation and failure, the types of mass movements that potentially form landslide dams should be categorized based on their
oo
f
varying characteristic sizes and geotechnical properties. With this review we aim at offering a unified landslide dam terminology, propose a new
pr
classification scheme based on existing ones, compile existing landslide dam
e-
inventories and thereby create a worldwide database, present the state of the art on formation and stability, and, review the impact of landslide dams over multiple
Pr
timescales (see 1.1 Structure of the review).
Focus
ur n
Previous benchmark reviews
al
Table 1 A summary of key papers, reviews and meta-analyses on landslide dams Key references Costa and Schuster (1988); Korup (2002); Korup and Tweed (2007); Evans et al. (2011a) Volcanic dams: Capra (2007); Capra (2011);
formed by different types of landslides and
Rockslide and rock avalanche dams: Dufresne et al.
Jo
Papers focus on studying landslide dams in different materials
(2018); Hermanns et al. (2006); Hermanns et al. (2011a); Hermanns et al. (2011b); Hewitt (2009a); Hewitt (2011) Unconsolidated sediments: Geertsema (1998)
Worldwide landslide dam datasets;
Schuster and Costa (1986), 187 cases;
event-based landslide dam inventory of the
Costa and Schuster (1991), 463 cases;
2008 Wenchuan earthquake;
Ermini and Casagli (2003), 350 cases;
other regional datasets are reviewed in
Peng and Zhang (2012), 1239 cases
Section 3
Strom and Abdrakhmatov (2018), 190 cases Fan et al. (2012a); Fan et al. (2012b); Fan et al. (2012c), 828 cases from the Wenchuan earthquake
Empirical indices predicting landslide dam
Swanson et al. (1986), Ermini (2003b), Dal Sasso et
formation
al. (2014), Tacconi Stefanelli et al. (2016) 5
Journal Pre-proof Empirical classification schemes and
Casagli and Ermini (1999); Dal Sasso et al. (2014);
geomorphic indices for landslide dam
Dong et al. (2009); Ermini (2003a); Ermini and
stability
Casagli (2003); Korup (2002); Korup (2004); Tacconi Stefanelli et al. (2016)
Landscape implications of landslide dams,
Hewitt (2002); Korup (2005a); Schuster (2006);
including their impacts on valley
Korup (2012); Jansen (2006); Korup et al. (2006)
morphology, sediment transport, fluvial incision etc.
1.1
Structure of the review
This review covers recent developments in understanding the formation and impacts of
oo
f
landslide dams. The lifecycle of a landslide dam involves the origin (the landslide),
landscape impacts throughout all stages.
pr
survival (longevity or stability), breach (sudden failure or slow erosion), and related
e-
The terminology used for characterizing landslide dams, their lakes and stability are
Pr
discussed in Section 2. Researchers around the world use different terminologies and definitions to characterize landslide dams, but no classification scheme with a near
al
standardized use comparable to the one devised by Varnes (1978) and Hungr et al.
ur n
(2014) for landslides themselves is available. In Section 2, an attempt is made to review, condense and supplement the common terminology and dam classifications ,
Jo
and to provide definitions that will make scientific communication easier if applied widely. Additionally, a complimentary classification scheme is presented based on the internal structure of landslide deposits that affects dam stability, although the scheme may not apply in many cases until the landslide dam is eroded. Section 3 presents an analysis of our worldwide landslide dam database. We collected and reviewed data from all known landslide dam hotspots around the world: the Tibetan mountains, the Himalayan regions, Central Asia including Pamir and Tien Shan mountains, tectonically active mountain ranges of Japanese and New Zealand, the European Alps, the Rocky Mountains and the Andes. Some event-based inventories 6
Journal Pre-proof are also included, such as the 2008 Wenchuan (China), the 2015 Gorkha (Nepal), and the 2016 Kaikoura (New Zealand) earthquakes. We identify inconsistencies and problems in the complied inventories and recommend a template for future data collection—provided as supplementary material. Various aspects of landslide dam stability are summarized in Section 4. We give a synopsis of the geomorphic stability indices proposed in the literature and illustrate their applicability based on our new database.
oo
f
A review of the geomorphic impacts of landslides dams in short-, medium-, and longterms are the focus of Section 5. Under short-term impacts, importance has been given
pr
to dam formation and stability within the first year after formation (often the most critical
e-
for immediate hazard assessment). River morphological changes and sediment fluxes lasting years to decades are considered medium-term and presented with key
Pr
examples. The long-term geomorphic impacts through centuries and millennia are
al
presented from a geological perspective, considering also the factors causing sudden collapse of dams that remained (apparently stable) for centuries to millennia.
ur n
Despite the decades-long research and associated advances in our understanding of the formation and failure of natural dams, many challenges still remain. Section 6
Jo
summarises the key research challenges and suggests future work to answer them.
2 Terminology and classification 2.1
Terms characterizing landslide dam and stability
As with landslide processes, landslide dam terminology varies by author, region and research focus, with different classifications existing in parallel. A common-sense approach is required when searching for literature pertaining to landslide dams and landslide-dammed lakes, and integrating the reported metrics. Terms common in studies describing landslide dams include landslide blockage, stream/river blockage, 7
Journal Pre-proof and, landslide barrier. Lakes that form behind a landslide dam have been variably termed landslide lake, impoundment, landslide-dammed lakes, landslide-barrier lake, reservoir, backwater lake, and earthquake -dammed lake. The term earthquakedammed lake, however, ignores the fact that river damming might be caused by coseismic faulting or folding, as during the 1980 El-Asnam earthquake in Algeria (Ruegg et al., 1982). Though ‘landslide lake’ seems to be the commonly accepted choice for lakes that formed on top of a landslide or landslide dam rather than behind it upstream,
oo
f
we recommend the term ‘supra-landslide lake’ (following the definition of 'super lake'; Crozier and Pillans, 1991) for future use since it clearly separates it from any lake
pr
formed behind the blockage and is more consistent with other fields (e.g. supra-glacial
e-
lake). An impoundment is any body of water formed by confinement. This term is, however, commonly used to describe the area upstream of a dam regardless of the
Pr
presence of a lake. The term ‘reservoir’ is often found in the context of processes
al
affecting the lake, akin to hydropower dam studies, whereas ‘backwater lake’ or ‘landslide-dammed lake’ are most commonly found in geomorphological studies.
ur n
Herein, we adopt ‘impoundment’ for the upstream area, and ‘landslide-dammed lake’ when processes pertaining to the water body are discussed and in all general,
Jo
descriptive situations.
The empirical classification schemes of landslide dam stability by Ermini and Casagli (2003), Korup (2004), and Tacconi Stefanelli et al. (2016) currently in use (Table 2) build upon the initial definition by Casagli and Ermini (1999), which is restricted to landslide dams that have either failed catastrophically or have not failed. In these studies, stability is a momentary definition of the state of the dam and the landslidedammed lake at the time of observation but does not include the duration of non-failure (longevity). Per that definition, a landslide dam is termed stable if a landslide-dammed lake still exists or is filled in with sediments at the time of the analysis (Table 3). The 8
Journal Pre-proof latter indicates that the dam was able to retain the lake water (maintaining in- and outflow balance by seepage through the dam or flow along a spillway) and allowed continued sedimentation into the lake until the lake was silted up. Those classified ‘not stable’ have breached catastrophically, evidenced by deep gullies, an impoundment devoid of significant sediments, and/or with tell-tale erosional features within left-over sediments that indicate rapid drawdown, and/or flood sediments downstream – although evidence of outburst floods are often superimposed or eroded by later valley
f
processes. We have recognized that a deep gulley cut through a dam, or bypassing a
oo
dam, can also be formed by long-term gradual erosion as well – many features are not
pr
conclusively diagnostic for sudden failure or for slower erosion. Between ’stable’ and
e-
‘unstable’ exists a large number of cases where the cause of lake drainage (catastrophic or gradual) remains unknown.
Pr
Table 2 Terms characterizing landslide dam stability Casagli and Ermini (1999)
Artificially controlled Slow erosion
Existing
Includes all the lakes extinguished due to progressive filling Includes lakes existing at present
Jo
Filling
Includes all the cases of a catastrophic dam collapse. Includes all the cases where the remedial works drastically changed the natural evolution of the phenomenon. Includes all the lakes extinguished because of slow progressive erosion of the dam without catastrophic events.
ur n
Collapsed dam
al
Includes all the cases of partial blockage, channel deviation, erosion of the landslide toe, in which the formation of a complete dam is prevented.
Dam not formed*
Ermini and Casagli (2003) Unstable dam Stable dam:
Landslide dam that has undergone erosion or collapse leading to a catastrophic breach, with the subsequent release of the impounded lake waters. Landslide dam that has remained stable and has not encountered a catastrophic breach thus still impounding an existent or relict lake.
Korup (2004) Stable The landslide-dammed lake persists for over a decade. Tacconi Stefanelli et al. (2016) The landslide reached the riverbed, but only a partial damming is realized (no lake Not-formed* formed). Over times that can range from hours to centuries, the dam collapsed or was Formed-unstable breached by the river, including the artificially breached or removed landslide dams. The blockage is complete with the formation of a dam and a lake, which are still Formed-stable existing or disappeared for sediment filling. The dam could have been overtopped during its life, but no total failure or destructive flooding wave occurred. Note: *Even though these landslides did not create full river blockages, they are considered by the respective authors, and included herein, since they may still alter the ri ver flow and sediment delivery or form partial blockages, 9
Journal Pre-proof which could be of relevance in risk assessment and is of interest for inventory analyses (i.e. factors relevant to the full dam formation and failure statistics).
A potential history of multiple breaches or predictions of future stability (or longevity) is rarely taken into hazard analyses (Hermanns et al., 2013). For a landslide dam to remain stable, it needs to be able to hold the impounded water indefinitely and/or release it gradually so that no catastrophic flood may impact the downstream area. From a practical point of view, and for a first-order assessment of a dam’s potential
f
threat, we propose the use of ‘apparent’ stability (Table 3), which applies to any dam
oo
that still holds a lake, but which either has not been investigated yet or has been stable
pr
for a long enough time for any potential risks (e.g. due to external causes) to be forgotten (all red categories in Table 3). The only proven ‘stable’ dams are those in the
e-
category ‘infilled’ since these demonstrably maintained a lake without catastrophic
Pr
breaching. This classification scheme furthermore takes into consideration the potential for partial and episodic breaching. It should also prompt further investigations of the
al
landslide dam in question, and assist in communication between scientific disciplines
ur n
and stakeholders. Furthermore, it should minimize the risk of inferring long-term stability in situations that are not fully investigated.
Jo
Longevity in this study refers simply to how long a landslide dam has existed before its lake emptied, regardless of the cause (catastrophic breach or gradual erosion). Stability, on the other hand, relates to whether a dam is prone to sudden collapse—in the landslide dam literature, this is mainly applied to the momentary situation wherein dams have either failed or not at the time of data acquisition and publication. The time-scales before/until failure and those at which the effects of a landslide dam are noticeable in the landscape system are classed as follows in this paper. Short -term refers to the first days to months after formation, since this is the most crucial for hazard assessment. Most dams that have failed catastrophically did so within the first year after 10
Journal Pre-proof formation (Clague and Evans, 1994; Costa and Schuster, 1988; Fan et al., 2012c; Peng and Zhang, 2012); hence we consider up to one year as short-term. It is also the time of the most pronounced landscape impacts by the blockage itself, disrupting river flow, and the filling of the impoundment creating a new landslide-dammed lake. Medium-term encompasses years to decades, a time where mitigation measures might still be necessary, a landslide-dammed lake might still exist and catastrophic breaching could still be possible (e.g. through external causes like earthquakes or strong/prolonged
oo
f
rainfall). Long-term involves centuries to millennia, a time-scale in which the dam (or remnant) can still influence the landscape, but in which, from a human lifespan point of
pr
view, hazards are either diminished by erosion or, if a landslide-dammed lake still
e-
exists, vanished from the collective consciousness. Database analyses often, by necessity, deviate from these first-order timelines to either investigate processes at
al
Pr
higher time-resolution or because data availability determines the intervals for analysis.
Table 3 Classification scheme of apparent landslide dam stability
ur n
Infilled
Jo
Long-term stable
Long-term eroded dam
Short-term stable
Short-term eroded dam Dam breached
Impoundment infilled with sediment (i.e. retained landslide-dammed lake throughout its lifetime) and there is no evidence for any catastrophic breaching during its existence. Landslide dam that has remained stable and has not encountered a breach over centuries or longer and thus still impounds an existent or relict lake and no immediate change is expected. However, extreme weather / climatic conditions or other effects (e.g. landsliding into the landslide-dammed lake or unusually high inflow) might cause catastrophic failure. Entirely/partially eroded, i.e. did not retain the impoundment lake waters, but released it gradually over a hundred to thousands of years. Existing landslide-dammed lake with no sign of partial breaching, however, there is a high threat for catastrophic failure in the near future (days to months) based on the material of the dam and/or geomorphic parameters. Entirely/partially eroded, i.e. did not retain the landslide-dammed lake waters, but released it gradually in days to months. Breached (or partially) with evidence of catastrophic outburst flood(s) – includes internal and external causes for breaching (e.g. seepage erosion or erosion caused by overtopping)
Note: blue categories are dams where no remediation is needed; red categories advice hazard assessment and monitoring. 11
Journal Pre-proof 2.2
Geomorphometric landslide dam classification
The dams discussed in this paper are mainly formed by slides, falls, and flows according to the Varnes (1978) classification as modified by Hungr et al. (2014). A landslide dam, therefore, is the result of a landslide of any form whose runout and interaction with local topography allows its deposits to form a barrier that can hold back water above ‘normal’ river level. The shape and size of the landslide dam in relation to the size of the blocked valley is
oo
f
one of the most widely accepted morphological classifications of landslide dams and was proposed by Costa and Schuster (1988). It is based on 225 world-wide case
pr
studies (Table 4) and defines six classes. Hermanns et al. (2011b) modified it and
e-
added four further classes, also listed in Table 4; however, these refer to very specific
Pr
and rare cases at river confluences and drainage divides. Table 4 Geomorphometric landslide dam classifications.
Description
Type (Costa and Schuster, 1988)
Type (Hermanns et al., 2011b)
Small dams relative to the valley size that does not reach the opposite slope
I
II c
Large dams, but still small compared to the valley dimensions. This is one of the most widespread types , and they are frequently formed by rotational or translational landslides
II
II a
Very large dams that fill the valley from side to side. The collapsed materials are distributed both upstream and downstream from the release area, at times with a Tshaped deposit. Casagli and Ermini (1999) introduced a distinction in this class, separating the phenomena in which the dam also affects the tributary valleys of the main one (III a) from those in which these are not involved (III b).
III
IV a
Jo
ur n
al
Sketch
12
Journal Pre-proof
IV
II b
Several dams are formed by multiple tongues of a singular landslide. They are not very common for landslides but may form by glaciers (glacier tongues).
V
III b
Landslides with sub-channel rupture surfaces.
VI
II d
-
I
-
III a
Type of a landslide dam affecting water divides; with the landslide deposit depositing directly in the drainage divide.
-
Va
Single landslide dam having different forms distinguished between a rockslide crossing the valley partially or entirely, a rockslide involving the valley floor, and rockslides which, after impacting the valley bottom, spread in an up- and down valley direction (almost the same as III of Costa and Schuster).
-
II e
More complex multi-river geometries, focusing on the morphological relations of the dam in relation to the river valley
-
IV b-e, VII a-b, VIII, IX
pr
oo
f
Those very rare dams are formed from the contemporary movement of two landslides detached from opposite sides of the same valley.
Pr
e-
Supra-landslide lakes on landslide deposits are frequent on larger rockslide deposits , but can occur on all types of landslides. They are not connected to located within the channel of the dammed river.
Jo
ur n
al
Sub-type of plan view distribution of multiple landslide dams in a single valley in a line formed by a single landslide. Another type is III b.
-
13
Journal Pre-proof Supra-landslide lakes (Type-I of the Hermanns et al. (2011b) classification (Table 4)) are very frequent phenomena and their deposits often help to date prehistoric events. However, those water bodies are, with exceptions (i.e. Los Rojos supra-landslide lake— a large pond on top of a landslide deposit in North Patagonia, Hermanns et al., 2011a), comparatively small in size and rarely represent any threat to society. Hereafter, stability considerations for the small supra-landslide lakes are not considered.
Compositions and origins of river-damming landslides
oo
f
2.3
Dam geometry and composition strongly depend on the type of landslide forming the
pr
blockage. In the following, we briefly outline the types, materials and origins of
e-
landslides forming dams. Whereas all the river-damming landslides reviewed belong to
Pr
the slide and flow movement categories in the classifications by (Varnes, 1978) and Hungr et al. (2014), further attributes such as size and saturation with water are taken
al
into account herein since they play important roles in debris spreading and hence in the
ur n
final dam plan geometry, thickness, and erodibility. We used elements of the classifications of Cruden and Varnes (1996) and Hungr et al. (2014), but added volcanic
Jo
categories, because volcanic landslides and their unique materials (e.g. pyroclastics or wide variety of components and grain sizes) are important enough for the formation of landslide dams to include separate mention. Therefore, the following adapts classifications to include landslide size, degree of saturation, and origins.In the prevalent literature, the landslides forming dams cluster into: (1) large rockslides and rock avalanches (volumes >10 6 m3) composed of crushed rock with broad grain size distributions (Figure 1a), (2) volcanic mass movements (volcanic debris avalanches and pyroclastic flows; Figure 1b-c), (3) debris flows (saturated flows, often emplaced in pulses, typically thinner than slides/flows in sediment) (Figure 1d), (4) slides and flows 14
Journal Pre-proof in unconsolidated sediments (various sizes, unsaturated) (Figure 1e), (5) debris avalanches (unsaturated mixtures of (variably) rock, soil, sediment, vegetation), and (6) complex landslides, where one type of landslide transforms into another during
Jo
ur n
al
Pr
e-
pr
oo
f
emplacement.
Figure 1 (a) Aerial view of the Usoi landslide dam and Lake Sarez (photo: Mergili, M. (2019): The world in images. Digital media (https://www.mergili.at/worldimages)). (b) Sentinel 2 image (RGB composite image created with 843 bands, 10 m in resolution, from GloVis, USGS) showing the distribution of the debris avalanche deposit originated from the 1888 collapse of the Bandai volcano (Japan). The deposit blocked several branches of the Nagase River, forming three main lakes. (c) View from the southwest toward El 15
Journal Pre-proof Chichón Volcano and the Tuspac Valley. On 4 April, block-and-ash flows rushed along the Agua Tibia and Tuspac gullies, coalescing in the Magdalena River to form a dam 25– 30 m in maximum height. The lake breakout occurred on 26 May 1982, originating hot lahars up to 35 km of distance. (Picture courtesy of R. Mota taken on 24 April 1982). Notice the steaming pyroclastic flow deposits at the surface of the hot lake. (d) The dam formed by a debris flow at Brenot Creek in British Columbia. (e) Landslide dam at Wrigley Creek sediment type --- matrix-supported diamicton having sub-rounded to angular clasts at Wrigley Creek in Northwest Territories, Canada occurred probably after 1940 (Jermyn and Geertsema, 2015). Photos in (d) and (e) taken by Marten Geertsema.
2.3.1 Dams formed by large rockslides and rock avalanches
oo
f
Large, long-runout rock avalanches (RA) and rockslides (RS) are specific types of extremely rapid (m/s) landslides which arise from sudden failure of high rock slopes
pr
with failure volumes exceeding ~10 6 m3 (e.g. Davies and McSaveney, 2011; Davies and
e-
McSaveney, 2012; Erismann and Abele, 2013; Heim, 1932; Hewitt et al., 2008; Scheidegger, 1973). For rock avalanches below failure volumes of ~106 m3, the
Pr
characteristic long runout, flow-like motion, and intense fragmentation are commonly
al
absent. Both rock avalanches and rockslides are sourced in (relatively) intact bedrock
ur n
but differ in their spreading behaviour. The former break up into highly fragmented flowtype movements (“Sturzstrom”, Heim, 1932) and often spread radially (where unconfined) into sheets of dispersed surface features (ridges, lobes and hummocks),
Jo
whereas the latter slide, often in a roughly linear trajectory, as more coherent ‘blocks’ of highly fragmented debris (e.g. Erismann and Abele, 2013; Gruber et al., 2009; Hutchinson, 2006; Pollet et al., 2005). RA/RS dams have been reported with thicknesses up to 1,000 m (Hewitt, 2009b) and causing impoundments on the order of cubic kilometres (e.g. the 17 km3 Lake Sarez behind the Usoi dam; Alford and Schuster, 2000; Ischuk, 2011; Schuster and Alford, 2004) over tens of kilometres of valley length. The landslide dams formed by these events are often some of the most long-lived and the largest and are the most 16
Journal Pre-proof implicated in causing long-term disruptions to landscapes over a long time- and spacescales. 2.3.2 Dams formed by volcanic mass movements A comprehensive review on identification, stability and secondary effects of volcanic dams is given by Capra (2007). In their study, granulometric characteristics of formation and breach of volcanic dams were reviewed explicitly, and the stability of volcanic dams was analysed from a textural point of view. Volcanic eruptions, as well as lateral
oo
f
collapses, can produce gravity-driven granular flows, mobilizing large volumes (>106 m3) of debris. The most common events in causing river blockages are (1) volcanic
pr
debris avalanches (VDA) (Figure 1b), (2) pyroclastic density currents (e.g. Figure 1c),
e-
and (3) lahars.
Pr
Emplacement of VDAs from volcano lateral collapse commonly involves large volumes (>1 km3 ) and rapid emplacement (few minutes). A notable example is the 1980 Mt. St.
al
Helens eruption and lateral collapse with the emplacement of a 2.3 km3 VDA that
ur n
caused the formation of five main landslide dams along the North Fork Toutle River and raised the level of the Spirit Lake by 60 m (Meyer et al., 1986). These landslide dams
Jo
can be up to few hundred meters high (e.g. the 150-m high dam of Mt Spurr, Alaska, USA), impounding a volume of water in the order of few km 3 (e.g. 1 km3 lake volume at Volcán de Colima, México) (Table 5). Table 5 Examples of landslide dams in volcanic environments. Volcano
date
landslide type
18 May 1980
volcanic debris avalanche
Lake name Mt. St. Helens, USA Spirit Lake Elk Rock Lake Castle Lake Coldwater Lake Jackson Creek Lake Bandai, Japan
15 July 1888
volcanic debris 17
dam height [m]
lake volume [km 3]
time to failure
69 9 37 71 4.5
0.33 0.0003 0.024 0.83 0.0025
---unfailed--63 days ---unfailed-----unfailed--644 days
ref.
1
2
Journal Pre-proof
Iriga, Philippines Buhi Lake Nevado del Huila, Colombia Nevado de Colima, Mexico Volcán de Colima, Mexico
1628 AD 46,000-200,000 yrs BP 18,500 yrs BP 3,600 yrs BP
Parinacota, Chile
8,000 yrs BP
Hualca Hualca, Peru
Late Pleistocene
Mt. Bakening, Kamchatka Bezymyannoe Lake Verkhneavacha Lake Mt. Hood, USA Mt. Spurr, USA
8,000-8,500 yrs BP
1991 1982
lahar lahar lahar lahar and pyroclastic flows pyroclastic flow
---unfailed-----unfailed-----unfailed---
150
1.2
---unfailed---
3
30
0.1224
---unfailed---
4
60
0.5
> 3 days
5
80
1.0
> 1 year
6
60
0.4
> 11 days
7
40
0.4
---unfailed---
8
-
-
-
9
---unfailed-----unfailed--12 minutes
10
10 60 20
0.45 0.1-0.01
24
0.075
< 1 month
12
55
0.04
2 months
13
e-
Pinatubo, Philippine Mapanuepe Lake Chichón, Mexico
25 Dec. 1980 Late Holocene 1953-1992
volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche
0.45 0.6 0.165
f
10,000-38,000 yrs BP
Mt. Spurr, USA
30 30 30
oo
avalanche
pr
Hiab ara Lake Onogana Lake Akimoto Lake
11 3
al
Pr
Referenc es: 1) Glick en and Nak amura (1988) and Nak amura and Glick en (1988); 2) S wans on et al. (1982); 3) Waythomas and Nye (2001); 4) Lagmay et al. (2000); 5) Pulgarín et al. (2004); 6) Capra and Macıas (2002) ; 7) Cortés (2002) and Cortés et al. (2010); 8) Clavero et al. (2004); 9) Gómez-Avalos et al. (2004); 10) Melek estsev et al. (1999); 11) Gallino and Pierson (1985); 12) Umbal and Rodolfo (1996); 13) Macías et al. (2004).
ur n
Pyroclastic flows and lahars can also form landslide dams, but those are generally short-lived. All the known cases failed a few days to months after their formation,
Jo
possibly because of their smaller size and higher erodibility (Carey et al., 1996; Gallino and Pierson, 1985; Macías et al., 2004; Schuster and Crandell, 1984; Umbal and Rodolfo, 1996; Waythomas and Kaufman, 1991; Waythomas and Nye, 2001). 2.3.3 Dams formed by debris flows Hungr et al. (2001) define a debris flow as ‘very rapid to the extremely rapid flow of saturated non-plastic debris in a steep channel’. Debris flow volumes typically range from several m3 to ~107 m3, whereby lahars and combined rock avalanche -debris flows can reach volumes up to ~10 9 m3 (Fort, 1987; Fort and Peulvast, 1995; Griswold and 18
Journal Pre-proof Iverson, 2008; Iverson, 1997; Rickenmann, 2005). Debris flows often grow during runout through entrainment of water and sediment from the runout path (Frank et al., 2015). The rheology can vary substantially within a single debris flow event, with surgelike behaviour common in a single ‘event’, and the resulting deposits are poorly-sorted mixtures of grain sizes ranging from clay-sized to large boulders (Hungr et al., 2001; Iverson, 1997). The fluid nature promotes spreading and thinning once unconstrained. As a result, debris flow dams, where not topographically confined, form quite low dams
oo
f
that often fan out from tributary valleys into the main channel network. Debris-flow dams (Figure 1d) are often part of a geohazard cascade (Gallino and
pr
Pierson, 1985; Wei et al., 2018; Yin et al., 2016) or process chain, related to the need
e-
for a debris source that can be transformed into a slurry-like flow. Multiple debris flow dams can become chronic problems (Fan et al., 2019a) until the debris sources are
Pr
exhausted, with a clear correlation to the temporal patterns of rainfall that both trigger
al
debris flows and increase river flow to dammed reaches.
ur n
2.3.4 Dams formed by slides/flows in unconsolidated sediments Landslides which are neither dominated by granular rock texture (2.3.1) nor are fully
Jo
water-saturated (2.3.3), yet can be of large volume, are those formed in unconsolidated sediments (e.g. till; Figure 1e). Examples include the glacial and associated water-lain sediments, common in former glaciated terrains (Geertsema, 1998; R. L. Hermanns et al., 2012; Reginald Leonhard Hermanns et al., 2012; Locat et al., 2017) which often produce landslide dams (Clague and Evans, 1994; Geertsema et al., 2018; Geertsema and Clague, 2006; Geertsema et al., 2006a; Geertsema and Cruden, 2008; Geertsema et al., 2006b; Geertsema et al., 2009; Geertsema et al., 2006c; Lefebvre et al., 1991; Locat et al., 2017; Miller et al., 2018; Miller and Cruden, 2002). One of the most infamous of these is the 2014 Oso landslide in Washington State, USA, which ran over 19
Journal Pre-proof a rural residential area and killed more tha n 40 people (Iverson et al., 2015; Wartman et al., 2016). Rupture surfaces can daylight far above stream levels (Evans et al., 1996; Fletcher et al., 2002; Geertsema, 1998), but they may also occur near or below river channels (Cruden et al., 1993; 1997; Miller et al., 2018; Miller and Cruden, 2002). Those that are located below the channels will form Type VI dams (Costa and Schuster, 1988), which can be impervious. For these latter dams, the channel alluvium may be elevated almost
oo
f
vertically (Miller et al., 2018; Miller and Cruden, 2002). Upon overtopping of the Type VI dams, an incision into the sediment comprising the dams can be rapid; however, where
pr
the post-landslide stream encounters the pre-landslide alluvium, incision can be
e-
considerably delayed.
Pr
2.3.5 Debris avalanches and complex landslides Debris avalanches can occur anywhere on a slope and are not confined to an
al
established channel; which debris flows commonly are Hungr et al. (2014). They may
ur n
initiate directly on the slope or be triggered by the impact from a rockfall or rockslide if liquefiable material exists downslope from these other landslides (Hungr et al., 2014).
Jo
Or they may be a consequence of erosion and entrainment of soil and sediment into another landslide type. Where one landslide transforms into or triggers another during emplacement, the term ‘complex landslide’ (Cruden and Varnes (1996) applies. The latter include ‘rockslide-debris avalanches’, ‘rotational slide-earth flow’, ‘rock avalanchedebris flow’, and many more combinations.
20
Journal Pre-proof 2.4
Dam failure susceptibility
The material and sedimentological structures of the dams are not taken into account in geomorphometric
classification
schemes,
even
though
the
resulting
variable
geotechnical properties play an important role in dam stability (Casagli et al., 2003; Dunning et al., 2005). Therefore, Fan et al. (2017) proposed a three-fold classification scheme based on material and internal geological structure from an analysis of 32 landslide dams resulting from the 2008 Wenchuan earthquake (Table 6).
Jo
ur n
al
Pr
e-
pr
oo
f
Table 6 Landslide dam classification based on dam composition and internal structures (modified from Fan et al., 2017).
The classifications by Costa and Schuster (1988) and Fan et al. (2017) can be combined to obtain a preliminary indication of dam stability by considering the topographic setting and the sedimentological features of the valley-landslide system. 21
Journal Pre-proof Since dam types I to III of Costa and Schuster (1988) taken together constitute between 84% and 96% of known dams (Figure 2), most landslide dams can be assigned to
Pr
e-
pr
oo
f
these classes.
ur n
al
Figure 2 Landslide dam classification (Type I to VI) according to Costa and Schuster (1988) (% of landslide dams in each class) for previous inventories: moving from the outer ring to inner ring: (a) Worldwide - 225 cases Costa and Schuster (1988); (b) 19 cases by Ermini (2000); (c) Italy - 300 cases Tacconi Stefanelli et al. (2015); (d) Cordillera Blanca, Peru - 51 cases Tacconi Stefanelli et al. (2018).
Table 7 is an example of how, by combining two classifications methods , it could be
Jo
possible to provide a preliminary assessment of landslide-dam stability a very short time (days) after their formation that is largely independent of landslide type. For a rapid initial assessment of a dam’s failure probability, a three-class scale can be used: green for apparently stable, orange for stable in the short-term but unstable in the medium- to long-term, and red for potential instability in the short-term (immediate mitigation advised). Stability of landslide dams is also affected by many other factors, which cannot be directly considered in the initial, rapid assessment, 2-dimensional matrix solution in Table 7. Other factors such as catchment area, dam height and volume, lake volume, discharge, and erosive power can positively or negatively be assigned 22
Journal Pre-proof accordingly within such a stability matrix. For example, dams with small catchment areas/inflow, and large heights tend to be more stable; whereas dams with a large volume of impoundment, in a river of high discharge and high erosive power may be less stable and vice versa. Landslide triggering mechanisms are not considered in this assessment. Apparently stable dams below very active catchments or those situated in active
seismic
zones
(especially
during
an
aftershock
sequence
post-dam
emplacement) need to be considered more likely to breach. We emphasize that Table 7
oo
f
is for a rapid assessment of landslide dams within a large number of cases, but it cannot satisfy all conditions—at this stage, it is not a conclusive suggestion, rather a
pr
first attempt at the design of more comprehensive hazard levels.
Pr
e-
Table 7 Preliminary landslide dams stability classification matrix combining Costa and Schuster (1988) and Fan et al. (2017) methods. Green is for basically stable, orange for the stable in short time (but unstable in the medium- to long-term), red for unstable in short time. Fan et al. (2017)
I
Type
Sub-type B
Sub-type C
Small dam blocking valley composed by loose fine materials, formed by debris flows, avalanches, and slides with long long-runout
Small dam blocking valley formed by large boulders and blocks, originated by rock avalanches or rock falls
Small dam blocking valley composed by intact rock strata topped by coarse materials, formed by deepseated rock slides or avalanches
Large dam able to block the valley composed by loose fine materials, formed by debris flows, avalanches, and slides with long long-runout
Large dam able to block the valley formed by large boulders and blocks, originated by rock avalanches or rock falls
Large dam able to block the valley composed by intact rock strata topped by coarse materials, formed by deep-seated rock slides
Very large dam that fills the valley composed by loose fine materials, formed by debris flows, avalanches, and slides with long long-runout
Very large dam that fills the valley formed by large boulders and blocks, originated by rock avalanches or rock falls
Very large dam that fills the valley composed by intact rock strata topped by coarse materials, formed by deep-seated rock slides
al
Type
Sub-type A
ur n
Type
Jo
Schuster (1988)
II
Type III
23
Journal Pre-proof
3 Worldwide landslide dam distribution and database compilation 3.1
Problems of current inventories
Current datasets and inventories of landslide dams encompass dams formed by the full variety of landslides, i.e. from small, slow-moving landslides to water-saturated debris flows and to large (106 – 109 m3 ) catastrophic failures of entire rock or volcano slopes. Material types range from soil and debris through to variably broken rock, and can include any mixture of those materials plus vegetation and/or water /ice. The landslide
oo
f
deposits show the variability of their morphometric characteristics (e.g. plan area,
pr
height, width, volume), and differ in geotechnical parameters (e.g. permeability, grain size distributions, water content, armo uring, erodibility). Many of these parameters are
e-
difficult to obtain before the dam is cut by erosion or a breach. A comprehensive
Pr
database allows for sensible transferability of characteristics from those that have
al
breached, to those that have just formed--within error bounds. Landslide (dam) inventories are: (1) typically geomorphological inventory maps
ur n
containing little geotechnical information, (2) event-based, i.e. including all the slope movement types that occurred after a particular triggering event (e.g. earthquake,
Jo
rainstorm), or (3) historical and regional, consisting of all landslides (dams) in a specific area going back in time tens to even tens of thousands of years (Malamud et al., 2004). The available datasets furthermore differ in two critical aspects: (1) quality/resolution of data sampling and (2) age of the events (Tacconi Stefanelli et al., 2015). The morphometric information for each inventory is generally extracted from a wide range of remote sensing data (satellite imagery, aerial photos, and digital terrain models), possibly with different spatial resolution and with limited ground-truthing. This often leads to measurement errors of morphometric parameters , and comparisons between 24
Journal Pre-proof datasets (and even individual events within a given dataset) need to consider this level of inaccuracy. Many cases in the historical inventories are not ‘recent’ events and, generally, the farther the case lies in the past, the less reliable the data is due to deposit modifications over time (erosion, vegetation, consecutive landslides, and anthropogenic modifications; even more so for smaller sizes). However, the estimate of ‘stability’ is likely better when based on long-term evidence than for rapidly collected event inventories. The direct consequence of this is a general underestimation of the
oo
f
morphometric parameters of the landslide dam body (i.e. height, length, width and volume), difficulties in landslide-dam reconstruction for any back-analyses of stability,
pr
or, biased preservation of the largest events that resist long-term removal (Sanhueza-
e-
Pino et al., 2011). Therefore, we take the first attempt to compile the existing
(see supplementary material).
Worldwide landslide dam database
al
3.2
Pr
inventories and datasets and create templates and standards for further data collections
ur n
We use the term ‘inventory’ for regional, world-wide or event-based catalogues; ‘dataset’ for smaller collections, and ‘database’ for the catalogue compiled within this
Jo
research since the goal of our data collection efforts is to make the resulting database accessible, searchable and open to new records in the future. Table 8 contains the published landslide dam inventories from which our database was compiled. They are historical, geological and event-based inventories. Those in bold are the most comprehensive in terms of the required key information, rather than simply number of events, for each geographical region (see Section 4.3 and supplementary material for the required key information). Table 8 Summary of reviewed landslide dam inventories and datasets. In bold letters, the most important/recent ones. Blue shaded rows are the reviewed datasets from which 25
Journal Pre-proof individual records were taken to the new database. In brackets of the blue shaded rows are the numbers of landslide dams selected for the new database compilation. Region, Country
N. of case s
Information on the inventory
Historical or event-based
Source or reference
1
world
187
Conclude the failure and lasting time.
Historical
Schuster (1986)
2
world
13
Most of these dams have been caused by relatively high-velocity debris flows originating from tributary valleys to briefly block rivers in main valleys. Generally, dams formed in this manner are neither high, nor are they particularly resistant to erosion.
Historical
Costa and Schuster (1988)
3
world
463
Including location, triggering mechanism, landslide type , size, failure time and mechanism, etc.
Historical
Costa and Schuster (1991)
Historical
[~100]
f
No.
world
18
Using a model to predict peak discharge of floods caused by the failure of natural landslide dams and artificial dams.
5
world
350
Most of the collected cases are from the western USA, Japan, and Italy. A common feature is a high frequency of damming events along or near the active margin areas of tectonic plates .
Historical
Ermini and Casagli (2003)
6
world
18
Between years 1840-2010, dams having impoundment volume exceeding 100 million m 3.
Historical
Evans et al. (2011a)
7
world
1239 [290]
Based on records of 52 landslide dam cases with breaching information in the database, empirical models for estimating the breaching parameters of landslide dams are developed.
Historical / Eventbased
Peng and Zhang (2012)
8
Argentina (Patagonia)
41
Overview of prehistoric rockslide dams, with geomorphic parameters extracted from GIS analysis.
Geological
Hermanns et al. (2011a)
9
Argentina (Central Andes)
20
Overview of prehistoric rockslide dams, with geomorphic parameters extracted from GIS analysis.
Geological
Hermanns et al. (2011b)
10
Canada
38
16 existing and 22 landslide dams.
historical
Historical
Clague and Evans (1994)
11
Canada
8
Geomorphic parameters and stability indices for recent landslide dams in glaciolacustrine sediments.
Historical
Miller et al. (2018)
12
Central Asia
190
Include location, state of the dammed lake (existing, silted, breached), data on the size and type of the 190 river-damming landslides exceeding 1 million m3 in volume for which parameters of the dammed lakes were indicated. The actual number of past rockslide dams is much larger.
Prehistoric mainly and several historical events
Strom and Abdrakhmatov (2018)
13
China
147
Including name, location, date,
Historical
Chai et al. (1995)
Jo
ur n
al
Pr
e-
pr
oo
4
26
Walder and O'Connor (1997)
Journal Pre-proof landslide volume, blocked river, last period, triggering factors and impacts. China
33
All are caused by the Wenchuan earthquake. Risks of dam collapsing and estimated collapse mechanism are given in the table.
Event-based
Yin et al. (2009)
15
China
257
All triggered by M8.0 Wenchuan earthquake.
Event-based
Cui et al. (2009)
16
China
256
Because of limited access, relatively detailed data are available for only 32. The paper presents statistical analyses of the distribution, classification, characteristics, and hazard evaluation of these 32 dams.
Event-based
Xu et al. (2009)
17
China
828 [20]
All in the Longmenshan mountains, China, triggered by the M w 7.9 2008 Wenchuan earthquake. It is the largest of its kind attributable to a single regional-scale triggering event.
Eventbased
Fan et al. (2012c)
18
Karakoram Himalaya
322
In surve ys between 1993 and 2009, in the Karakoram and Hindu Raj Mountains, a total of 322 Hewitt (2004); Hewitt (2006); rock avalanches were identified; this total has been increased to more than 370 in recent work (Hewitt (2011). 65% dammed rivers of the Upper Indus (all are rock avalanches).
Geological
19
Italy, Apennine
70
Several geomorphic parameters have been measured, during field investigations and through air photo interpretation, or estimated, using historical and bibliographic information.
Event-based
Casagli and Ermini (1999)
20
Italy, Northern Apennines
68
Including grain size distribution of dammed lakes only.
Historical
Casagli et al. (2003)
21
Italy
300 [49]
Several geomorphic parameters have been measured, during field investigations and through air photo interpretation, or estimated, using historical and bibliographic information.
Historical
Tacconi Stefanelli et al. (2015)
22
Japan
43
This study utilizes 43 landslide dams to determine the significant variables, including logtransformed peak flow (or catchment area), and logtransformed dam height, width, and length in hierarchical order, which affect the stability of a landslide dam.
Historical
Dong et al. (2009)
23
New Zealand
21
Eleven small lakes were formed by landslides caused by the 1929 Buller earthquake; four others were formed by other historical earthquakes in New Zealand. At least nine other New Zealand lakes are also dammed by landslides and were probably formed by prehistoric earthquakes.
Historical / Event-based
Adams (1981)
Hewitt (2011)
Jo
ur n
al
Pr
e-
pr
oo
f
14
27
Journal Pre-proof New Zealand
232
Geomorphometric data on landslide dams, associated lakes, and contributing catchment characteristics
Historical
Korup (2004)
25
New Zealand
39
Geomorphic parameters, landslide types and features of very large landslides in Fiordland. Only landslides with volumes greater 6 3 than 10 m are listed. Comparison with 11 other giant landslides from around the world.
Historical / Event-based
Hancox and Perrin (2009)
26
New Zealand
240
More than half of all recorded landslide dams in New Zealand were formed by large rock-slope failures, with most of them occurring in the Southern Alps and adjacent Fiordland. Almost 30% of all recorded landslide dams in New Zealand were formed by rock avalanches
Historical
Korup (2011)
27
Peru
51 [7]
Geomorphic parameters have been measured, during field investigations and through air photo interpretation.
28
Venezuela
35
Overview of historic and prehistoric rockslide dams in the Andes of Venezuela.
29
Switzerland
35
List the natural lakes in Switzerland, which cover an area above approximately 0.5 km 2.
30
Taiwan, China
17
31
Turkey
32
33
oo
f
24
pr
Historical
Tacconi Stefanelli et al. (2018)
Ferrer (1999)
Historical
Evans et al. (2011a)
This paper gives the risk level, present condition, and its authority.
Eventbased
Chen et al. (2011)
3
Only one remains. All are induced by a rainfall event.
Eventbased
Duman (2009)
Nepal
*
The 25 April 2015 M w 7.8 Gorkha earthquake triggered many shallow failures with only a few landslide dams that were either very short-lived or only partially blocked the river course.
Eventbased
Dhital et al. (2016)
New Zealand
#
196 significant drainage blocking or drainage constricting landslides were identified after the 14 September 2017 M7.8 Kaikoura earthquake.
Eventbased
Massey et al. (2018)
ur n
Few
al
Pr
e-
Geological
Jo
196 [4]
Note: * Inventories reporting the landslide dams are not available. #Only 4 large landslide dams with > 106 m3 landslide volume were reported.
The inventories in Table 8 total >1,800 entries. In addition, another dataset from Strom and Abdrakhmatov (2018) has 1,000 pre-historic rockslides from the Central Asian region 549 of which had formed dams, though parameters were measured for 190 cases only. As the numbers of known landslide dams increases, we have found uncertainty enlarges and statistical representativeness of the dataset diminishes 28
Journal Pre-proof because of missing information. Therefore, to minimize uncertainties, our analysis focuses only on a sub-set taken from the entire database (>1 ,800 dams), comprising 410 large (i.e. more than 106 m3 in landslide volume) and recent (i.e. time after 1900) landslide dams with detailed information. From hazard perspectives, more attention needs to be paid to these recent, large dams. Evans et al. (2011a) also considered rockslide dams with volume in excess of 106 m3 . These can be potentially dangerous to the downstream population considering the potentially large volume of water present in
oo
f
the impoundment. The distribution of the existing landslide dam inventories and the 410
Jo
ur n
al
Pr
e-
pr
landslide dams is shown in Figure 3.
29
Journal Pre-proof
f o
l a n
o r p
e
r P
r u o
J
Figure 3 Worldwide distributions of existing landslide inventories and recent large landslide dams in different regions. The yellow stars are landslide dam locations; the circles show the order sourced from Table 7, and the numbers inside brackets are the total number of landslide dams. The background is modified from Amante and Eakins (2009) . 30
Journal Pre-proof The distribution in Figure 3 seems to suggest that most landslide dams are located the ice-free areas of Asia, and many are in North America, Europe, Australia, and South America, and that next to none have been reported on the African continent. However, strong biases are introduced by (a) researchers’ regional preferences and citizenship, (b) single events with many landslides (such as the 2008 Wenchuan earthquake) skewing the distribution, and (c) our bias towards large landslide dams. However, geologically this distribution makes sense . In Asia, affected by the plate collision
oo
f
between the Indian Ocean and the Eurasian Plates, the Himalayas and surrounding areas suffer from large earthquakes due to intense tectonic movements (Blöthe et al.,
pr
2015; Gong et al., 2018; Kirby et al., 2000; Kirby et al., 2003). Earthquake-induced
e-
landslide dams are distributed along collision areas and subduction zones, such as the
and Nepal; Japan; Central Asia.
Pr
Tibetan Plateau bordering Western China; Himalayan ranges bordering India, Pakistan
al
Taking China as an example, the historical landslide dams are predominantly found in
ur n
Sichuan Province, where the destructive 2008 Wenchuan earthquake occurred (Shen et al., 2009; Yi et al., 2008; Yin et al., 2009). Landslide dams are widely distributed in
Jo
Japan and are well documented by Swanson et al. (1986). Most of them were triggered by earthquakes and rainfall, while some cases triggered by snowmelt are also reported (Costa and Schuster, 1991; Dong et al., 2009; Koi et al., 2008; Tabata et al., 2002; Wang et al., 2019). In North America, landslide dams are mainly concentrated in the west of Canada and the United States, close to the Pacific Ocean (Clague and Evans, 1994). With the influence of marine climate, rainfall is the most prominent triggering factor here. In Europe, landslide dams are abundant in western Italy, Switzerland, and Norway. The European Alps, Norway, and the Apennines in Italy are characterized by wide geo31
Journal Pre-proof tectonic and climatic variability where landslide dams and related flooding are frequent (Bonnard, 2006; Casagli and Ermini, 1999; Guzzetti and Tonelli, 2004; Salvati et al., 2010; Tacconi Stefanelli et al., 2015). In New Zealand, located in a deforming plate boundary zone (Adams, 1980; Kingma, 1959) where steady tectonic deformation is occurring (King, 2000), many historical landslide dams are oft-assumed triggered by earthquakes. For around 59 % of the 232 landslide dams examined by Korup (2004), the trigger mechanism remains unexplained, whereas 29 % are believed to have
oo
f
formed by earthquakes and another 11 % are tentatively believed co-seismic. The very strong rainfall typical for the Southern Alps of New Zealand could account for some of
pr
the unknown triggers (Henderson and Thompson, 1999). In South America, influenced
e-
by tectonics, volcanism, and climate change (Carey et al., 2012; Schauwecker et al., 2014; Worni et al., 2014), landslide dams are prominent along the Andes in Peru,
Pr
Argentina, and Venezuela (Ferrer, 1999; Hermanns et al., 2011a; Tacconi Stefanelli et
al
al., 2018). The Andes range, a part of the continental divide, is tectonically active
plates.
Recommendations for data collection
Jo
3.3
ur n
through the collision of the Nazca, South American, and a part of Antarctic lithospheric
To ensure databases are as complete as possible, it is important to standardize how key variables are collected for each landslide dam. Taking how dam height is measured as a simple example: dam surfaces are typically highly irregular, and it is not always clear whether the maximum or the minimum (where water may flow) elevation is reported in the literature. Ideally, all the information should be recorded (for example see Table 2 of Korup (2002), but in practice, few papers record all, either due to unavailable data or a lack of a well-established recording protocol. We recommend:
32
Journal Pre-proof (1)
Creating the inventory using geographic information system (GIS), which can show the accurate location linked with an attribute table to compile all the dam parameters, and it can also facilitate further geospatial and geomorphological analysis;
(2)
Creating metadata in order to ensure correct and proper use and interpretation of the inventory by its users; Using standard terminologies and providing clear definitions of the terms used;
(4)
Creating an Open Access portal in which collated records can be updated as
oo
f
(3)
new data are available, and wherein new events can be verified and added.
pr
We also create a data collection template (including over 60 diverse geological,
e-
geomorphological, geotechnical, and climate factors) for anyone to use and upload the
Pr
new collection of landslide dam inventories in the planned open-access portal. The template is updated after the versions of Casagli and Ermini (1999), Korup (2002) and
al
Tacconi Stefanelli et al. (2015), and is available as supplementary information. We
ur n
further provide an example of how to use the template by filling in information of a recent landslide dam event in eastern Tibet, China (Fan et al., 2019b) (see
Jo
supplementary information).
4 Formation and Stability 4.1
Controlling factors
In Figure 4, we compare the statistics of the triggers and types of river-damming landslides, movement and material, and the longevity of landslide dams based on the inventory from Costa and Schuster (1991) and the new landslide dam database from this study. About 100 cases from Costa and Schuster (1991) are included in the new database. The intention of this comparison is to visualise whether different sampling 33
Journal Pre-proof strategies with thresholds of size and age (as used in the new database) influence the results. Rainfall was the trigger for 42% and 44% of landslide dams, and the earthquake was the second major trigger for 29% and 20%, respectively in Costa and Schuster (1991) and in the new database (Figure 4a). Even though only large landslide dams are included in the datasets, the overall statistics do not differ much from Costa and Schuster (1991). The categories listed here are based on the classifications provided in each of the
oo
f
inventories used to compile our new database. As discussed in sections 2.3 and 3.3, terminologies are not always clear and classifications sometimes exclude specific
pr
phenomena or they were updated after an inventory was published. For example,
e-
‘complex’ is listed as a movement type in Varnes (1978) and as a modifier (e.g.
Pr
complex rock slide – debris flow) in Cruden and Varnes (1996), but is excluded in Hungr et al. (2014). Careful checking of each entry in the new database and updating
al
the classification schemes for landslides forming dams are two of the overarching
Jo
ur n
research goals of our ongoing database compilation and research.
34
Journal Pre-proof Figure 4 (a) Triggering mechanisms and (b) landslide, movement and material types of landslide dams compared between Costa and Schuster (1991) and our new database
In the inventories and datasets shown in Figure 5, of those landslide dams that did fail, about 80% did so within the first year after formation (Costa and Schuster, 1991; Fan et al., 2012c; Peng and Zhang, 2012). If a landslide dam survives the first year, it is statistically likely to survive for several decades or even centuries as it has found internal stability until some external influence, such as earthquakes, modifies that equilibrium. However, it is always unknown how many will remain stable and which
oo
f
ones will experience partial or complete breaching despite being classified as stable at the time of inclusion in any inventory. In most of the (historical) cases, the failure cause
pr
is unknown, and it is not possible to differentiate external (mainly earthquakes) or
e-
internal influences. The Kummersee rockslide dam in the Italian Alps that had blocked
Pr
the valley up to 150 m high is an example of multiple breaching episodes. It was formed in 1404 and experienced six partial failures until its final, catastrophic breach in 1774
al
(Ermini and Casagli, 2003). Other examples are the 1914 breach of the rockslide dam
ur n
in the Rio Barrancas Valley in Argentina that had existed at least 425 years before breaching (Hermanns et al., 2011a), the 1963 breach of the Issyk rockslide dam near
Jo
Almaty City in Kazakhstan (failure triggered by a debris flow that was produced by the breach of the Jarsay glacial lake upstream), and the 1966 breach of the Yashinkul rockslide dam in Kyrgyzstan (Strom and Abdrakhmatov, 2018). Both dams in Central Asia had existed for centuries, maybe even millennia before their catastrophic failure.
35
pr
oo
f
Journal Pre-proof
e-
Figure 5 Age of landslide dams at failure from different inventories; data modified from Fan et al. (2012c).
Pr
Understandably, these concerns drive much of the research, one of the major objectives of which is to understand the stability of the dam, to estimate when or if it is
al
likely to fail catastrophically. However, although the known population of dams can be
ur n
separated into those that have failed and those that have not, those in the latter category - may still fail in the future (e.g. Nash et al., 2008). Thus, while some large
Jo
dams seem unlikely ever to fail, based on either empirical or other investigations, the possibility of a catastrophe cannot be ignored completely. Most of the reviewed inventories and datasets (Table 7) are heterogeneous, as are the methods and biases of different scientific disciplines when it comes to identifying the factors controlling landslide dam stability. Key factors usually recorded include dam dimension, composition, and internal structures (post-erosion or breach), the river hydraulic gradient, inflow rate, and, as a proxy for inflow, the size of the upstream catchment area (e.g. Capra, 2011; Casagli and Ermini, 1999; Casagli et al., 2003; 36
Journal Pre-proof Dunning and Armitage, 2011; Dunning et al., 2007; Strom and Abdrakhmatov, 2018). Some data are readily available in the field or from remote sensing images, such as the dam dimensions, whereas other geotechnical parameters of the dam material and its internal structures, which control the stability, e.g. dam permeability, are rarely available (Casagli et al., 2003; Fan, 2013; Wang et al., 2016). As a result, some parameters affecting stability can be directly assessed from the landslide dam, whereas others must be inferred based on similarity to other landslides or landslide dams.
oo
f
In Table 9, we outline the characteristics that past works have deemed most important for landslide dam stability in each of their inventories. The apparent differences in
pr
emphasis highlight the need for continued, focused research, detailed observations,
e-
and comprehensive data collection to enable statistically significant analyses that go
Pr
beyond individual or regional case studies (cf. 3.1).
Table 9 Key stability-controlling variables used to infer (in) stability.
Worldwide and regional inventories of various landslide types
ur n
Landslide dam volume
Notes on the landslides that the referenced studies are based on
al
Factors identified as key in dam stability
Catchment; taken as proportional to river discharge
Landslides in unconsolidated sediments
References Dal Sasso et al. (2014); Ermini and Casagli (2003); Korup (2004); Tacconi Stefanelli et al. (2016) Geertsema and Clague (2008); Geertsema et al. (2006a); Miller et al. (2018) Capra (2006); Capra (2007); Capra (2011); Davies and McSaveney (2011)
Geometry, which influences the hydrologic gradient and location of the highest landslide dam point
Long-runout rockslides and rock avalanches
Korchevskiy et al. (2011); Strom (2006); Strom and Abdrakhmatov (2018)
Jo
Sedimentology/texture
Volcanic debris avalanches (in comparison to lahars and pyroclastic flows); long-runout rockslides and rock avalanches
4.2
Geomorphic indices
Geomorphic indices have been developed by various researchers to quantitatively assess the formation probability of a dam and predict its state once formed by employing morphometric parameters (Figure 6); most of which are easy and rapid to 37
Journal Pre-proof collect. The stability thresholds are then determined by analysing the geomorphic parameters in a bi- or multivariate way and graphically delineating domains in which ‘stable’ or ‘unstable’ (i.e. with existing or silted-up lakes or without lakes) dams exist
Jo
ur n
al
Pr
e-
pr
oo
f
exclusively (often with a large area of uncertainty between the two domains).
Figure 6 Cartoon illustrating the various geometric parameters of a dam and catchment
4.2.1 Dam formation indices Whether a landslide can form a complete dam when it reaches a river valley depends on several factors, including landslide velocity and deposit parameters (volume, 38
Journal Pre-proof thickness, and shape; i.e. resulting dam geometry), topography (valley width and shape), and hydrology (stream discharge and transport capacity) (Fan et al., 2014a; Fan et al., 2014b). Four indices are published that investigate dam formation probability (Table 10). The Annual Constriction Ratio (ACR) relates the width of the valley (Wv) with landslide velocity (v, here in m/a): ACR = v/Wv and was initially intended to estimate channel response (e.g. erosion of constricting material, channel diversion, lake
oo
f
formation) to delivery of material by slope processes (Swanson et al., 1986). Applied to an inventory of different landslide types from slow (1.5 m/a) to extremely rapid ones (>
pr
3m/s) in Japan, an ACR > 100 was assumed to be required to cause a lake to form
e-
(Swanson et al., 1986). Using an inventory of landslides in Italy, Tacconi Stefanelli et al.
Pr
(2016) redefined the annual constriction ratio into ACR = log (W v/v), with v expressed in m/s this time. They found a threshold of 4.26 below which all landslides formed dams.
al
The index took into account landslide velocity as an empirically evaluated parameter
ur n
(Cruden and Varnes, 1996) that is not always possible to measure, but for which calculated velocities based on runup heights (Chow, 1959; Jibson et al., 2006) and
Jo
seismic signal analyses delineate a range of values (Coe et al., 2016; Dufresne et al., 2019; Guthrie et al., 2012; Jiskoot, 2011). The Dimensionless Morpho-Invasion Index (DMI; Dal Sasso et al., 2014) is a derivative and more complex formulation of the ACR, relating landslide velocity (v), the density of the landslide material (ρL) and the landslide volume (VL) to the valley width (WV), the density of water (ρw), the gravitational acceleration (g), the hydraulic level (h), which is the elevation (m) of water surface from the bottom of river bed, and the landslide width (WL) in the relationship DMI = (2·ρL·v2 ·VL)/(ρw·g·h2·Wv·WL).
39
Journal Pre-proof Ermini (2003a) formulated the Dimensionless Constriction Index (DCI) to include the erodibility of the dam material in the stability estimates, taking into account landslide velocity (v), landslide width and height (WL and HL), cumulative grain size distribution (D30), the water discharge with a return period of 5 years (QP), and the width of the valley (WV) in DCI = (v·WL·HL·D 30) / (QP·WV). Correlating landslide volume (VL) with the width of the valley (W V), Tacconi Stefanelli et al. (2016) furthermore designed the Morphological Obstruction Index, MOI = log
oo
f
(VL/WV).
Dimensionless MorphoInvasion Index Dimensionless Constriction Index
> 100
log (WV/v)
< 4.26
(2·ρL·v2 ·VL) /
(QP·WV)
Morphological Obstruction
<1
> 0.002
log (VL/WV)
Region Japan Italy
Reference Swanson et al. (1986) Tacconi Stefanelli et al. (2016)
Italy
Dal Sasso et al. (2014)
Italy
Ermini (2003a)
<3
Italy
< 3.08
Peru
> 4.6
Jo
Index
> 6.88
>1
(ρw·g·h 2·Wv·WL)
(v·WL·HL·D30)/
e-
Ratio
v/ WV
No dam forms
Pr
Constriction
Dam forms
al
Annual
Equation
ur n
Index
pr
Table 10 Indices formulated to assess the probability of a landslide to block a river and form a dam.
Tacconi Stefanelli et al. (2016) Tacconi Stefanelli et al. (2018)
4.2.2 Dam stability indices Once a dam has formed, its stability depends on morphometric (size and shape of the dam and valley system), hydrogeological (pertaining to the river, the catchment and impoundment areas, and to climatic systems), and engineering geological (i.e. dam geotechnical properties) factors. The parameters used in the indices all interlink, and their relative weighting and significance regarding their ease of measure requires further investigation. 40
Journal Pre-proof With landslide dam volume (VD) in the numerator, the Blockage Index, BI = log(VD/AC) takes into account the catchment area, AC (Canuti et al., 1998; Casagli and Ermini, 1999), whilst the Impoundment Index (Casagli and Ermini, 1999) considers the volume of the impoundment, i.e. the landslide-dammed lake (VL), in II = log (VD/VL). Introducing dam height (HD) to the Blockage Index, Ermini and Casagli (2003) designed the Dimensionless Blockage Index, DBI = log (AC*H D/VD). Based on dam height (HD), Korup (2004) defined the Backstow Index IS = log (HD3/VL)
oo
f
and the Basin Index IA = log (H D2/AC). Note that in IS both parameters are in m3 , whereas in IA the dam height is in m and the catchment area in km2 . In the Relief Index
pr
IR = log (HD/HR) (IR; Korup, 2004), the dam height is correlated with the relief upstream
e-
of blockage, HR = Emax – Emin – HD; where Emax and Emin are the highest and lowest
Pr
altitudes in the catchment respectively, where the lowest altitude is corrected for landslide dam height. In IR (HD and HR), the units are meters.
al
Tacconi Stefanelli et al. (2016) proposed to consider dam destabilization by the river
ur n
using a simplified expression of stream power (AC*S) per unit channel length (where S is the local longitudinal slope of the channel bed), in their Hydromorphological Dam
Jo
Stability Index, HDSI = log (VD/AC*S). Table 11 categorizes all the above-mentioned geomorphic stability indices and their thresholds for apparent stability. Table 11 Indices drafted to categorize stable and unstable dams. In bold are the indices used for the analysis shown in Fig. 7. Index
Blockage Index (BI)
Equation
Stable
Uncertain
Unstable
Region
>5
4-5
3-4
Italy
>5
4-5
<4
Italy
log (VD/AC)
41
Reference Canuti et al. (1998) Casagli and
Journal Pre-proof Ermini (1999)
>0
Na
<0
>1
Na
<1
< 2.75
2.75-3.08
> 3.08
<3
3-5
>5
< 2.43
2.43-3.98
> 3.98
log (H D3/VL)
>0
0-(-3)
Basin Index (BI)
log (H D2/AC)
>3
Na
Relief Index (RI)
log (H D/HR)
> -1
Impoundment Index (II)
Italy
Casagli and Ermini (1999)
log (VD/VL) New Zealand
world-wide
Korup (2004)
Ermini and Casagli (2003)
Index (HDSI)
log (VD/AC*S)
ur n
Dam Stability
<3
New Zealand
Italy
New Zealand New Zealand New
Na
< -1
5.74-7.44
< 5.74
Italy
5.26-8.07
< 5.26
Peru
Zealand
al
> 7.44 Hydromorphic
oo
Index (IS)
< -3
pr
Backstow
e-
Index (DBI)
log (AC*HD/VD)
Pr
Blockage
f
Dimensionless
> 8.07
Korup (2004)
Tacconi Stefanelli et al. (2016)
Korup (2004)
Korup (2004)
Korup (2004)
Tacconi Stefanelli et al. (2016) Tacconi Stefanelli et al. (2018)
Jo
Owing to the empirical nature of HDSI, the thresholds vary with the dataset used and depend on regional differences (as demonstrated in studies in Italy and Peru by Tacconi Stefanelli et al. (2016); Tacconi Stefanelli et al. (2018)), and perhaps are also influenced by data acquisition methods. The thresholds will continue to be modified as more data becomes available. In order to evaluate a specific dam’s stability, however, these thresholds should be narrowed down for each distinct landslide type. For example, Miller et al. (2018) and Dufresne et al. (2018) tested the indices on a set of landslides in glaciolacustrine sediments (Peace River region, Canada), and large rock avalanches and rockslides (Oetz Valley, Austria), respectively. The calculations of dam 42
Journal Pre-proof stability remained largely inconclusive in both cases. However, the two studies are based on very limited cases and need to be reproduced with larger datasets (such as the one compiled herein) since some dams fall into “wrong” domains. The same issues of transferability were noted for dams formed by large-scale rockslides in Central Asia (Strom, 2010). As a first step, we chose a set of four very different landslides but of comparable volume, and in a similar geomorphic setting to contrast the different stability indices.
oo
f
The (1) Tangjiashan landslide dam (Sichuan, China) was formed by a rockslide triggered by the Mw 7.9 2008 Wenchuan earthquake. An artificial spillway was excavated
pr
to mitigate the dam-breach flood (Fan et al., 2012a). The (2) Signatico landslide (Italy)
e-
is complex debris and earth flow that was reactivated more than once over time. In
Pr
1945, a large debris flow formed a landslide-dammed lake in Parma Creek, which survived for several years (Tacconi Stefanelli et al., 2015). (3) Between 28 March and 4
al
April 1982, El Chichón erupted, producing pyroclastic deposits which dammed the
ur n
Magdalena River (Mexico). On 26 May, the landslide dam was breached releasing sediments and hot water downstream (Macías et al., 2004). (4) On 22 March 2014,
Jo
after a long period of heavy rainfall, the Oso landslide involving glacial and lacustrine sediments dammed the North Fork of the Stillaguamish River (Washington, USA) (Iverson et al., 2015).
All these four landslide dams have volumes of about 107 m3 and, apart from the Oso landslide, have had a landslide-dammed lake some km in length. Almost all the indices correctly classified the formation of the four landslide dams, with the exception of the Annual Constriction Ratio (with the formulation of Swanson et al., 1986) and, for the Signatico dam, the Dimensionless Morpho-Invasion Index (Table 12). In the stability calculations, nearly all the cases were classified as unstable except for the 43
Journal Pre-proof Dimensionless Blockage Index, which classified the Tangjiashan landslide as stable (Table 13). However, the Tangjiashan landslide dam was artificially breached, and its stability without human intervention remains unknown, though qualified to be potentially hazardous considering dam height, dam composition and a maximum capacity of the impoundment (Xu et al., 2009). Table 12 Index results for the prediction of the dam formation for the Tangjiashan, Signatico, El Chichón, and Oso landslides.
Constriction Ratio
1.68
Morphological Obstruction
Formed Formed
5.78
Formed Formed
4.85
2.38E-02 1.62
Formed
Formed
1.96E-06
12.80
Formed
n.a.
2.00E-
Not
Formed
02
Formed
Formed
1.70
Formed
2
Uncertain
2
Uncertain
3
Formed
Not
Formed
1
3.92 Formed
59.91
Formed
208.30
Formed
4
n.a.
-
n.a.
-
5
Pr
81.07
Ref
Not
4.62
5.30 Formed
Dimensionless
Invasion Index
-
al
Dimensionless Constriction
Formed
Formed
Index
Morpho-
Not
1.67E-06
Oso
f
Not
El Chichón
oo
2.08E-02
Signatico
pr
Annual
Tangjiashan
e-
Index
Index
et al. 2014; 5: Ermini 2003
ur n
References, 1: Swanson et al. 1986; 2: Tacconi Stefanelli et al. 2016; 3: Tacconi Stefanelli et al. 2018; 4: Dal Sasso
Index
Jo
Table 13 Index results for the prediction of the stability of the dam for the Tangjiashan, Signatico, El Chichón, and Oso landslides. Tangjiashan
Signatico
Unstable Blockage Index
Impoundment Index
Dimensionless Blockage Index
0.53
El Chichón
Unstable -1.33
Oso
Unstable -1.14
Ref. Unstable
1
-2.32
Unstable
Unstable
Unstable
Unstable
2
Unstable
Unstable
Unstable
-
2
-1.29
0.06
-0.81
n.a.
Unstable
Unstable
Unstable
-
3
Stable
Uncertain
Uncertain
Unstable
4
Uncertain
3
1.39
2.81 Stable
2.84 Stable 44
3.62 Stable
Journal Pre-proof Stable
Uncertain
Uncertain
Uncertain
5
Backstow Index
-2.62
Unstable
-2.47
Uncertain
-2.51
Uncertain
n.a.
-
3
Basin Index
-2.72
Unstable
-5.22
Unstable
-4.53
Unstable
-6.48
Unstable
3
Relief Index
n.a.
-
n.a.
-
n.a.
-
n.a.
-
3
Unstable
5
Unstable
6
Hydromorphic Dam Stability Index
Unstable 2.44
Unstable 1.54
Unstable
Unstable 0.90
Unstable
0.17 Unstable
oo
f
References, 1: Canuti et al. (1998); 2: Casagli and Ermini (1999); 3: Korup (2004); 4: Ermini and Casagli (2003); 5: Tacconi Stefanelli et al. (2016); 6: Tacconi Stefanelli et al. (2018).
In addition to the four cases above, the Impoundment Index, Blockage Index, Backstow
pr
Index, and the Dimensionless Blockage Index were calculated for cases with sufficient
e-
details from our new database (Figure 7). While none of the indices clearly distinguish stable and unstable dams precise enough to make well-informed quantitative risk
Pr
decisions, the overall performance of the DBI is found better at identifying some of the
al
stable dams which lasted between one and >20 years. However, the DBI also identifies
ur n
some landslide dams that lasted one-day-to-one-month and one-month-to-one-year as stable. The Impoundment Index identifies some cases as stable. The other two indices
Jo
mostly identify the landslide dams as either unstable or uncertain. Info rmation to calculate other indices, i.e. Relief Index and Hydromorphic Dam Stability Index are not available from the database, a common issue which hampers evaluation. The basis on which these indices were developed does not include the time a dam has remained stable. The Impoundment Index and Dimensionless Blockage Index show increases in identifying more dams as stable the longer these had existed (with some fluctuation, particularly concerning the insignificant number of dams in the 10-to-20years category). Other factors, such as climatic conditions or seismicity also do not factor into these indices. These observations suggest that geomorphic stability indices 45
Journal Pre-proof may be of help to classify past landslide dams and tune parameters, but they are not reliable to make decisions for present cases , especially during an emergency. It is therefore time to develop geotechnical indices or stability criteria that consider the internal dam structure and dam composition (information that is hard to gather in the required risk management timeframes) for rapid assessment and prioritization of
Jo
ur n
al
Pr
e-
pr
oo
f
resources in an event (e.g. earthquake trigger) scenario.
Figure 7 Geomorphic stability indices, categorizing stable and unstable landslide dams verified against longevity for available cases from our new worldwide database. Key to read figure: numbers after longevity denote the number of data-available cases. Read left to right, i.e. for Dimensionless Blockage Index, for > 20 years longevity, 58 % cases are predicted stable, 17% cases are uncertain and 25% cases are falsely predicted as unstable.
46
Journal Pre-proof 4.2.3 Multivariate statistics on predicting the failure probability of a landslide dam Data-driven models employing geomorphometric parameters of landslide dams and lakes have demonstrated to be a useful approach for assessing landslide dam stability (e.g., Chen and Chang, 2016; Dong et al., 2011a; Dong et al., 2009; Hua and Shaoyu, 2015; Korup, 2004). Logistic regression (LR), support vector machine (SVM), and Fisher’s discriminant analysis (FDA) are typically used in deriving landslide
oo
f
susceptibility maps and are also applied for landslide dams. These multivariate techniques use a dependent variable which is categorical (e.g., stable dam or unstable
pr
dam), and often have a large number of the independent variables (geomorphometric
e-
parameters), which can be either categorical, numerical (most of them obtained from a digital elevation model, not always available immediately) or both (Dong et al., 2011b).
Pr
A list of geomorphometric parameters used in the data-driven models and their
al
definitions are provided in Table S2 (supplementary material).
ur n
The data should be divided into training data (consisting at least half, usually 70% of the total cases) and testing data (the remaining). The training cases with selected variables
Jo
are used to train the model, and the test dataset is for validating the model. The performance evaluations of prediction rates are evaluated using ROC (receiver operating characteristic) diagram and AUC (area under the curve) values. Alternatively, kappa values, accuracy or a confusion matrix which shows true positive, false positive, true negative, and false negative are often used as evaluation criteria. Hua and Shaoyu (2015) compared different data-driven methods using 121 landslide dams documented in Ermini and Casagli (2003) and Tabata et al. (2002). They show that SVM has about 10% higher predictive performance than LR and about 23% higher performance than DBI. Nevertheless, these approaches should always be used with caution since their 47
Journal Pre-proof performance in regions beyond the geographical location of inventories can produce failure probability maps that contain several false positives and false negatives (Dong et al., 2011a).
5 Landscape implications of river-blocking landslides Here we examine the effects of river-blocking landslides on broader landscape processes across a spectrum of spatial and temporal scales. Following the scheme of Hewitt (2002) (Figure 8), the impacts of landslide dams on the landscape can be
oo
f
categorized temporally as short- (days to months), medium- (years to decades), and long-term (centuries to millennia), and spatially as upstream- and downstream-effects.
pr
Long-term landslide dams naturally bring more important implications to the broader
e-
landscape and thus it is to those we devote lengthier discussion. Hewitt (2002) outlines
Jo
ur n
al
final removal (Table 14; Figure 8).
Pr
five stages that trace landscape responses from the time of dam emplacement to its
Figure 8. Timescales of landslide dam and lake processes outlining stages of response to river-blocking landslides (after Hewitt, 2002).
48
Journal Pre-proof Individual cases may vary widely and display some or all of these stages, and, the evidence of some stages may be absent from the geomorphic and geological record due to censuring by subsequent erosional and depositional processes. When compared to basin sedimentation, landslide dam impacts are local and unlikely to be preserved in the stratigraphic record (Miall, 2013; Reading and Levell, 1996). However, over the Himalaya at least 177 (+/- 31) km3 of valley fill (15-20% of all valley fill) is associated with mass wasting, mostly as 57 large landslide deposits (Blöthe and Korup, 2013).
oo
f
Impacts are significant over human timescales, and this framework provides insights to hazard cascades associated with landslide-dam formation, failure, and beyond. In the
pr
following, we review the stages, link them to the indices and inventories discussed
e-
above, and suggest future research directions aimed at filling data and knowledge
Pr
gaps.
al
Table 14. Stages in the evolution of a landslide-interrupted river, identifying characteristic features and how they combine to underpin a land systems approach described below (modified after Hewitt, 2002; Korup, 2005a; Korup, 2005b). MAIN DEVELOPMENT
Stage 1:
Stage 2:
Landslide emplacement and creation of a dam
Landslide debris lobes, hummocks, pressure ridges, furrows, ‘brandung’; rudaceous or fragmented deposits
Aggradation and constructional landforms upstream of dam, submergence/inundation of upstream areas, lake sedimentation, reduction of discharge, changes in flow regime, decrease in upstream gradient.
Axial, fluvio-lacustrine flats, multi-channel (braided, anastomosing) streams; delta(s) and fan deltas, cross -bedded scour and fill, scattered swamp deposits; redistribution of fine-grained sediment by valley winds and dust storms; dune fields, loess, stream piracy
Jo
Landslide dam formation
TYPICAL LANDFORMS AND DEPOSITS
ur n
STAGE
Lake impoundment
Valley side and trib utaries; talus, colluvium, sediment fans, debris flow cones, disturbed confluence environments. On-landslide; knickpoints, megaclasts, ponds, loess.
49
Journal Pre-proof
Stage 4: Superimposed interruption
Stage 5:
Strath terraces superimposed (epigenetic) gorges, pothole swarms ‘defended’ landforms, outwash fan
Channels and straths superimposed in bedrock; minor depositional Remnants in sheltered locations, remnant irregularities in river thalwegs
Exhumed, pre-landslide valley floor sediments, epigenetic gorges Migrating knickpoints in bedrock
Short-term impacts of landslide dams
e-
5.1
Incision into the pre-landslide floor, an exhumation of buried valley fill & channels, increasingly scattered, smaller remnants of impoundment complex, increased flood frequency
Reworked and mixed sediments
pr
‘Shadow’ interruption
Spillway across the dam, segmented fans, trenched valley fill, discontinuous erosion terraces; migrating knickpoints in valley fill, alluvial flats.
f
Erosional interruption
Overtopping or piping, outburst flood, excavation, trenching and removal of the impoundment complex, changes in channel morphology, downstream sedimentation
oo
Stage 3:
Pr
5.1.1 Dam formation, impoundment, and inundation The most immediate disruption is the blockage itself (Hammack and Wohl, 1996; Korup,
al
2005b; Ohmori, 1992). This occurs synchronously with emplacement of the landslide
sediments
ur n
over seconds to minutes (Stage 1, Figure 8). The landslide dam retains water and from upstream, consequently causing
water and sediment deficit
1997).
Jo
downstream, which may experience erosion due to the lack of sediment load (Kondolf,
Filling a landslide-formed impoundment may take from days to years, though in some cases just hours (Stage 2). The rate of filling is a function of the input (runoff) versus output (evaporation and seepage) (Nash, 2003). For instance, the breached Köfels rockslide dam (Austria) with an estimated lake volume of ~2.4 km3 would have taken 46 months to fill assuming an annual (modern-day average) inflow of 19.8 m3 /s (Dufresne et al., 2018). The Sarez Lake (Figure 1a) formed by the Usoi landslide in 1911 attained close to its present-day level in the 1940s (Ischuk, 2011) and it appears that inflow now 50
Journal Pre-proof almost equals the outflow through seepage to maintain an equilibrium condition (Delaney and Evans, 2011). Conversely, a large (~60 million m3) rockslide dam formed ~5,000 years BP at Contrada Monte, Italy, is not yet filled due to its small (6.4 km 2 ) catchment area (Nicoletti and Parise, 2002). Impoundments in deep, narrow valleys or in areas of high runoff fill faster. A well-known example is the Hunza lake (600 million m3) that formed behind the January 2010 Attabad landslide (45 million m 3 ) in Pakistan, filled within 5 months, inundating 240 houses, 5 villages and about 25 km of the
oo
f
Karakoram national highway (Schneider et al., 2013). 5.1.2 Dam failure and flooding
pr
Examination of previous dam failures shows that ~80-90% fail within the first year of
e-
formation (Costa and Schuster, 1988; Clague and Evans, 1994; Tacconi Stefanelli et
Pr
al., 2018) (see Figs. 5 and 9); an observation confirmed in the aftermath of the 2008 Wenchuan earthquake (Fan et al., 2012c). Specific cases vary widely; for example, in
al
the northern Apennines the fraction is just 58 % (Canuti et al. (1998). The compositions
ur n
of landslide dams in the Costa and Schuster (1991) inventory are not known in detail, but observations from Himalayan landslide dams reported by Weidinger (2011) suggest
Jo
a relationship exists between landslide type (cf. Fig. 9), constituent grain size and dam longevity: dams containing fine-grained sediment (e.g., debris-flow dams) failing shortly after formation, whereas large rock avalanche/rockslide dams with overall larger grain sizes (up to a few 10s meters in diameter, especially typical of the upper carapace) have the potential to last up to tens of thousands of years. Our new database (Section 3 and Figure 9) also suggests that complex and debris flow type landslide dams tend to last only for brief periods (i.e., 1-month-to-1-year and 1-to10-years) followed by unconsolidated sediments. For example, 80% of complex, 65% of debris flows, and ~60% of unconsolidated sediments type of landslide dams have a 51
Journal Pre-proof longevity of less than 1 year, whereas rock/debris avalanches and rockslides can last longer. The high longevity of volcanic dams might be biased due to the limited number of cases presented in Table 2, most of which lasted more than 100 years, and the typically larger volumes of volcanic debris avalanches (up to 100s km3, whereas landslides from non-volcanic sources rarely exceed 1 km3). ‘Complex landslides’ initiate as one landslide type (e.g. rockslide) and successively transform into one (e.g. debris avalanche) or more (e.g. debris flow, earth flow) other
oo
f
types during emplacement. The landslide type relevant for the river blockage is typically one of the later, distal ones that reached the valley floor . It therefore makes sense that
pr
their longevity (except those formed by rock avalanches that also convert from
e-
rockslides) is at the lower end together with debris flows and flows/slides in
Pr
unconsolidated sediments. It forms one end-member in this distribution by never exceeding 1-year longevity. At the other end are dams from volcanic sources, all of
al
which lasted longer than 1-month. The difference of accumulated percentage of the
ur n
other 4 types (rock avalanches and rockslides, debris avalanches, debris flows, flows/slides in unconsolidated sediments) decreases with increasing time lapsed after
Jo
the dam formed. That is, the influence of landslide type or material blocking the river on the longevity of a landslide dam apparently decreases beyond the first year (25% difference when longevity < 1year; 15% difference for longevity between 1-to-10 years; 10% difference for longevity between 10-to-20 year). Beyond 10 years, about 30-40% of the dams survived regardless of the landslide type of rock/debris avalanches, rockslides, debris flows, or slides/flows in unconsolidated sediments. Post-emplacement processes that may favourably influence dam stability over time are consolidation of the dam body, minerals precipitating into pore spaces, or fine sediments washing into the landslide-dammed lake sealing the upstream face of the 52
Journal Pre-proof dam (e.g. Weidinger, 2011; Hewitt, 2011). Another factor that is seldom investigated or reported, but would affect river incision through the dam, is the across-valley dam profile governed by the ‘along-way’ debris distribution (Korchevsky et al., 2011). However, more systematic analyses that consider all influencing factors (dam volume, type of transformed landslide in the ‘complex’ category, river mean and peak discharge, dam internal structure, post-emplacement processes, and dam and impoundment morphometric parameters, etc.) as well as the number of data points per category for
oo
f
statistical significance and biases in data collection are needed for more reliable
Jo
ur n
al
Pr
e-
pr
inferences.
Figure 9. The longevity of landslide dams relative to landslide type according to our database. Landslide dams for which material type is unknown are not shown. The sooner 53
Journal Pre-proof cumulative percentage marches towards 100 %, the lesser the longevity of the landslide dam.
In dam breach scenarios, the severity of the dam-breach flood depends upon several factors, including (i) volume of outburst flows; (ii) geotechnical properties of damming material (i.e., grain size distribution, internal structure, friction angle, cohesion etc.); (iii) sediment load in the landslide-dammed lake; and (iv) amount of loose, easily erodible materials in the channel pathway (Schuster (2000). Dam breaching leads to stored
f
sediments being partially rerouted downstream either gradually or catastrophically—
oo
outburst floods are capable of transporting millions of m3 of debris (King et al., 1989).
pr
Some of the largest known cases of catastrophic dam failures are reported from China
e-
(Li, 1994; Li et al., 1986). For example, the 1967 breaching of the Tanggudong landslide on the Yangtze River generated > 50 m high waves and the flood travelled
Pr
1,000 km downstream (Nash et al., 2008; Nash, 2003). Earthquake-induced landslide
al
dams and consequent dam-beach floods can cause severe damage. For example, the
ur n
Dadu River flood after the 1786 earthquake (Dai et al., 2005), flooding after the 1920 Haiyuan earthquake in the loess plateau of northwest China (Havenith et al., 2015;
Jo
Huang, 2009), and the Minjiang River flood after the 1933 Diexi earthquake (Chai et al., 1995) all affected the downstream areas killing numerous people. Debris flows or river flooding triggered by breaching of large landslide dams can cause rare secondary blockages within the valley, but these are of short persistence. Examples include the 2004 catastrophic breach of the Tsatichhu landslide dam, Bhutan (Dunning et al., 2006), and the Ram Creek Dam, New Zealand (Nash et al., 2008).
54
Journal Pre-proof 5.2
Medium-term impacts of landslide dams
5.2.1 River morphology effects Mountain valleys have relatively steep long-profiles and commonly irregular width where confined sections alternate with discontinuous floodplains. Landslide dams typically impose spatial variation on rates of the fluvial incision (Korup, 2005a). Upstream of landslide dams, sedimentation reflects the raised local base level (to lake height). For instance, Ouimet et al. (2007) observed lower river gradients and small
oo
f
alluviated tributaries upstream while downstream tributaries continued to incise. Huang and Fan (2013) showed that five years after the 2008 Wenchuan earthquake some river
pr
beds remained aggraded by > 10 m, which may be detrimental to hydropower schemes,
e-
and also increases the probability of floods in the future.
Pr
In landslide-prone areas, Type-I landslide dams (according to Costa and Schuster, 1988) frequently exert strong control on river planform. For example, the Obi -Khingow
al
River in the northern Pamir, Tajikistan, has been repeatedly blocked by large debris
ur n
flows, either partially or completely (Luknitskiy, 1955), causing the channel thalweg to be displaced laterally between alternate valley margins. Similarly, the Chingshui River,
Jo
Taiwan, has undergone five successive blockages in the past 140 years downstream of Chingshui, resulting in channel width expansion, decreased channel curvature, and development of laterally-mobile channels and anabranches (Chen et al., 2008). 5.2.2 Sediment flux effects The mechanisms and range of extreme sediment transport events in fluvial settings and the global recycling of sediments were recently reviewed by Korup (2012). The study underscores that sediment fluxes in mountain rivers are significantly modulated by large landslide dams and their associated outburst floods. That said, few documented cases 55
Journal Pre-proof compare pre- and post-event sediment load. Pratt-Sitaula et al. (2007) reported that in steep rivers subject to blocking landslides, bedload accounts for ∼35% of the sediment transport and suspended load to ~50%. Large landslide dams with sizeable impoundment capacity can act to buffer peak river discharge downstream. The effects on sediment flux vary upstream and downstream of the landslide dam, but remain essentially the product of climate, basin relief and hillslope angle, lithology, and seismo-tectonics, as well as the environmental changes
oo
f
that have occurred in the geological past (Leeder et al., 1998). Upstream of the barrier dam, fluvial deltas and lacustrine sediments are commonly observed (Montgomery et
pr
al., 2004). Downstream of the dam, abrupt peaks in sediment flux during outburst floods
e-
can pose a significant risk to settlements and infrastructure, especially those developed
Pr
on alluvial fans subject to channel avulsion and increased flood frequency (Korup, 2004). In the wet tropics of Papua New Guinea, the 1985 Bairaman landslide dam
Long-term impacts of landslide dams
ur n
5.3
al
failure flushed sediment at a rate of ~5×10 9 t km−2 yr−1 (King et al., 1989; Korup, 2012).
For long-lived landslide dams, attention shifts from the dam itself to the modifications of
Jo
valley forms and deposits. Long-lived dams can control intermontane sedimentation, and the patterns and rates of denudation, notably by creating local base levels , and constraining sediment delivery (e.g. Korup, 2005a; Korup, 2005b) and peak discharge downstream. Hundreds of landslide dams have been identified as persisting for several millennia at least—in the European Alps (Abele, 1974; Heim, 1932), the Canadian Rockies (Cruden, 1985a; Cruden, 1985b; Eisbacher, 1979), the Southern Alps (Korup, 2002; Read et al., 1992; Whitehouse and Griffiths, 1983), the southern Andes (Hermanns et al., 2015; Hermanns et al., 2011a; Strecker and Marrett, 1999), in High Asia, including the Pamir, Tien Shan and Dzungaria (Abdrakhmatov and Strom, 2006; 56
Journal Pre-proof Strom, 2010; Strom, 2013; Strom, 2015; Strom and Abdrakhmatov, 2018), and the Nepal Himalaya (Weidinger, 2011). In the upper Indus basin, over 400 prehistoric landslide dams are now recognized, perhaps no more than one-third of those still awaiting identification (Evans et al., 2011a; Hewitt, 1988; Hewitt, 2011; Hewitt, 2014), yet their histories of longevity and failure/erosion remain unknown. Here, we outline what we may learn from paleo-dams by examining sediment archives recording their existence and their long-term impact on mountain rivers.
oo
f
5.3.1 Sedimentary archives of long-term landslide dams
Sedimentary archives of fluvial and lacustrine activity associated with numerous long -
pr
lived lakes can indicate high sediment loads during the Late Pleistocene and Holocene.
e-
For example, the thickness of lacustrine sediment (~15 ka to 30 ka BP) behind the Diexi
Pr
landslide dam on the Minjiang River, Sichuan, China, reaches 240 m (Wang and Han, 2014). Another example is the lake created by the Karakul rockslide dam on the Karasu
al
River, Tien Shan, which is now completely silted-up to a depth of 200 m and occupied
ur n
by the town of Karakul (Strom, 2010). The longevity of river damming can be assessed (at least to first approximation) based upon the sediment yield of the dammed river if
Jo
such data exist. Such an approach was used to estimate a ~2,200-3,100 lifespan for the 11-15 km3 sediment-filled lake created by the 8-12 km3 Beshkiol landslide on the Naryn River, central Tien Shan (Korup et al., 2006; Strom and Abdrakhmatov, 2018). Another important aspect of long-lived dams and their sedimentation record has been discussed in Korup and Montgomery (2008) and Korup et al. (2010). They pointed out that damming causes substantial hindering in river incision (impeded headward fluvial erosion), which protects the basins in the upper part of the river course. Thanks to sedimentary archives, we can also be alerted to cases in which a landslide dam and its lake remained stable for centuries to millennia before breaching 57
Journal Pre-proof catastrophically. The Casa de Los Loros dam o n the Las Conchas River, northwest Argentina, existed for ~7 ka before breaching 4.8 ± 0.5 ka BP, owing to a 225 × 106 m3 rock avalanche crashing into the lake 2 km upstream of the dam and causing an overtopping wave (Hermanns et al., 2004). Similar breaches caused by a landslide into a landslide-dammed lake are reported elsewhere (Glancy and Bell, 2000; Penna et al., 2008). The failure of the rockslide dam impounding Laguna Carri Lauquén, triggering an outburst flood over >900 km downstream is a good example of the impact of climate
oo
f
variability on dam stability. This dam existed for at least 425 yrs (Costa and Díaz, 2007; Hermanns et al., 2004) before breaching during heavy summer rains (González Díaz et
pr
al., 2001; Groeber, 1916); although the true cause may have been the recorded high
e-
temperatures of the preceding winter that promoted glacier melt and increased inflow to
Pr
the lake (Evans et al., 2011a; Evans et al., 2011b; Hermanns et al., 2004). Even once breached, landslide dams can impact the fluvial landscape for millennia as
al
they often persist in some remnant form, and they can add to or amplify the patchiness
ur n
that is typical of mountain rivers (Burchsted et al., 2014b; Burchsted and Daniels, 2014). The large Flims rockslide (8-12 km3 ), dated to around 8900 ± 700 yr B.P. (Ivy-Ochs et
Jo
al., 2009), is much-eroded, but still influences the upper Rhine River (von Poschinger et al., 2006), as do several landslides in the Ötztal Basin in the Alps (Dufresne et al., 2018; Hermanns et al., 2006), where knickpoints, extensive lake sediments, and outsized fans still dominate the valley some 3 ,000 years since the last blockage occurred. 5.3.2 Fragmentation of mountain river systems In many of the world’s high mountains, streams are repeatedly blocked, constricted, deflected or split into multiple channels by non-fluvial processes (Montgomery and Buffington, 1997; Wohl and Pearthree, 1991). Various studies find heterogeneity, 58
Journal Pre-proof discontinuities, or patchiness to be typical of mountain rivers (Benda et al., 2003; Benda et al., 2004; Burchsted et al., 2014a). However, some large landslide-related controls have been missed or misread as due to tectonic activity, climate change and, especially, glaciation (Fauqué et al., 2000; Fauqué and Tchilinguirian, 2002; Hewitt, 1999; Porter and Orombelli, 1981). Landslide dams have impacts comparable in scale and complexity to those identified with the ‘fragmented’ river systems concept (Hewitt, 2011); the ‘fragments’ being channel and valley sections above, below and between
oo
f
landslide dams. A majority of the obstructions may be one-time events that initiate local epicycles of
pr
deposition and erosion. And there has been a tendency to treat them as singular,
e-
extreme events, and their main geomorphic consequence as increases in sediment
Pr
yields (Gabet et al., 2008; Petley et al., 2007). However, the extent and significance of fragmented drainage in high mountains derive especially from multiple blockages along
landslides
occurring
upstream
and
downstream.
An
important
factor
ur n
other
al
the same streams. Individual cases endure long enough to affect and be affected by
predetermining landslide dam evolution is the presence of several dammed lakes along
Jo
the same river. Such multiple damming of a river predetermine the stability and longevity of the blockages at a large extent (Strom and Abdrakhmatov, 2018). If the upstream dam forms a large lake, it can transform flood runoff, accumulating summer water flow and releasing it gradually, thereby decreasing the erosional capacity supporting long-term stability of the downstream dam(s) that otherwise would be breached (see (Strom and Abdrakhmatov, 2018) for the Kulun rockslide dam and other cascading effects in the Central Asian region). On the other hand, breaching of an upstream dam may lead to the destruction of those downstream, as occurred during the
59
Journal Pre-proof 1963 Issyk Lake breach linked to the breach of the small Jarsay glacial lake upstream (Gerasimov, 1965; Strom and Abdrakhmatov, 2018). Landslide obstructions impose temporary local base levels of erosion that may survive for millennia. They are recurring aspects of some of the world’s highest mountain headwaters. In formerly glaciated highlands, they can be critical parts of, and controls over, Holocene conditions (eds. Evans et al., 2011b). Hewitt (2011) noted many landslide dam sites in the trans-Himalayan Upper Indus basin where streams mainly
oo
f
flow over thick valley fill. Bedrock channels comprise <2 % of the rivers in the Karakoram Himalaya. Large-scale storage of intermontane sediment and irregular
pr
channel slopes in the region mainly reflect drainage interruption by landslides. For
e-
example, the Yugo-Korphak (Y-K) rockslide-rock avalanche in the eastern Karakoram
Pr
(Figure 10a) and the Skardu Basin of the Indus River (Figure 10b) illustrate the large scale and scope of aggradation, as well as extensive reworking, trenching and removal
al
of valley fill. The extent of aggradation is unusually high for a region widely identified
ur n
with exceptional relief, high rates of tectonic uplift, and globally extreme erosion (Burbank, 2002; Cornwell et al., 2003; Shroder, 1994; Shroder Jr, 1998; Shroder Jr and
Jo
Bishop, 1998; Zeitler, 1985). There is a challenge for high mountain work that has emphasized steep channels incised in bedrock, or adjustments and evolution based on tectonics and climate change. These are surely important, but large landslide interruptions and associated deep valley-fills can counteract their influence (Korup and Montgomery, 2008; Korup et al., 2010). Sets of long-lived dams create opportunities to reconstruct (minimum) frequencies and spatial patterns of events, as well as landslide responses to changing climate and other conditions (Ho et al., 2017; Strom, 2015) if the inventory is large and complete enough.
60
al
Pr
e-
pr
oo
f
Journal Pre-proof
5.4
Jo
ur n
Figure 10. (a) The Yugo-Korphak (Y-K) rock avalanche dam on the Shyok River, eastern Karakoram (N 35.11, E 76.10), which blocked the Shyok (middle ground). Right panel: upstream view from where Y-K boulders mark the upper limits of the landslide dam. (b) The Indus River in the Skardu Basin, Baltistan, where the present river level is backed up behind the Katzarah rockslide (N 35.42, E 75.46). The original lake level of the landslide dam is represented by the terraces (50-70 m high) surrounding the basin.
Landslide dams and landscape evolution
Landslide dams can change the way fluvial systems function over geological timescales. Here, we outline how landslide dams cause fluvial knickpoint retreat and reorganisation of the river network. 5.4.1 Knickpoint development Knickpoints are local convexities or steepenings in river longitudinal profiles that occur as two main types: static and mobile. Static, or non-propagating knickpoints develop i n 61
Journal Pre-proof response to a resistant substrate, such as strong bedrock (Jansen, 2006) or where river-blocking landslide debris exceeds the sediment transport capacity of the river (Korup, 2006; Korup et al., 2010). Mobile knickpoints are typically the product of a change in climate-related river discharge or base-level fall due to tectonism (Crosby and Whipple (2006). These knickpoints behave as transient ‘incisional waves’ that propagate upstream at rates determined by discharge, bed slope, substrate erodibility, and sediment load (Gardner, 1983; Jansen et al., 2011; Yunus, 2016; Yunus et al.,
oo
f
2016). Most commonly, knickpoints are identified through visual analysis of ri ver profiles either
pr
mapped from topography or derived from a digital elevation model (DEM). Major
e-
knickpoints can be mapped quantitatively from an analysis of χ-space (Neely et al.,
Pr
2017), or channel slope versus drainage area (Foster and Kelsey, 2012; Zhang et al., 2011), and several application tools have recently appeared to facilitate this approach,
al
such as Stream Profiler tool (Whipple et al., 2007), Topotoolbox (Schwanghart and
ur n
Scherler, 2014; 2017)), KZPicker (Neely et al., 2017), GISknickfinder (Rengers, 2017), and LSDTopoTools (Gailleton et al., 2019).
Jo
Numerous examples of well-developed knickpoints associated with late Pleistocene, Holocene and recent landslide dams are reported from the Himalayas (Walsh et al., 2012), the Tibetan Mountains (Fan et al., 2019c), Central Asia (Korup, 2006; Strom and Abdrakhmatov, 2018) and the Southern Alps of New Zealand (Korup, 2006). In such landscapes, a downstream succession of multiple river-blocking landslides is rather common (Figure 11). Given that knickpoints can be generated by several different mechanisms, care is required for reliable interpretation. A normalised index of channel steepness known as 62
Journal Pre-proof Ksn has been used successfully to differentiate knickpoints at landslide dams from other types of knickpoints (Harrison et al., 2015; Korup, 2006). For example, detailed field surveys and DEM-based analysis of Ksn in the Nepal Himalayas indicate that steepness maxima and transitions to lower values reflect knickpoint formation associated with damming, subsequent alluviation, breaching, and incision (Walsh et al., 2012). The numbers and volumes of large landslides identified in some mountains, or parts of them, suggest they are a major factor in the unroofing of the orogen (Korup and Clague,
oo
f
2009; e.g. Malamud et al., 2004). And yet, the long-term effects of landslide dams would seem to retard fluvial incision and overall landscape denudation. One of the most
pr
critical effects of landslide dams concerns their role in potentially slowing down or
e-
stalling the upstream retreat of mobile knickpoints. Long-term blocking of mobile
Pr
knickpoints can halt the transmission of new base-level information throughout the landscape affecting long-profile development. Upstream of the landslide dam, river
al
incision abruptly declines and this altered coupling between rivers and hillslopes may
Jo
ur n
lead to a long-term reduction in the frequency of landsliding (Korup et al., 2010).
Figure 11 (a) Multiple knickpoints identified using ‘knickpointfinder’ (Schwanghart and Scherler, 2017) from the Minjiang River basin along 50 km stretch near Diexi, Sichuan, China—some of these are formed by massive rock avalanches. (b) Landslide boundary 63
Journal Pre-proof (white dashed lines) and knickpoint (red dot) location for a landslide dam induced by the 1933 Diexi earthquake. Image is a LiDAR DEM.
5.4.2 River piracy and network reorganisation River piracy, or stream capture, is the diversion of the headwaters of one stream into another, leading to a re-organisation of the stream network (Bishop, 1955). Piracy has been examined in terms of tectonic motion (Perucca et al., 2018; Valdiya, 1996), rapid base-level fall (Calvache and Viseras, 1997), and glacial environments (Carter et al.,
oo
f
2013; Shugar et al., 2017), but few studies have focused on river piracy caused by river-blocking landslides. River catchments are especially vulnerable to piracy when a
pr
landslide dam stands at a higher elevation than the lowest catchment drainage divide
e-
(Figure 12). Piracy has profound implications for sediment flux, hazard assessment,
Pr
and for sedimentary records of environmental change (Mather, 2003). Korup (2005b) recognized stream piracy as a generalized landslide impact in the alpine
al
river systems of southwest New Zealand. Cases of piracy have also been observed in
ur n
the Patagonian and the central Andes related to rock avalanches of volumes ranging 0.5-3 km3 (Fauqué et al., 2000; González Díaz et al., 2000; Hermanns et al., 2011a).
Jo
Yet more ancient cases are difficult to track due to their low preservation potential. However, investigation of lacustrine sediments tied to landslide debris provides the key to identifying stream captures from the past, but is not explored much. Abandonment of the Unaweep Canyon in the upper Colorado River, western USA, is thought to reflect stream piracy linked to a landslide dam (Aslan et al., 2014; Hood et al., 2014). Balco et al. (2013) and Oesleby (2005) suggested that the abandonment was caused by landslide-damming and the subsequent spillway in the western part of the Unaweep Canyon diverted the Gunnison River. 64
Journal Pre-proof
Figure 12 Schematic representation of river piracy caused by a landslide dam
oo
f
6 Conclusions and recommendations for future research Firstly, this review of landslide dams unified the wide range of terminologies used in the
pr
literature and recommended their use in future standardization for sensible cross-
e-
comparison of methods and data. In addition, we suggest the creation of a citable
Pr
encyclopaedia of landslide dam terminologies for cross -reference. Secondly, we reviewed the existing classification schema and proposed a preliminary landslide dams
al
stability classification matrix which can be used for emergency response. Further, we
ur n
compiled the existing landslide dam inventories and created a worldwide landslide dam database comprising 410 cases (since 1900 and with volumes >10 6 m3 ). Based on this
Jo
new database, various geomorphic indices for landslide dam formation and stability were reviewed and compared. None of these indices is found satisfactory; the challenge that remains is to develop a well-performing dam stability index with the least number of input parameters; mainly because the geotechnical properties of dam composition materials are often not available. Despite the growing number of publications on landslide dam case studies and event-based and regional landslide inventories, few landslide dams have complete data entries (e.g. triggers, failure mechanisms, accurate longevities, types of damming landslides, key parameters of dam stability, etc.), which hamper better understanding o f dam formation conditions and 65
Journal Pre-proof stability assessment. Finally, the short- to long- term impacts of the landslide dam were reviewed. The following are the research challenges and possible future research directions that we identify as most important: (1) Most landslide dam classifications are based on morphologic features, and neglect dam composition and internal geological structure. More well-studied cases formed by different types of landslides and new field measurement methods (e.g. geophysical methods) are encouraged, which will help to create a new classification that better
oo
f
serves landslide dam stability assessment.
pr
(2) Maintain and update the worldwide database with more data entries for as many case studies as possible to provide a sound statistical basis for any failure mechanism
e-
and longevity studies, and to produce adequate and representable input parameters for
Pr
models and hazard assessment. Climatic events and earthquake-triggered landslides
al
should be separated in order to examine their formation and failure.
ur n
(3) Landslide dam stability indices are so far based on heterogeneous datasets and need to be refined and tested for landslide dams of different materials, geo-tectonic and
Jo
climatic settings, and differentiated for events of different sizes. Physically-based and numerical models of dam formation, failure, and routing of water and sediment outbursts are highly required. (4) More work needs to be done on the influence of multiple dams, and whether they cluster spatially and temporally (and how dating methods can truly resolve a cluster versus a sequence), and how they, in turn, affect the stability of their downstream neighbours and the hazard cascades. (5) Catastrophic failures of landslide dams and resultant outburst floods increase sediment transport in mountain rivers. Quantifying this terrestrial sediment flux and 66
Journal Pre-proof linking it to the total sediment budget, and studies on the long-term orogenic processes warrant special attention. Future studies should be linked to approaches involving detachment-limited or transport-limited models that systematically estimate the deposit volumes and fractions to allow us to address key controls on incision rates and understand the sediment cascade in terms of temporal fluxes. (6) To detect ancient landslide dams in river profiles and differentiate knickpoints resulting from landslide-damming from those of other drivers (tectonic, climatic, base
oo
f
level, and lithological signals) is the key in developing a complete understanding of how large-scale landslide dams affect erosion and fluvial geomorphology. More research
pr
efforts are needed for developing a scalable landscape evolution model.
e-
(7) Further exploitation of lacustrine sediments will shed light on the reconstruction of
Pr
paleoenvironmental and paleoseismic events, as well as their signal in long -term
al
landscape evolution.
ur n
Acknowledgements
This research is financially supported by the National Science Fund for Outstanding
Jo
Young Scholars of China (Grant No. 41622206), the Funds for Creative Research Groups of China (Grant No. 41521002), and the Fund for International Cooperation (NSFC-RCUK_NERC), Resilience to Earthquake-induced landslide risk in China (grant No.
41661134010),
National Key R&D Program of China (No. 2017YFC1501002).
Sincere thanks to the editor, Prof. Chris Fielding, for the kind acceptance of our proposal and encouragement to submit this research work. Many thanks also go to editorial assistant Mr Tim Horscroft for the kind support. We thank the three anonymous reviewers whose comments improved the clarity and structure of this paper.
67
Journal Pre-proof
Competing interests Authors declare no competing interests.
Supplementary data The supplementary data for this manuscript is attached separately. A summary of the attached files is given below. 1. Supplementary_readme: A word file detailing the supplementary data and tables.
oo
f
2. Database: An excel file having details of the comprehensive database. 3. List_of_references_database: A word file listing the references cited in the
pr
database.
Jo
ur n
al
Pr
e-
4. Data_collection_template: A word file for the data collection template.
68
Journal Pre-proof
References
Jo
ur n
al
Pr
e-
pr
oo
f
Abdrakhmatov, K. and Strom, A., 2006. Dissected rockslide and rock avalanche deposits; Tien Shan, Kyrgyzstan, Landslides from Massive Rock Slope Failure. Springer, pp. 551-570. Abele, G., 1974. Bergsturze in den Alpen. Ihre Verbreitung, Morphologie und Folgeerscheinungen. Wiss. Alpenvereinshefte, 25. Adams, J., 1980. Contemporary uplift and erosion of the Southern Alps, New Zealand. Geological Society of America Bulletin, 91(1_Part_II): 1-114. Adams, J., 1981. Earthquake-dammed lakes in New Zealand. Geology, 9(5): 215-219. Alford, D. and Schuster, R.L., 2000. Usoi Landslide Dam and Lake Sarez—An Assessment of Hazard and Risk in the Pamir Mountains. Tajikistan: United Nations Secretariat for International Strategy for Disaster Reduction, Geneva. Amante, C. and Eakins, B.W., 2009. ETOPO1 Global Relief Model converted to PanMap layer format. NOAA-National Geophysical Data Center: Boulder, CO, USA, 10. Aslan, A., Hood, W.C., Karlstrom, K.E., Kirby, E., Granger, D.E., Kelley, S., Crow, R., Donahue, M.S., Polyak, V. and Asmerom, Y., 2014. Abandonment of Unaweep Canyon (1.4–0.8 Ma), western Colorado: Effects of stream capture and anomalously rapid Pleistocene river incision. Geosphere, 10(3): 428-446. Balco, G., Soreghan, G.S., Sweet, D.E., Marra, K.R. and Bierman, P.R., 2013. Cosmogenic-nuclide burial ages for Pleistocene sedimentary fill in Unaweep Canyon, Colorado, USA. Quaternary Geochronology, 18: 149-157. Benda, L., Miller, D., Bigelow, P. and Andras, K., 2003. Effects of post-wildfire erosion on channel environments, Boise River, Idaho. Forest Ecology and Management, 178(1-2): 105-119. Benda, L.E.E., Andras, K., Miller, D. and Bigelow, P., 2004. Confluence effects in rivers: interactions of basin scale, network geometry, and disturbance regimes. Water Resources Research, 40(5). Bishop, A.W., 1955. The use of the slip circle in the stability analysis of slopes. Geotechnique, 5(1): 7-17. Blöthe, J.H. and Korup, O., 2013. Millennial lag times in the Himalayan sediment routing system. Earth and Planetary Science Letters, 382: 38-46. Blöthe, J.H., Korup, O. and Schwanghart, W., 2015. Large landslides lie low: Excess topography in the Himalaya-Karakoram ranges. Geology, 43(6): 523-526. Bonnard, C., 2006. Technical and human aspects of historic rockslide dammed lakes and landslide dam breaches. Italian Journal of Engineering Geology and Environment(Special Issue 1): 21-30. Burbank, D.W., 2002. Rates of erosion and their implications for exhumation. Mineralogical Magazine, 66(1): 25-52. Burchsted, D., Daniels, M. and Wohl, E.E., 2014a. Introduction to the special issue on discontinuity of fluvial systems. Burchsted, D., Daniels, M. and Wohl, E.E., 2014b. Introduction to the special issue on discontinuity of fluvial systems. Discontinuities in Fluvial Systems, 205: 1-4. Burchsted, D. and Daniels, M.D., 2014. Classification of the alterations of beaver dams to headwater streams in northeastern Connecticut, USA. Geomorphology, 205: 36-50. Calvache, M.L. and Viseras, C., 1997. Long‐term control mechanisms of stream piracy processes in Southeast Spain. Earth Surface Processes and Landforms: The Journal of the British Geomorphological Group, 22(2): 93-105. 69
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Canuti, P., Casagli, N. and Ermini, L., 1998. Inventory of landslide dams in the Northern Apennine as a model for induced flood hazard forecasting. Managing HydroGeological Disasters in a Vulnerable Environment. CNR-GNDCI and UNESCO IHP, Perugia: 189-202. Capra, L., 2006. Abrupt climatic changes as triggering mechanisms of massive volcanic collapses. Journal of Volcanology and Geothermal Research, 155(3-4): 329-333. Capra, L., 2007. Volcanic natural dams: identification, stability, and secondary effects. Natural Hazards, 43(1): 45-61. Capra, L., 2011. Volcanic natural dams associated with sector collapses: textural and sedimentological constraints on their stability, Natural and artificial rockslide dams. Springer, pp. 279-294. Capra, L. and Macıas, J.L., 2002. The cohesive Naranjo debris-flow deposit (10 km3):: A dam breakout flow derived from the Pleistocene debris-avalanche deposit of Nevado de Colima Volcano (México). Journal of Volcanology and Geothermal Research, 117(1-2): 213-235. Carey, M., Huggel, C., Bury, J., Portocarrero, C. and Haeberli, W., 2012. An integrated socio-environmental framework for glacier hazard management and climate change adaptation: lessons from Lake 513, Cordillera Blanca, Peru. Climatic Change, 112(3): 733-767. Carey, S., Sigurdsson, H., Mandeville, C. and Bronto, S., 1996. Pyroclastic flows and surges over water: an example from the 1883 Krakatau eruption. Bulletin of Volcanology, 57(7): 493-511. Carter, S.P., Fricker, H.A. and Siegfried, M.R., 2013. Evidence of rapid subglacial water piracy under Whillans Ice Stream, West Antarctica. Journal of Glaciology, 59(218): 1147-1162. Casagli, N. and Ermini, L., 1999. Geomorphic analysis of landslide dams in the Northern Apennine. Transactions of the Japanese Geomorphological Union, 20(3): 219-249. Casagli, N., Ermini, L. and Rosati, G., 2003. Determining grain size distribution of the material composing landslide dams in the Northern Apennines: sampling and processing methods. Engineering geology, 69(1-2): 83-97. Chai, H.J., Liu, H.C. and Zhang, Z.Y., 1995. The catalog of Chinese landslide dam events. Journal of Geological Hazards and Environment Preservation, 6(4): 1-9. Chen, C.-Y. and Chang, J.-M., 2016. Landslide dam formation susceptibility analysis based on geomorphic features. Landslides, 13(5): 1019-1033. Chen, S.C., Hsu, C.L., Wu, T.Y., Chou, H.T. and Cui, P., 2011. Landslide dams induced by typhoon Morakot and risk assessment, 5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment, Padua, Italy, pp. 653-660. Chen, X., Wu, J. and Hu, Q., 2008. Simulation of climate change impacts on streamflow in the Bosten Lake Basin using an artificial neural network model. Journal of Hydrologic Engineering, 13(3): 180-183. Chow, V.T., 1959. Open-channel hydraulics, Open-channel hydraulics. McGraw-Hill. Clague, J.J. and Evans, S.G., 1994. Formation and failure of natural dams in the Canadian Cordillera. Geological Survey of Canada Bulletin, 464 pp. Clavero, J.E., Sparks, R.S.J., Pringle, M.S., Polanco, E. and Gardeweg, M.C., 2004. Evolution and volcanic hazards of Taapaca volcanic complex, Central Andes of northern Chile. Journal of the Geological Society, 161(4): 603-618. Coe, J.A., Baum, R.L., Allstadt, K.E., Kochevar Jr, B.F., Schmitt, R.G., Morgan, M.L., White, J.L., Stratton, B.T., Hayashi, T.A. and Kean, J.W., 2016. Rock-avalanche dynamics revealed by large-scale field mapping and seismic signals at a highly 70
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
mobile avalanche in the West Salt Creek valley, western Colorado. Geosphere, 12(2): 607-631. Cornwell, K., Norsby, D. and Marston, R., 2003. Drainage, sediment transport, and denudation rates on the Nanga Parbat Himalaya, Pakistan. Geomorphology, 55(1-4): 25-43. Cortés, A., 2002. Depósitos de avalancha y flujos de escombros originados hace 3,600 años por el colapso del sector Suroeste del Volcán de Colima. Cortés, A., Macías, J.L., Capra, L. and Garduño-Monroy, V.H., 2010. Sector collapse of the SW flank of Volcán de Colima, México: the 3600 yr BP La Lumbre–Los Ganchos debris avalanche and associated debris flows. Journal of Volcanology and Geothermal Research, 197(1-4): 52-66. Costa, C.H. and Díaz, E.F.G., 2007. Age constraints and paleoseismic implication of rock avalanches in the northern Patagonian Andes, Argentina. Journal of South American Earth Sciences, 24(1): 48-57. Costa, J.E. and Schuster, R.L., 1988. The formation and failure of natural dams. Geological society of America bulletin, 100(7): 1054-1068. Costa, J.E. and Schuster, R.L., 1991. Documented historical landslide dams from around the world. 2331-1258, US Geological Survey. Crosby, B.T. and Whipple, K.X., 2006. Knickpoint initiation and distribution within fluvial networks: 236 waterfalls in the Waipaoa River, North Island, New Zealand. Geomorphology, 82(1-2): 16-38. Crozier, M.J. and Pillans, B.J., 1991. Geomorphic events and landform response in south-eastern Taranaki, New Zealand. CATENA, 18(5): 471-487. Cruden, D.M., 1985a. Destructive mass movements in high mountains: hazard and management, Geological Survey of Canada Paper 84-16: Book review. Canadian Geotechnical Journal, 22(3): 426-426. Cruden, D.M., 1985b. Rock slope movements in the Canadian Cordillera. Canadian Geotechnical Journal, 22(4): 528-540. Cruden, D.M., Keegan, T.R. and Thomson, S., 1993. The landslide dam on the Saddle River near Rycroft, Alberta. Canadian Geotechnical Journal, 30(6): 1003-1015. Cruden, D.M., Lu, Z.Y. and Thomson, S., 1997. The 1939 Montagneuse River landslide, Alberta. Canadian Geotechnical Journal, 34(5): 799-810. Cruden, D.M. and Varnes, D.J., 1996. Landslides: investigation and mitigation. Chapter 3-Landslide types and processes. Transportation research board special report(247). Cui, P., Zhu, Y.-y., Han, Y.-s., Chen, X.-q. and Zhuang, J.-q., 2009. The 12 May Wenchuan earthquake-induced landslide lakes: distribution and preliminary risk evaluation. Landslides, 6(3): 209-223. Dai, F.C., Lee, C.F., Deng, J.H. and Tham, L.G., 2005. The 1786 earthquake-triggered landslide dam and subsequent dam-break flood on the Dadu River, southwestern China. Geomorphology, 65(3-4): 205-221. Dal Sasso, S.F., Sole, A., Pascale, S., Sdao, F., Bateman Pinzón, A. and Medina, V., 2014. Assessment methodology for the prediction of landslide dam hazard. Natural Hazards and Earth System Sciences, 14(3): 557-567. Davies, T.R. and McSaveney, M.J., 2011. Rock-avalanche size and runout–implications for landslide dams, Natural and Artificial Rockslide Dams. Springer, pp. 441-462. Davies, T.R. and McSaveney, M.J., 2012. Mobility of long-runout rock avalanches. Landslides–types, mechanisms and modeling. Edited by JJ Clague and D. Stead. Cambridge University Press: 50-58.
71
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Davies, T.R.H. and Korup, O., 2010. Sediment cascades in active landscapes. Sediment Cascades: An Integrated Approach. John Wiley and Sons, Chichester: 89-115. Delaney, K.B. and Evans, S.G., 2011. Rockslide Dams in the Northwest Himalayas (Pakistan, India) and the Adjacent Pamir Mountains (Afghanistan, Tajikistan), Central Asia, Natural and Artificial Rockslide Dams. Springer, pp. 205-242. Delaney, K.B. and Evans, S.G., 2015. The 2000 Yigong landslide (Tibetan Plateau), rockslide-dammed lake and outburst flood: review, remote sensing analysis, and process modelling. Geomorphology, 246: 377-393. Dhital, M.R., Rijal, M.L. and Bajracharya, S.R., 2016. Landslide dams and their hazard after the 25 April 2015 Gorkha Earthquake in Central Nepal. Lowland technology international: the officisl journal of the International Association of Lowland Technology (IALT)/Institute of Lowland Technology, Saga University, 18(2): 129140. Dong, J.-J., Li, Y.-S., Kuo, C.-Y., Sung, R.-T., Li, M.-H., Lee, C.-T., Chen, C.-C. and Lee, W.-R., 2011a. The formation and breach of a short-lived landslide dam at Hsiaolin village, Taiwan—part I: post-event reconstruction of dam geometry. Engineering geology, 123(1-2): 40-59. Dong, J.-J., Tung, Y.-H., Chen, C.-C., Liao, J.-J. and Pan, Y.-W., 2009. Discriminant analysis of the geomorphic characteristics and stability of landslide dams. Geomorphology, 110(3-4): 162-171. Dong, J.-J., Tung, Y.-H., Chen, C.-C., Liao, J.-J. and Pan, Y.-W., 2011b. Logistic regression model for predicting the failure probability of a landslide dam. Engineering Geology, 117(1-2): 52-61. Dufresne, A., Ostermann, M. and Preusser, F., 2018. River-damming, late-Quaternary rockslides in the Ötz Valley region (Tyrol, Austria). Geomorphology, 310: 153167. Dufresne, A., Wolken, G.J., Hibert, C., Bessette-Kirton, E.K., Coe, J.A., Geertsema, M. and Ekström, G., 2019. The 2016 Lamplugh rock avalanche, Alaska: deposit structures and emplacement dynamics. Landslides, 16(12): 2301-2319. Duman, T.Y., 2009. The largest landslide dam in Turkey: Tortum landslide. Engineering Geology, 104(1-2): 66-79. Dunning, S.A. and Armitage, P.J., 2011. The grain-size distribution of rock-avalanche deposits: implications for natural dam stability, Natural and artificial rockslide dams. Springer, pp. 479-498. Dunning, S.A., Mitchell, W.A., Rosser, N.J. and Petley, D.N., 2007. The Hattian Bala rock avalanche and associated landslides triggered by the Kashmir Earthquake of 8 October 2005. Engineering Geology, 93(3-4): 130-144. Dunning, S.A., Petley, D.N., Rosser, N.J. and Strom, A.L., 2005. The morphology and sedimentology of valley confined rock-avalanche deposits and their effect on potential dam hazard, Proceedings of the International Conference on Landslide Risk Management, Edited by O. Hungr, R. Fell, R. Couture, and E. Eberhardt. Taylor & Francis, Balkema, London, pp. 691-701. Dunning, S.A., Rosser, N.J., Petley, D.N. and Massey, C.R., 2006. Formation and failure of the Tsatichhu landslide dam, Bhutan. Landslides, 3(2): 107-113. Eisbacher, G.H., 1979. Cliff collapse and rock avalanches (sturzstroms) in the Mackenzie Mountains, northwestern Canada. Canadian Geotechnical Journal, 16(2): 309-334. Erismann, T.H. and Abele, G., 2013. Dynamics of rockslides and rockfalls. Springer Science & Business Media. 72
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Ermini, L., 2000. Elaborazione di un modello per la precisione dell’evoluzione di sbarramenti fluviali causati da frane, unpublished. Unpublished PhD thesis, University of Florence, 159. Ermini, L., 2003a. Gli sbarramenti d’alveo da frane: Criteri speditivi per la stesura di scenari evolutivi derivanti dalla loro formazione. AIGA-IConvegno Nazionale: 355-367. Ermini, L., 2003b. Gli sbarramenti fluviali causati da frane: criteri speditivi per la stesura di scenari evolutivi derivanti dalla loro formazione, I Congresso Associazione Italiana Geologia Applicata, Chieti, pp. 355-367. Ermini, L. and Casagli, N., 2003. Prediction of the behaviour of landslide dams using a geomorphological dimensionless index. Earth Surface Processes and Landforms: The Journal of the British Geomorphological Research Group, 28(1): 31-47. Evans, S.G., Delaney, K.B., Hermanns, R.L., Strom, A. and Scarascia-Mugnozza, G., 2011a. The formation and behaviour of natural and artificial rockslide dams; implications for engineering performance and hazard management, Natural and artificial rockslide dams. Springer, pp. 1-75. Evans, S.G., Hermanns, R.L., Strom, A. and Scarascia-Mugnozza, G., 2011b. Natural and artificial rockslide dams, 133. Springer Science & Business Media. Evans, S.G., Hu, X.Q. and Enegren, E.G., 1996. The 1973 Attachie slide, Peace River Valley, near Fort St. John, British Columbia, Canada: a landslide with a highvelocity flowslide component in Pleistocene sediments, Proceedings, 7th International Symposium on Landslides, pp. 715-720. Fan, X., 2013. Understanding the causes and effects of earthquake-induced landslide dams. Fan, X., Huang, R., van Westen, C.J., Xu, Q., Havenith, H.-B. and Jetten, V., 2014a. A Conceptual Event-Tree Model for Coseismic Landslide Dam Hazard Assessment. In: K. Sassa, P. Canuti and Y. Yin (Editors), Landslide Science for a Safer Geoenvironment. Springer International Publishing, Cham, pp. 605-608. Fan, X., Rossiter, D.G., van Westen, C.J., Xu, Q. and Görüm, T., 2014b. Empirical prediction of coseismic landslide dam formation. Earth Surface Processes and Landforms, 39(14): 1913-1926. Fan, X., Scaringi, G., Korup, O., West, A.J., van Westen, C.J., Tanyas, H., Hovius, N., Hales, T.C., Jibson, R.W., Allstadt, K.E., Zhang, L., Evans, S.G., Xu, C., Li, G., Pei, X., Xu, Q. and Huang, R., 2019a. Earthquake-Induced Chains of Geologic Hazards: Patterns, Mechanisms, and Impacts. Reviews of Geophysics, 0(0). Fan, X., Tang, C.X., Van Westen, C.J. and Alkema, D., 2012a. Simulating dam-breach flood scenarios of the Tangjiashan landslide dam induced by the Wenchuan Earthquake. Natural hazards and earth system sciences, 12(10): 3031. Fan, X., van Westen, C.J., Korup, O., Gorum, T., Xu, Q., Dai, F., Huang, R. and Wang, G., 2012b. Transient water and sediment storage of the decaying landslide dams induced by the 2008 Wenchuan earthquake, China. Geomorphology, 171: 58-68. Fan, X., van Westen, C.J., Xu, Q., Gorum, T. and Dai, F., 2012c. Analysis of landslide dams induced by the 2008 Wenchuan earthquake. Journal of Asian Earth Sciences, 57: 25-37. Fan, X., van Westen, C.J., Xu, Q., Gorum, T., Dai, F., Wang, G. and Huang, R., 2013. Spatial Distribution of Landslide Dams Triggered by the 2008 Wenchuan Earthquake. In: C. Margottini, P. Canuti and K. Sassa (Editors), Landslide Science and Practice. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 279285. 73
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Fan, X., Xu, Q., Alonso-Rodriguez, A., Siva Subramanian, S., Li, W., Zheng, G., Dong, X. and Huang, R., 2019b. Successive landsliding and damming of the Jinsha River in eastern Tibet, China: prime investigation, early warning, and emergency response. Landslides, 16(5): 1003-1020. Fan, X., Xu, Q., van Westen, C.J., Huang, R. and Tang, R., 2017. Characteristics and classification of landslide dams associated with the 2008 Wenchuan earthquake. Geoenvironmental Disasters, 4(1): 12. Fan, X., Yunus, A.P., Jansen, J.D., Dai, L., Strom, A. and Xu, Q., 2019c. Comment on ‘Gigantic rockslides induced by fluvial incision in the Diexi area along the eastern margin of the Tibetan Plateau’ by Zhao et al. (2019) Geomorphology 338, 27–42. Geomorphology: 106963. Fauqué, L., Cortés, J.M., Folguera, A. and Etcheverría, M., 2000. Avalanchas de rocas asociadas a neotectónica en el valle del río Mendoza, al sur de Uspallata. Revista de la Asociación Geológica Argentina, 55(4): 419-423. Fauqué, L. and Tchilinguirian, P., 2002. Villavil rockslides, Catamarca Province, Argentina. Catastrophic Landslides: Effects, Occurrence, and Mechanism: 30324. Ferrer, C., 1999. Represamientos y rupturas de embalses naturales (lagunas de obstrución) como efectos cosísmicos: Algunos ejemplos en los Andes venezolanos. Revista Geográfica Venezolana, 40(1): 109-121. Fletcher, L., Hungr, O. and Evans, S.G., 2002. Contrasting failure behaviour of two large landslides in clay and silt. Canadian Geotechnical Journal, 39(1): 46-62. Fort, M., 1987. Geomorphic and Hazards mapping in the dry, continental Himalaya: 1: 50,000 maps of Mustang District, Nepal. Mountain Research and Development: 222-238. Fort, M. and Peulvast, J.P., 1995. Catastrophic mass-movements and morphogenesis in the Peri-Tibetan Ranges: examples from West Kunlun, East Pamir and Ladakh. Steepland Geomorphology: 171-198. Foster, M.A. and Kelsey, H.M., 2012. Knickpoint and knickzone formation and propagation, South Fork Eel River, northern California. Geosphere, 8(2): 403416. Frank, F., McArdell, B.W., Huggel, C. and Vieli, A., 2015. The importance of entrainment and bulking on debris flow runout modeling: examples from the Swiss Alps. Natural Hazards and Earth System Sciences, 15(11): 2569-2583. Gabet, E.J., Burbank, D.W., Pratt-Sitaula, B., Putkonen, J. and Bookhagen, B., 2008. Modern erosion rates in the High Himalayas of Nepal. Earth and Planetary Science Letters, 267(3-4): 482-494. Gailleton, B., Mudd, S.M., Clubb, F.J., Peifer, D. and Hurst, M.D., 2019. A segmentation approach for the reproducible extraction and quantification of knickpoints from river long profiles. Earth Surface Dynamics, 7(1): 211-230. Gallino, G.L. and Pierson, T.C., 1985. Polallie Creek debris flow and subsequent dambreak flood of 1980, East Fork Hood River basin, Oregon, 2273. US Geological Survey. Gardner, T.W., 1983. Experimental study of knickpoint and longitudinal profile evolution in cohesive, homogeneous material. Geological Society of America Bulletin, 94(5): 664-672. Geertsema, M., 1998. Flowslides in waterlain muds of northwestern British Columbia, Canada, Proceedings of the 8th Congress of the International Association of Engineering Geology and the Environment, pp. 1913-1921. Geertsema, M., Blais-Stevens, A., Kwoll, E., Menounos, B., Venditti, J.G., Grenier, A. and Wiebe, K., 2018. Sensitive clay landslide detection and characterization in 74
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
and around Lakelse Lake, British Columbia, Canada. Sedimentary geology, 364: 217-227. Geertsema, M. and Clague, J.J., 2006. 1,000-year record of landslide dams at Halden Creek, northeastern British Columbia. Landslides, 3(3): 217-227. Geertsema, M. and Clague, J.J., 2008. Natural dams, temporary lakes, and outburst floods in western Canada, Proceedings of the 1st World Landslide Forum, pp. 211-214. Geertsema, M., Clague, J.J., Schwab, J.W. and Evans, S.G., 2006a. An overview of recent large catastrophic landslides in northern British Columbia, Canada. Engineering Geology, 83(1-3): 120-143. Geertsema, M. and Cruden, D.M., 2008. Travels in the Canadian Cordillera, Proceedings of 4th Canadian conference on geohazards, Quebec City, pp. 383390. Geertsema, M., Cruden, D.M. and Schwab, J.W., 2006b. A large rapid landslide in sensitive glaciomarine sediments at Mink Creek, northwestern British Columbia, Canada. Engineering Geology, 83(1-3): 36-63. Geertsema, M., Highland, L. and Vaugeouis, L., 2009. Environmental impact of landslides, Landslides–disaster risk reduction. Springer, pp. 589-607. Geertsema, M., Hungr, O., Schwab, J.W. and Evans, S.G., 2006c. A large rockslide– debris avalanche in cohesive soil at Pink Mountain, northeastern British Columbia, Canada. Engineering Geology, 83(1-3): 64-75. Gerasimov, V., 1965. The Issyk catastrophe in 1963 and its effects on geomorphology of the Issyk River Valley. Transactions of the All-union Geographic Society, 97(6): 541-547. Glancy, P.A. and Bell, J.W., 2000. Landslide-induced flooding at Ophir Creek, Washoe County, western Nevada, May 30, 1983. 2330-7102, US Geological Survey. Glicken, H. and Nakamura, Y., 1988. Restudy of the 1888 eruption of Bandai volcano, Japan, pp. 392-395. Gómez-Avalos, J.C., Macias, J.L., Siebe, C. and Ocola, A.L., 2004. Debris avalanche deposit of Hualca Hualca volcano and the formation of a volcanic dam in the Colca Valley, Arequipa e Peru. International Association of Volcanology and Chemistry of the Earth Interior General Assembly, Chile. Gong, W., Jiang, X., Zhou, H., Xing, J., Li, C. and Yang, K., 2018. Varied thermorheological structure, mechanical anisotropy and lithospheric deformation of the southeastern Tibetan Plateau. Journal of Asian Earth Sciences, 163: 108-130. González Díaz, E.F., Fauqué, L.E., Giaccardi, A.D. and Costa, C., 2000. Las lagunas de Varvar Co Campos y Varvar Co Tapia (N del Neuquén, Argentina): su relación con avalanchas de rocas. Revista de la Asociación Geológica Argentina, 55(3): 147-164. González Díaz, E.F., Giaccardi, A.D. and Costa, C.H., 2001. La avalancha de rocas del río Barrancas (Cerro Pelán), norte del Neuquén: su relación con la catástrofe del río Colorado (29/12/1914). Revista de la Asociación Geológica Argentina, 56(4): 466-480. Griswold, J.P. and Iverson, R.M., 2008. Mobility statistics and automated hazard mapping for debris flows and rock avalanches. US Department of the Interior, US Geological Survey. Groeber, P., 1916. Informe sobre las causas que han producido las crecientes del río Colorado (Territorios del Neuquén y La Pampa) en 1914. 1-29 p. Buenos Aires: Dirección General de Minas, Geología e Hidrogeología. Gruber, N., Gloor, M., Mikaloff Fletcher, S.E., Doney, S.C., Dutkiewicz, S., Follows, M.J., Gerber, M., Jacobson, A.R., Joos, F. and Lindsay, K., 2009. Oceanic 75
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
sources, sinks, and transport of atmospheric CO2. Global Biogeochemical Cycles, 23(1). Guthrie, R.H., Friele, P., Allstadt, K., Roberts, N., Evans, S.G., Delaney, K.B., Roche, D., Clague, J.J. and Jakob, M., 2012. The 6 August 2010 Mount Meager rock slide-debris flow, Coast Mountains, British Columbia: characteristics, dynamics, and implications for hazard and risk assessment. Natural Hazards and Earth System Sciences, 12(5): 1277-1294. Guzzetti, F. and Tonelli, G., 2004. Information system on hydrological and geomorphological catastrophes in Italy (SICI): a tool for managing landslide and flood hazards. Hammack, L. and Wohl, E., 1996. Debris-fan formation and rapid modification at Warm Springs rapid, Yampa River, Colorado. The Journal of Geology, 104(6): 729-740. Hancox, G.T. and Perrin, N.D., 2009. Green Lake Landslide and other giant and very large postglacial landslides in Fiordland, New Zealand. Quaternary Science Reviews, 28(11-12): 1020-1036. Harrison, L.M., Dunning, S.A., Woodward, J. and Davies, T.R.H., 2015. Post-rockavalanche dam outburst flood sedimentation in Ram Creek, Southern Alps, New Zealand. Geomorphology, 241: 135-144. Havenith, H.-B., Torgoev, I., Torgoev, A., Strom, A., Xu, Y. and Fernandez-Steeger, T., 2015. The Kambarata 2 blast-fill dam, Kyrgyz Republic:blast event, geophysical monitoring and dam structure modelling. Geoenvironmental Disasters, 2(1): 11. Heim, A., 1932. Landslides and human lives (Bergsturz und Menschenleben). Bi -Tech Publishers, Vancouver, BC: 196. Henderson, R.D. and Thompson, S.M., 1999. Extreme rainfalls in the Southern Alps of New Zealand. Journal of Hydrology (New Zealand): 309-330. Hermanns, R.L., 2013. Landslide dam. Encyclopedia of Natural Hazards: 602-606. Hermanns, R.L., Blikra, L.H., Naumann, M., Nilsen, B., Panthi, K.K., Stromeyer, D. and Longva, O., 2006. Examples of multiple rock-slope collapses from Köfels (Ötz valley, Austria) and western Norway. Engineering Geology, 83(1-3): 94-108. Hermanns, R.L., Dahle, H., Bjerke, P.L., Crosta, G.B., Anda, E., Blikra, L.H., Saintot, A. and Longva, O., 2013. Rockslide dams in møre og romsdal county, Norway, Landslide science and practice. Springer, pp. 3-12. Hermanns, R.L., Fauqué, L. and Wilson, C.G.J., 2015. 36Cl terrestrial cosmogenic nuclide dating suggests Late Pleistocene to Early Holocene mass movements on the south face of Aconcagua mountain and in the Las Cuevas–Horcones valleys, Central Andes, Argentina. Geological Society, London, Special Publications, 399(1): 345-368. Hermanns, R.L., Folguera, A., Penna, I., Fauqué, L. and Niedermann, S., 2011a. Landslide dams in the Central Andes of Argentina (northern Patagonia and the Argentine northwest), Natural and artificial rockslide dams. Springer, pp. 147176. Hermanns, R.L., Hansen, L., Sletten, K., Böhme, M., Bunkholt, H., Dehls, J.F., Eilertsen, R., Fischer, L., Lheureux, J.S. and Høgaas, F., 2012. Systematic geological mapping for landslide understanding in the Norwegian context. Landslide and engineered slopes: protecting society through improved understanding. Taylor & Francis Group, London: 265-271. Hermanns, R.L., Hewitt, K., Strom, A., Evans, S.G., Dunning, S.A. and ScarasciaMugnozza, G., 2011b. The classification of rockslide dams, Natural and artificial rockslide dams. Springer, pp. 581-593. Hermanns, R.L., Niedermann, S., Ivy-Ochs, S. and Kubik, P.W., 2004. Rock avalanching into a landslide-dammed lake causing multiple dam failure in Las 76
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Conchas valley (NW Argentina)—evidence from surface exposure dating and stratigraphic analyses. Landslides, 1(2): 113-122. Hermanns, R.L., Valderrama, P., Fauqué, L., Penna, I.M., Sepúlveda, S., Moreiras, S. and Zavala Carrión, B., 2012. LandsLIdes In tHe andes and tHe need to CommunICate on an InteRandean LeVeL on LandsLIde maPPIng and ReseaRCH. Revista de la Asociación Geológica Argentina, 69(3): 321-327. Hewitt, K., 1988. Catastrophic landslide deposits in the Karakoram Himalaya. Science, 242(4875): 64-67. Hewitt, K., 1999. Quaternary moraines vs catastrophic rock avalanches in the Karakoram Himalaya, northern Pakistan. Quaternary Research, 51(3): 220-237. Hewitt, K., 2002. Postglacial landform and sediment associations in a landslidefragmented river system: The TransHimalayan Indus streams, central Asia, Landscapes of transition. Springer, pp. 63-91. Hewitt, K., 2009a. Catastrophic rock slope failures and late Quaternary developments in the Nanga Parbat–Haramosh Massif, Upper Indus basin, northern Pakistan. Quaternary Science Reviews, 28(11-12): 1055-1069. Hewitt, K., 2009b. Rock avalanches that travel onto glaciers and related developments, Karakoram Himalaya, Inner Asia. Geomorphology, 103(1): 66-79. Hewitt, K., 2011. Rock avalanche dams on the Trans Himalayan upper Indus streams: a survey of late Quaternary events and hazard-related characteristics, Natural and artificial rockslide dams. Springer, pp. 177-204. Hewitt, K., 2014. Regions of risk: A geographical introduction to disasters. Routledge. Hewitt, K., Clague, J.J. and Orwin, J.F., 2008. Legacies of catastrophic rock slope failures in mountain landscapes. Earth-Science Reviews, 87(1-2): 1-38. Ho, K., Lacasse, S. and Picarelli, L., 2017. Slope Safety Preparedness for Impact of Climate Change. CRC Press. Hood, W.C., Aslan, A. and Betton, C., 2014. Aftermath of a stream capture: Cactus Park lake spillover and the origin of East Creek, Uncompahgre Plateau, western Colorado. Geosphere, 10(3): 447-461. Hua, B. and Shaoyu, W., 2015. An Improved Classification Algorithm Applied on Landslide Dam Disaster Events Detection. Huang, R., 2009. Some catastrophic landslides since the twentieth century in the southwest of China. Landslides, 6(1): 69-81. Huang, R. and Fan, X., 2013. The landslide story. Nature Geoscience, 6(5): 325. Hungr, O., Evans, S.G. and Hutchinson, I.N., 2001. A Review of the Classification of Landslides of the Flow Type. Environmental & Engineering Geoscience, 7(3): 221-238. Hungr, O., Leroueil, S. and Picarelli, L., 2014. The Varnes classification of landslide types, an update. Landslides, 11(2): 167-194. Hutchinson, J.N., 2006. Massive rock slope failure: perspectives and retrospectives on state-of-the-art, Landslides from Massive Rock Slope Failure. Springer, pp. 619662. Ischuk, A.R., 2011. Usoi rockslide dam and lake Sarez, Pamir mountains, Tajikistan, Natural and Artificial Rockslide Dams. Springer, pp. 423-440. Iverson, R.M., 1997. The physics of debris flows. Reviews of geophysics, 35(3): 245296. Iverson, R.M., George, D.L., Allstadt, K., Reid, M.E., Collins, B.D., Vallance, J.W., Schilling, S.P., Godt, J.W., Cannon, C.M. and Magirl, C.S., 2015. Landslide mobility and hazards: implications of the 2014 Oso disaster. Earth and Planetary Science Letters, 412: 197-208. 77
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Ivy-Ochs, S., Poschinger, A.V., Synal, H.A. and Maisch, M., 2009. Surface exposure dating of the Flims landslide, Graubünden, Switzerland. Geomorphology, 103(1): 104-112. Jansen, J.D., 2006. Flood magnitude–frequency and lithologic control on bedrock river incision in post-orogenic terrain. Geomorphology, 82(1-2): 39-57. Jansen, J.D., Fabel, D., Bishop, P., Xu, S., Schnabel, C. and Codilean, A.T., 2011. Does decreasing paraglacial sediment supply slow knickpoint retreat? Geology, 39(6): 543-546. Jermyn, C. and Geertsema, M., 2015. An Overview of Some Recent Large Landslide Types in Nahanni National Park, Northwest Territories, Canada, Engineering Geology for Society and Territory-Volume 1. Springer, pp. 315-320. Jibson, R.W., Harp, E.L., Schulz, W. and Keefer, D.K., 2006. Large rock avalanches triggered by the M 7.9 Denali Fault, Alaska, earthquake of 3 November 2002. Engineering geology, 83(1-3): 144-160. Jiskoot, H., 2011. Long‐runout rockslide on glacier at Tsar Mountain, Canadian Rocky Mountains: potential triggers, seismic and glaciological implications. Earth Surface Processes and Landforms, 36(2): 203-216. King, J., Loveday, I. and Schuster, R.L., 1989. The 1985 Bairaman landslide dam and resulting debris flow, Papua New Guinea. Quarterly Journal of Engineering Geology and Hydrogeology, 22(4): 257-270. King, P.R., 2000. Tectonic reconstructions of New Zealand: 40 Ma to the present. New Zealand Journal of Geology and Geophysics, 43(4): 611-638. Kingma, J.T., 1959. The tectonic history of New Zealand. New Zealand journal of geology and geophysics, 2(1): 1-55. Kirby, E., Whipple, K.X., Burchfiel, B.C., Tang, W., Berger, G., Sun, Z. and Chen, Z., 2000. Neotectonics of the Min Shan, China: Implications for mechanisms driving Quaternary deformation along the eastern margin of the Tibetan Plateau. Geological Society of America Bulletin, 112(3): 375-393. Kirby, E., Whipple, K.X., Tang, W. and Chen, Z., 2003. Distribution of active rock uplift along the eastern margin of the Tibetan Plateau: Inferences from bedrock channel longitudinal profiles. Journal of Geophysical Research: Solid Earth, 108(B4). Koi, T., Hotta, N., Ishigaki, I., Matuzaki, N., Uchiyama, Y. and Suzuki, M., 2008. Prolonged impact of earthquake-induced landslides on sediment yield in a mountain watershed: The Tanzawa region, Japan. Geomorphology, 101(4): 692702. Kondolf, G.M., 1997. PROFILE: hungry water: effects of dams and gravel mining on river channels. Environmental management, 21(4): 533-551. Korchevskiy, V.F., Kolichko, A.V., Strom, A.L., Pernik, L.M. and Abdrakhmatov, K.E., 2011. Utilisation of data derived from large-scale experiments and study of natural blockages for blast-fill dam design, Natural and artificial rockslide dams. Springer, pp. 617-637. Korup, O., 2002. Recent research on landslide dams-a literature review with special attention to New Zealand. Progress in Physical Geography, 26(2): 206-235. Korup, O., 2004. Geomorphometric characteristics of New Zealand landslide dams. Engineering Geology, 73(1-2): 13-35. Korup, O., 2005a. Geomorphic hazard assessment of landslide dams in South Westland, New Zealand: fundamental problems and approaches. Geomorphology, 66(1-4): 167-188. Korup, O., 2005b. Geomorphic imprint of landslides on alpine river systems, southwest New Zealand. Earth Surface Processes and Landforms, 30(7): 783-800. 78
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Korup, O., 2006. Rock-slope failure and the river long profile. Geology, 34(1): 45-48. Korup, O., 2011. Rockslide and rock avalanche dams in the Southern Alps, New Zealand, Natural and Artificial Rockslide Dams. Springer, pp. 123-145. Korup, O., 2012. Earth's portfolio of extreme sediment transport events. Earth-Science Reviews, 112(3-4): 115-125. Korup, O. and Clague, J.J., 2009. Natural hazards, extreme events, and mountain topography. Quaternary Science Reviews, 28(11-12): 977-990. Korup, O. and Montgomery, D.R., 2008. Tibetan plateau river incision inhibited by glacial stabilization of the Tsangpo gorge. Nature, 455(7214): 786. Korup, O., Montgomery, D.R. and Hewitt, K., 2010. Glacier and landslide feedbacks to topographic relief in the Himalayan syntaxes. Proceedings of the National Academy of Sciences, 107(12): 5317-5322. Korup, O., Strom, A.L. and Weidinger, J.T., 2006. Fluvial response to large rock-slope failures: Examples from the Himalayas, the Tien Shan, and the Southern Alps in New Zealand. Geomorphology, 78(1-2): 3-21. Korup, O. and Tweed, F., 2007. Ice, moraine, and landslide dams in mountainous terrain. Quaternary Science Reviews, 26(25-28): 3406-3422. Lagmay, A.M.F., de Vries, B.v.W., Kerle, N. and Pyle, D.M., 2000. Volcano instability induced by strike-slip faulting. Bulletin of Volcanology, 62(4-5): 331-346. Lee, C.F. and Dai, F.C., 2011. The 1786 Dadu River Landslide Dam, Sichuan, China, Natural and Artificial Rockslide Dams. Springer, pp. 369-388. Leeder, M.R., Harris, T. and Kirkby, M.J., 1998. Sediment supply and climate change: implications for basin stratigraphy. Basin research, 10(1): 7-18. Lefebvre, G., Rosenberg, P., Paquette, J. and Lavallée, J.G., 1991. The September 5, 1987, landslide on the La Grande River, James Bay, Quebec, Canada. Canadian Geotechnical Journal, 28(2): 263-275. Li, T., 1994. Landslide disasters and human responses in China. Mountain Research and Development: 341-346. Li, T., Schuster, R.L. and Wu, J., 1986. Landslide dams in south-central China, Landslide dams: processes, risk, and mitigation. ASCE, pp. 146-162. Locat, A., Demers, D., Locat, P. and Geertsema, M., 2017. Sensitive clay landslides in Canada, Proceedings of the 70th Canadian Geotechnical Conference, Ottawa, Paper. Luknitskiy, P.N., 1955. Travel over the Pamirs. Molodaya Gvardia Publishers, Moscow. Macías, J.L., Capra, L., Scott, K.M., Espíndola, J.M., García-Palomo, A. and Costa, J.E., 2004. The 26 May 1982 breakout flows derived from failure of a volcanic dam at El Chichón, Chiapas, Mexico. Geological Society of America Bulletin, 116(1-2): 233-246. Malamud, B.D., Turcotte, D.L., Guzzetti, F. and Reichenbach, P., 2004. Landslide inventories and their statistical properties. Earth Surface Processes and Landforms, 29(6): 687-711. Mason, K., 1929. Indus floods and Shyok glaciers. Himalayan Journal, 1: 10-29. Massey, C., Townsend, D., Dellow, G., Lukovic, B., Rosser, B., Archibald, G., Villeneuve, M., Davidson, J., Jones, K., Morgenstern, R. and et al., 2018. Kaikoura Earthquake Short-Term Project: landslide inventory and landslide dam assessments. GNS Science report; 2018/19, GNS Science, Lower Hutt (NZ). Mather, A., 2003. Book Review: Flood and megaflood processes and deposits: recent and ancient examples. The Holocene, 13(6): 954-955. Melekestsev, I.V., Dirksen, O.V. and Girina, O.A., 1999. A giant landslide-explosion circue and debris avalanche at Bakening volcano, Kamchatka. Journal of Volcanology and Seismology, 20(3): 265-279. 79
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Meyer, W., Sabol, M.A. and Schuster, R., 1986. LANDSLIDE DAMMED LAKES AT MOUNT ST. HELENS, WASHINGTON, Landslide Dams: Processes, Risk, and Mitigation. Proceedings of a Session in Conjunction with the ASCE Convention., pp. 21-41. Miall, A.D., 2013. Principles of sedimentary basin analysis. Springer Science & Business Media. Miller, B., Dufresne, A., Geertsema, M., Atkinson, N., Evensen, H. and Cruden, D., 2018. Longevity of dams from landslides with sub-channel rupture surfaces, Peace River region, Canada. Geoenvironmental Disasters, 5(1): 1. Miller, B.G. and Cruden, D.M., 2002. The Eureka River landslide and dam, Peace River Lowlands, Alberta. Canadian Geotechnical Journal, 39(4): 863-878. Montgomery, D.R. and Buffington, J.M., 1997. Channel-reach morphology in mountain drainage basins. Geological Society of America Bulletin, 109(5): 596-611. Montgomery, D.R., Hallet, B., Yuping, L., Finnegan, N., Anders, A., Gillespie, A. and Greenberg, H.M., 2004. Evidence for Holocene megafloods down the Tsangpo River gorge, southeastern Tibet. Quaternary Research, 62(2): 201-207. Nakamura, Y. and Glicken, H., 1988. Blast and debris-avalanche deposits of the 1888 eruption, Bandai volcano. Journal of Geography (Chigaku Zasshi), 97(4): 309316. Nash, T., Bell, D., Davies, T. and Nathan, S., 2008. Analysis of the formation and failure of Ram Creek landslide dam, South Island, New Zealand. New Zealand journal of geology and geophysics, 51(3): 187-193. Nash, T.R., 2003. Engineering geological assessment of selected landslide dams formed from the 1929 Murchison and 1968 Inangahua earthquakes. Neely, A.B., Bookhagen, B. and Burbank, D.W., 2017. An automated knickzone selection algorithm (KZ‐Picker) to analyze transient landscapes: Calibration and validation. Journal of Geophysical Research: Earth Surface, 122(6): 12361261. Nicoletti, P.G. and Parise, M., 2002. Seven landslide dams of old seismic origin in southeastern Sicily (Italy). Geomorphology, 46(3-4): 203-222. O'Connor, J.E. and Costa, J.E., 2004. The world's largest floods, past and present: their causes and magnitudes, 1254. US Geological Survey. Oesleby, T.W., 2005. Thick sediment fill in Unaweep Canyon: Implications for the history of the Uncompahgre Uplift, western Colorado, Geological Society of America Abstracts with Programs, pp. 16. Ohmori, H., 1992. Dynamics and erosion rate of the river running on a thick deposit supplied by a large landslide. Zeitschrift fur Geomorphologie Neue Folge, 36: 129-140. Ouimet, W.B., Whipple, K.X., Royden, L.H., Sun, Z. and Chen, Z., 2007. The influence of large landslides on river incision in a transient landscape: Eastern margin of the Tibetan Plateau (Sichuan, China). Geological Society of America Bulletin, 119(11-12): 1462-1476. Peng, M. and Zhang, L.M., 2012. Breaching parameters of landslide dams. Landslides, 9(1): 13-31. Penna, I.M., Derron, M.-H., Volpi, M. and Jaboyedoff, M., 2013. Analysis of past and future dam formation and failure in the Santa Cruz River (San Juan province, Argentina). Geomorphology, 186: 28-38. Penna, I.M., Hermanns, R.L. and Folguera, A., 2008. Remoción en masa y colapso catastrófico de diques naturales generados en el frente orogénico andino (36º38ºs): los casos Navarrete y Río Barrancas. Revista de la Asociación Geológica Argentina, 63(2): 172-180. 80
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Perucca, L.P., Espejo, K., Angillieri, M.Y.E., Rothis, M., Tejada, F. and Vargas, M., 2018. Neotectonic controls and stream piracy on the evolution of a river catchment: a case study in the Agua de la Peña River basin, Western Pampean Ranges, Argentina. Journal of Iberian Geology, 44(2): 207-224. Petley, D.N., Hearn, G.J., Hart, A., Rosser, N.J., Dunning, S.A., Oven, K. and Mitchell, W.A., 2007. Trends in landslide occurrence in Nepal. Natural hazards, 43(1): 2344. Pollet, N., Cojean, R., Couture, R., Schneider, J.-L., Strom, A.L., Voirin, C. and Wassmer, P., 2005. A slab-on-slab model for the Flims rockslide (Swiss Alps). Canadian geotechnical journal, 42(2): 587-600. Porter, S.C. and Orombelli, G., 1981. Alpine Rockfall Hazards: Recognition and dating of rockfall deposits in the western Italian Alps lead to an understanding of the potential hazards of giant rockfalls in mountainous regions. American Scientist, 69(1): 67-75. Pratt-Sitaula, B., Garde, M., Burbank, D.W., Oskin, M., Heimsath, A. and Gabet, E., 2007. Bedload-to-suspended load ratio and rapid bedrock incision from Himalayan landslide-dam lake record. Quaternary Research, 68(1): 111-120. Pulgarín, B., Capra, L., Cepeda, H. and Macías, J.L., 2004. Late Pleistocene deposits associated with a southern flank collapse of the Nevado del Huila volcanic complex (Colombia). Acta vulcanologica, 16(1/2): 1000-1022. Read, S.A.L., Beetham, R.D. and Riley, P.B., 1992. Lake Waikaremoana barrier–a large landslide dam in New Zealand, Proceedings of the Sixth International Symposium on Landslides, pp. 1481-1487. Reading, H.G. and Levell, B.K., 1996. Controls on the sedimentary rock record. Sedimentary environments: processes, facies and stratigraphy: 5-36. Rickenmann, D., 2005. Runout prediction methods, Debris-flow hazards and related phenomena. Springer, pp. 305-324. Ruegg, J.C., Kasser, M., Tarantola, A., Lepine, J.C. and Chouikrat, B., 1982. Deformations associated with the El Asnam earthquake of 10 October 1980: Geodetic determination of vertical and horizontal movements. Bulletin of the Seismological Society of America, 72(6A): 2227-2244. Salvati, P., Bianchi, C., Rossi, M. and Guzzetti, F., 2010. Societal landslide and flood risk in Italy. Natural Hazards and Earth System Sciences, 10(3): 465-483. Sanhueza-Pino, K., Korup, O., Hetzel, R., Munack, H., Weidinger, J.T., Dunning, S., Ormukov, C. and Kubik, P.W., 2011. Glacial advances constrained by 10 Be exposure dating of bedrock landslides, Kyrgyz Tien Shan. Quaternary Research, 76(3): 295-304. Schauwecker, S., Rohrer, M., Acuña, D., Cochachin, A., Dávila, L., Frey, H., Giráldez, C., Gómez, J., Huggel, C., Jacques-Coper, M., Loarte, E., Salzmann, N. and Vuille, M., 2014. Climate trends and glacier retreat in the Cordillera Blanca, Peru, revisited. Global and Planetary Change, 119: 85-97. Scheidegger, A.E., 1973. On the prediction of the reach and velocity of catastrophic landslides. Rock Mechanics and Rock Engineering, 5(4): 231-236. Schneider, J.F., Gruber, F.E. and Mergili, M., 2013. Recent cases and geomorphic evidence of landslide-dammed lakes and related hazards in the mountains of Central Asia, Landslide science and practice. Springer, pp. 57-64. Schuster, R.L., 1986. Landslide dams: processes, risk, and mitigation. ASCE. Schuster, R.L., 2000. Dams built on pre-existing landslides, ISRM International Symposium. International Society for Rock Mechanics and Rock Engineering. Schuster, R.L., 2006. IMPACTS OF LANDSLIDE DAMS ON MOUNTAIN VALLEY MORPHOLOGY. In: S.G. Evans, G.S. Mugnozza, A. Strom and R.L. Hermanns 81
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
(Editors), Landslides from Massive Rock Slope Failure. Springer Netherlands, Dordrecht, pp. 591-616. Schuster, R.L. and Alford, D., 2004. Usoi landslide dam and lake sarez, Pamir mountains, Tajikistan. Environmental & Engineering Geoscience, 10(2): 151-168. Schuster, R.L. and Costa, J.E., 1986. PERSPECTIVE ON LANDSLIDE DAMS, pp. 120. Schuster, R.L. and Crandell, D.R., 1984. Catastrophic debris avalanches from volcanoes, Fourth Int Symp Landslides Proc Toronto, pp. 567-572. Schwanghart, W. and Scherler, D., 2017. Bumps in river profiles: uncertainty assessment and smoothing using quantile regression techniques. Earth Surface Dynamics, 5(4). Shen, Z.-K., Sun, J., Zhang, P., Wan, Y., Wang, M., Bürgmann, R., Zeng, Y., Gan, W., Liao, H. and Wang, Q., 2009. Slip maxima at fault junctions and rupturing of barriers during the 2008 Wenchuan earthquake. Nature geoscience, 2(10): 718. Shroder, J.F., 1994. Book Review: Geology and tectonics of the Karakoram mountains. MP Searle. Wiley, New York, 1991. 358 pp., 218 illustrations, 1: 250,000 geology and structure map in color, slipcase. Price: $298. ISBN 0-471-92773-2. Geomorphology, 9: 79-81. Shroder Jr, J.F., 1998. Slope failure and denudation in the western Himalaya. Geomorphology, 26(1-3): 81-105. Shroder Jr, J.F. and Bishop, M.P., 1998. Mass movement in the Himalaya: new insights and research directions. Geomorphology, 26(1-3): 13-35. Shugar, D.H., Clague, J.J., Best, J.L., Schoof, C., Willis, M.J., Copland, L. and Roe, G.H., 2017. River piracy and drainage basin reorganization led by climate-driven glacier retreat. Nature Geoscience, 10(5): 370. Strecker, M.R. and Marrett, R., 1999. Kinematic evolution of fault ramps and its role in development of landslides and lakes in the northwestern Argentine Andes. Geology, 27(4): 307-310. Strom, A., 2006. Morphology and internal structure of rockslides and rock avalanches: grounds and constraints for their modelling, Landslides from Massive Rock Slope Failure. Springer, pp. 305-326. Strom, A., 2010. Landslide dams in Central Asia region. Journal of the Japan Landslide Society, 47(6): 309-324. Strom, A., 2013. Geological prerequisites for landslide dams’ disaster assessment and mitigation in Central Asia, Progress of Geo-Disaster Mitigation Technology in Asia. Springer, pp. 17-53. Strom, A., 2015. Natural river damming: climate-driven or seismically induced phenomena: basics for landslide and seismic hazard assessment, Engineering Geology for Society and Territory-Volume 2. Springer, pp. 33-41. Strom, A. and Abdrakhmatov, K., 2018. Rockslides and rock avalanches of Central Asia: distribution, morphology, and internal structure. Elsevier. Swanson, F.J., Fredriksen, R.L. and McCorison, F.M., 1982. Material transfer in a western Oregon forested watershed. Swanson, F.J., Oyagi, N. and Tominaga, M., 1986. Landslide dams in Japan, Landslide dams: processes, risk, and mitigation. ASCE, pp. 131-145. Tabata, S., Mizuyama, T. and Inoue, K., 2002. Landslide dams and disasters. Kokon Shoin: 205. Tacconi Stefanelli, C., Catani, F. and Casagli, N., 2015. Geomorphological investigations on landslide dams. Geoenvironmental Disasters, 2(1): 21. Tacconi Stefanelli, C., Segoni, S., Casagli, N. and Catani, F., 2016. Geomorphic indexing of landslide dams evolution. Engineering geology, 208: 1-10. 82
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Tacconi Stefanelli, C., Vilímek, V., Emmer, A. and Catani, F., 2018. Morphological analysis and features of the landslide dams in the Cordillera Blanca, Peru. Landslides, 15(3): 507-521. Umbal, J.V. and Rodolfo, K.S., 1996. The 1991 lahars of southwestern Mount Pinatubo and evolution of the lahar-dammed Mapanuepe Lake. Fire and mud: Eruptions and lahars of Mount Pinatubo, Philippines: Quezon, Philippines: 951-970. Valdiya, K.S., 1996. River piracy. Resonance, 1(5): 19-28. Varnes, D.J., 1978. Slope movement types and processes. Special report, 176: 11-33. von Poschinger, A., Wassmer, P. and Maisch, M., 2006. The Flims rockslide: history of interpretation and new insights, Landslides from massive rock slope failure. Springer, pp. 329-356. Walder, J.S. and O'Connor, J.E., 1997. Methods for predicting peak discharge of floods caused by failure of natural and constructed earthen dams. Water Resources Research, 33(10): 2337-2348. Walsh, L.S., Martin, A.J., Ojha, T.P. and Fedenczuk, T., 2012. Correlations of fluvial knickzones with landslide dams, lithologic contacts, and faults in the southwestern Annapurna Range, central Nepalese Himalaya. Journal of Geophysical Research: Earth Surface, 117(F1). Wang, F., Fan, X., Yunus, A.P., Siva Subramanian, S., Alonso-Rodriguez, A., Dai, L., Xu, Q. and Huang, R., 2019. Coseismic landslides triggered by the 2018 Hokkaido, Japan (M w 6.6), earthquake: spatial distribution, controlling factors, and possible failure mechanism. Landslides: 1-16. Wang, G., Furuya, G., Zhang, F., Doi, I., Watanabe, N., Wakai, A. and Marui, H., 2016. Layered internal structure and breaching risk assessment of the HigashiTakezawa landslide dam in Niigata, Japan. Geomorphology, 267: 48-58. Wang, X. and Han, M., 2014. Online sequential extreme learning machine with kernels for nonstationary time series prediction. Neurocomputing, 145: 90-97. Wartman, J., Montgomery, D.R., Anderson, S.A., Keaton, J.R., Benoît, J., dela Chapelle, J. and Gilbert, R., 2016. The 22 March 2014 Oso landslide, Washington, USA. Geomorphology, 253: 275-288. Waythomas, C.F. and Kaufman, D.S., 1991. Comment on:“Latest Pleistocene increase in wind intensity recorded in eolian sediments from central Alaska,” by N. Bigelow, JE Begét, and WR Powers. Quaternary Research, 36(3): 329-333. Waythomas, C.F. and Nye, C.J., 2001. Preliminary volcano-hazard assessment for Mount Spurr volcano, Alaska. 2331-1258, US Geological Survey. Wei, R., Zeng, Q., Davies, T., Yuan, G., Wang, K., Xue, X. and Yin, Q., 2018. Geohazard cascade and mechanism of large debris flows in Tianmo gully, SE Tibetan Plateau and implications to hazard monitoring. Engineering Geology, 233: 172-182. Weidinger, J.T., 2011. Stability and life span of landslide dams in the Himalayas (India, Nepal) and the Qin Ling Mountains (China), Natural and Artificial Rockslide Dams. Springer, pp. 243-277. Whipple, K.X., Wobus, C., Crosby, B., Kirby, E. and Sheehan, D., 2007. New tools for quantitative geomorphology: extraction and interpretation of stream profiles from digital topographic data. GSA short course, 506. Whitehouse, I.E. and Griffiths, G.A., 1983. Frequency and hazard of large rock avalanches in the central Southern Alps, New Zealand. Geology, 11(6): 331-334. Wohl, E.E. and Pearthree, P.P., 1991. Debris flows as geomorphic agents in the Huachuca Mountains of southeastern Arizona. Geomorphology, 4(3-4): 273-292.
83
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Worni, R., Huggel, C., Clague, J.J., Schaub, Y. and Stoffel, M., 2014. Coupling glacial lake impact, dam breach, and flood processes: A modeling perspective. Geomorphology, 224: 161-176. Xu, Q., Fan, X.-M., Huang, R.-Q. and Van Westen, C., 2009. Landslide dams triggered by the Wenchuan Earthquake, Sichuan Province, south west China. Bulletin of engineering geology and the environment, 68(3): 373-386. Yi, G.X., Wen, X.Z. and Su, Y.J., 2008. Study on the potential strong‐earthquake risk for the eastern boundary of the Sichuan‐Yunnan active faulted‐block, China. Chinese Journal of Geophysics, 51(6): 1151-1158. Yin, Y., Cheng, Y., Liang, J. and Wang, W., 2016. Heavy-rainfall-induced catastrophic rockslide-debris flow at Sanxicun, Dujiangyan, after the Wenchuan Ms 8.0 earthquake. Landslides, 13(1): 9-23. Yin, Y., Wang, F. and Sun, P., 2009. Landslide hazards triggered by the 2008 Wenchuan earthquake, Sichuan, China. Landslides, 6(2): 139-152. Yunus, A.P., 2016. Geomorphic and lithologic control on bedrock channels in drainage basins of the Western Arabian Peninsula. Arabian Journal of Geosciences, 9(2): 133. Yunus, A.P., Oguchi, T. and Hayakawa, Y.S., 2016. Remote identification of fluvial knickzones and their imprints on landscape morphology in the passive margins of Western Arabia. Journal of Arid Environments, 130: 14-29. Zeitler, P.K., 1985. Cooling history of the NW Himalaya, Pakistan. Tectonics, 4(1): 127151. Zhang, H., Zhang, P. and Fan, Q., 2011. Initiation and recession of the fluvial knickpoints: A case study from the Yalu River-Wangtian’e volcanic region, northeastern China. Science China Earth Sciences, 54(11): 1746.
84
Journal Pre-proof
Competing interests
Jo
ur n
al
Pr
e-
pr
oo
f
Authors declare no competing interests.
85
Xuanmei Fan, Anja Dufresne, Srikrishnan Siva Subramanian, Alexander Strom, Reginald Hermanns, Carlo Tacconi Stefanelli, Kenneth Hewitt, Ali P. Yunus, Stuart Dunning, Lucia Capra, Marten Geertsema, Brendan Miller, Nicola Casagli, John D. Jansen, Qiang Xu PII:
S0012-8252(19)30476-3
DOI:
https://doi.org/10.1016/j.earscirev.2020.103116
Reference:
EARTH 103116
To appear in:
Earth-Science Reviews
Received date:
15 July 2019
Revised date:
5 February 2020
Accepted date:
5 February 2020
Please cite this article as: X. Fan, A. Dufresne, S.S. Subramanian, et al., The formation and impact of landslide dams – State of the art, Earth-Science Reviews(2020), https://doi.org/ 10.1016/j.earscirev.2020.103116
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier.
Journal Pre-proof
The formation and impact of landslide dams – state of the art Xuanmei Fan1, Anja Dufresne2* , Srikrishnan Siva Subramanian1* , Alexander Strom3 , Reginald Hermanns4,5, Carlo Tacconi Stefanelli 6, Kenneth Hewitt7, Ali P. Yunus1, Stuart Dunning8, Lucia Capra9, Marten Geertsema10, Brendan Miller10 , Nicola Casagli 6, John D. Jansen1,11, Qiang Xu1 1
State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of
Technology, 610059, Chengdu, Sichuan, People’s Republic of China 2
Engineering Geology and Hydrogeology, RWTH-Aachen University, Lochnerstr. 4-20, 52066 Aachen,
Germany Geodynamics Research Centre, Volkolamskoe Shosse 2, 125080 Moscow, Russia
4
Geohazards and Earth Observation, Geological Survey of Norway, Trondheim, N-7491, Norway
5
Department of Geoscience and Petroleum, Norwegian University of Science and Technology, Norway
6
Department of Earth Sciences, University of Firenze, Via La Pira, 4, Florence 50121, Italy
7
Department of Geography and Environmental Studies, Wilfried Laurier University, 75 University Avenue
pr
oo
f
3
8
e-
West, Waterloo, Ontario N2L 3C5, Canada
School of Geography, Politics and Sociology, Newcastle University, Newcastle upon Tyne, NE1 7RU,
9
Pr
UK
Centro de Geosciencias, Universidad Nacional Autónoma de México, Mexico City
10
Ministry of Forests, Lands, Natural Resource Operations, and Rural Development, Prince George,
al
British Columbia, Canada 11
ur n
GFÚ Institute of Geophysics, Czech Academy of Sciences, Prague, Czechia.
*Corresponding author: Dr Anja Dufresne, email: [email protected], Dr.
Jo
Srikrishnan Siva Subramanian, email: [email protected]
Highlights:
Compilation of a new worldwide database of 410 landslide dams of > 1 million m3 in volume and design of a template for future data collection. A preliminary landslide dam stability classification matrix combining topographic setting and the sedimentological features. Comprehensive review and evaluation of landslide dam stability criteria and geomorphic indices. Landscape impacts of landslide dams discussed on different time scales.
1
Journal Pre-proof
Abstract The blocking of river courses by mass movements is very common in mountainous areas with deep and narrow valleys. Landslide dams may pose serious threats to people and their livelihoods downstream in the case of abrupt dam failure. Since the publication of benchmark reviews of Costa and Schuster (1988) and Korup (2002), there is a growing number of studies focusing on the formation, stability, and short-term impacts of landslide dams. This review combines the insights of all these studies, builds
oo
f
on current concepts of landslide dams, and suggests ways to unify terminologies and classifications. We furthermore present a new worldwide database compiled from
pr
literature data. It contains 410 landslide dams >1 million m3 in volume that were formed
e-
since 1900 since these have the most complete data entries. These data show that
Pr
dam longevity is, among other factors, correlated with the type of landslide forming the dam. Those formed by rock/debris avalanches and rockslides have longest lifespans.
al
However, the influence of landslide type or material on dam longevity decreases with
ur n
time after dam formation. To ensure consistency in the next database generation, we suggest guidelines for data collection to provide a solid basis for evaluating dam
Jo
stability and governing factors. A preliminary classification matrix for landslide dam stability that combines topographic setting and the internal structure of the dam body is another outcome of our review. Furthermore, an evaluation of the various geomorphic stability indices proposed in the literature regarding their suitability and limitations in assessing dam formation and stability shows that they predict the probability of dam formation reasonably well, but that their application to longevity estimates requires further assessment. The geomorphic impacts of landslide dams in the short-, mediumand long-term are summarized and illustrated with key examples. Finally, for a better understanding of the factors controlling dam stability, we recommend to (1) include dam 2
Journal Pre-proof composition and sedimentary structures in future case studies, (2) maintain and update the worldwide database for sound statistical analyses, (3) refine landslide dam stability indices and test them for different landslide types, and (4) study hazard cascades related to multiple dams in one watershed. For long-term landscape evolution studies, we suggest to (5) quantify terrestrial sediment flux related to landslide dams, (6) detect ancient landslide dams in river profiles, and (7) further exploit the sediment archives in former impoundment areas.
oo
f
Keywords: worldwide database; classification; stability; geomorphological impacts;
pr
landscape evolution
e-
1 Introduction
Natural damming of rivers by mass movements and their consequences pose a major
Pr
hazard in mountai nous areas with deep, narrow valleys (Capra, 2006; Costa and Schuster, 1988; Ermini and Casagli, 2003; Evans et al., 2011a; Fan et al., 2019a;
al
Hermanns, 2013; Hewitt, 1988; Korup, 2002; Korup and Tweed, 2007; Swanson et al.,
ur n
1986). Though natural river-damming events are induced not only by landslides (there are moraine dams and dams formed by fault ruptures), those formed by large landslides
Jo
are most hazardous (Mason, 1929; Shroder Jr, 1998), and this review is focused only on these. Large landslide dams often result in more complex threats to people than the landslides themselves, both in space and time, by inundating upstream and potentially flooding downstream areas (Davies and Korup, 2010; Groeber, 1916; Penna et al., 2013), making dam stability assessment crucial. The stability of landslide dams largely depends on the composition and geotechnical properties of the dam material (Casagli and Ermini, 1999; Korup, 2002; Schuster, 1986; Walder and O'Connor, 1997). However, until the landslide dam is cut—artificially or naturally—its internal geological structure and material properties are unknown in most cases, though some can be 3
Journal Pre-proof extrapolated from analogue or empirical case studies. Despite clear identification of the factors controlling landslide dam formation and stability, the actual mechanism of formation or failure in relation to the material composition, dam geometry, the impounded catchment area, lake volume, and water inflow rates still remain unclear and warrant a thorough review. Hazards and risks related to these dams have long been recognized and are documented in many historical accounts of catastrophic floods from natural dam failures
oo
f
(Dai et al., 2005; Delaney and Evans, 2015; Fan et al., 2012a; 2012b; 2012c; 2013; Groeber, 1916; Lee and Dai, 2011; O'Connor and Costa, 2004; Peng and Zhang,
pr
2012). Changes in climatic and meteorological conditions often play a part in
e-
‘unexpected’ dam failures. Examples include the Ram Creek landslide dam (New Zealand), which remained stable for 13 years before failure after heavy rain (Harrison et
Pr
al., 2015) or the Rio Barrancas rockslide dam in Argentina, which failed 425 years after
al
its formation due to increased inflow of glacial meltwater during the hottest austral summer on record and an antecedent, unusually wet winter (Evans et al., 2011a). The
ur n
many unknowns in assessing both immediate and long-term stability complicate the management of potential hazards from landslide dams.
Jo
Benchmark reviews on landslide dams by Costa and Schuster (1988) and Korup (2002) addressed the complex interactions between landslide morphometric characteristics and magnitude, geomorphological parameters of the valley, and hydrological parameters of the river at a given location. Since then, a large body of literature on specific aspects of landslide dams has emerged. The most widely recognized works, which are relevant for the preparation of our review, are summarized in Table 1 , including a book edited by Evans et al. (2011a) that presents a review focussed on natural and artificial rockslide dams mostly with volumes in excess of 106 m3 .
4
Journal Pre-proof Efforts were also made in creating both worldwide and regional landslide dam inventories, but there are still some difficulties. For example, inventories encompass dams formed by all types of landslides rather than detailed breakdowns of the subtypes. The other challenge is that the wide range of terminology and landslide dam classification schemes without a global standard makes cross-comparison of methods and data difficult. To understand dam formation and failure, the types of mass movements that potentially form landslide dams should be categorized based on their
oo
f
varying characteristic sizes and geotechnical properties. With this review we aim at offering a unified landslide dam terminology, propose a new
pr
classification scheme based on existing ones, compile existing landslide dam
e-
inventories and thereby create a worldwide database, present the state of the art on formation and stability, and, review the impact of landslide dams over multiple
Pr
timescales (see 1.1 Structure of the review).
Focus
ur n
Previous benchmark reviews
al
Table 1 A summary of key papers, reviews and meta-analyses on landslide dams Key references Costa and Schuster (1988); Korup (2002); Korup and Tweed (2007); Evans et al. (2011a) Volcanic dams: Capra (2007); Capra (2011);
formed by different types of landslides and
Rockslide and rock avalanche dams: Dufresne et al.
Jo
Papers focus on studying landslide dams in different materials
(2018); Hermanns et al. (2006); Hermanns et al. (2011a); Hermanns et al. (2011b); Hewitt (2009a); Hewitt (2011) Unconsolidated sediments: Geertsema (1998)
Worldwide landslide dam datasets;
Schuster and Costa (1986), 187 cases;
event-based landslide dam inventory of the
Costa and Schuster (1991), 463 cases;
2008 Wenchuan earthquake;
Ermini and Casagli (2003), 350 cases;
other regional datasets are reviewed in
Peng and Zhang (2012), 1239 cases
Section 3
Strom and Abdrakhmatov (2018), 190 cases Fan et al. (2012a); Fan et al. (2012b); Fan et al. (2012c), 828 cases from the Wenchuan earthquake
Empirical indices predicting landslide dam
Swanson et al. (1986), Ermini (2003b), Dal Sasso et
formation
al. (2014), Tacconi Stefanelli et al. (2016) 5
Journal Pre-proof Empirical classification schemes and
Casagli and Ermini (1999); Dal Sasso et al. (2014);
geomorphic indices for landslide dam
Dong et al. (2009); Ermini (2003a); Ermini and
stability
Casagli (2003); Korup (2002); Korup (2004); Tacconi Stefanelli et al. (2016)
Landscape implications of landslide dams,
Hewitt (2002); Korup (2005a); Schuster (2006);
including their impacts on valley
Korup (2012); Jansen (2006); Korup et al. (2006)
morphology, sediment transport, fluvial incision etc.
1.1
Structure of the review
This review covers recent developments in understanding the formation and impacts of
oo
f
landslide dams. The lifecycle of a landslide dam involves the origin (the landslide),
landscape impacts throughout all stages.
pr
survival (longevity or stability), breach (sudden failure or slow erosion), and related
e-
The terminology used for characterizing landslide dams, their lakes and stability are
Pr
discussed in Section 2. Researchers around the world use different terminologies and definitions to characterize landslide dams, but no classification scheme with a near
al
standardized use comparable to the one devised by Varnes (1978) and Hungr et al.
ur n
(2014) for landslides themselves is available. In Section 2, an attempt is made to review, condense and supplement the common terminology and dam classifications ,
Jo
and to provide definitions that will make scientific communication easier if applied widely. Additionally, a complimentary classification scheme is presented based on the internal structure of landslide deposits that affects dam stability, although the scheme may not apply in many cases until the landslide dam is eroded. Section 3 presents an analysis of our worldwide landslide dam database. We collected and reviewed data from all known landslide dam hotspots around the world: the Tibetan mountains, the Himalayan regions, Central Asia including Pamir and Tien Shan mountains, tectonically active mountain ranges of Japanese and New Zealand, the European Alps, the Rocky Mountains and the Andes. Some event-based inventories 6
Journal Pre-proof are also included, such as the 2008 Wenchuan (China), the 2015 Gorkha (Nepal), and the 2016 Kaikoura (New Zealand) earthquakes. We identify inconsistencies and problems in the complied inventories and recommend a template for future data collection—provided as supplementary material. Various aspects of landslide dam stability are summarized in Section 4. We give a synopsis of the geomorphic stability indices proposed in the literature and illustrate their applicability based on our new database.
oo
f
A review of the geomorphic impacts of landslides dams in short-, medium-, and longterms are the focus of Section 5. Under short-term impacts, importance has been given
pr
to dam formation and stability within the first year after formation (often the most critical
e-
for immediate hazard assessment). River morphological changes and sediment fluxes lasting years to decades are considered medium-term and presented with key
Pr
examples. The long-term geomorphic impacts through centuries and millennia are
al
presented from a geological perspective, considering also the factors causing sudden collapse of dams that remained (apparently stable) for centuries to millennia.
ur n
Despite the decades-long research and associated advances in our understanding of the formation and failure of natural dams, many challenges still remain. Section 6
Jo
summarises the key research challenges and suggests future work to answer them.
2 Terminology and classification 2.1
Terms characterizing landslide dam and stability
As with landslide processes, landslide dam terminology varies by author, region and research focus, with different classifications existing in parallel. A common-sense approach is required when searching for literature pertaining to landslide dams and landslide-dammed lakes, and integrating the reported metrics. Terms common in studies describing landslide dams include landslide blockage, stream/river blockage, 7
Journal Pre-proof and, landslide barrier. Lakes that form behind a landslide dam have been variably termed landslide lake, impoundment, landslide-dammed lakes, landslide-barrier lake, reservoir, backwater lake, and earthquake -dammed lake. The term earthquakedammed lake, however, ignores the fact that river damming might be caused by coseismic faulting or folding, as during the 1980 El-Asnam earthquake in Algeria (Ruegg et al., 1982). Though ‘landslide lake’ seems to be the commonly accepted choice for lakes that formed on top of a landslide or landslide dam rather than behind it upstream,
oo
f
we recommend the term ‘supra-landslide lake’ (following the definition of 'super lake'; Crozier and Pillans, 1991) for future use since it clearly separates it from any lake
pr
formed behind the blockage and is more consistent with other fields (e.g. supra-glacial
e-
lake). An impoundment is any body of water formed by confinement. This term is, however, commonly used to describe the area upstream of a dam regardless of the
Pr
presence of a lake. The term ‘reservoir’ is often found in the context of processes
al
affecting the lake, akin to hydropower dam studies, whereas ‘backwater lake’ or ‘landslide-dammed lake’ are most commonly found in geomorphological studies.
ur n
Herein, we adopt ‘impoundment’ for the upstream area, and ‘landslide-dammed lake’ when processes pertaining to the water body are discussed and in all general,
Jo
descriptive situations.
The empirical classification schemes of landslide dam stability by Ermini and Casagli (2003), Korup (2004), and Tacconi Stefanelli et al. (2016) currently in use (Table 2) build upon the initial definition by Casagli and Ermini (1999), which is restricted to landslide dams that have either failed catastrophically or have not failed. In these studies, stability is a momentary definition of the state of the dam and the landslidedammed lake at the time of observation but does not include the duration of non-failure (longevity). Per that definition, a landslide dam is termed stable if a landslide-dammed lake still exists or is filled in with sediments at the time of the analysis (Table 3). The 8
Journal Pre-proof latter indicates that the dam was able to retain the lake water (maintaining in- and outflow balance by seepage through the dam or flow along a spillway) and allowed continued sedimentation into the lake until the lake was silted up. Those classified ‘not stable’ have breached catastrophically, evidenced by deep gullies, an impoundment devoid of significant sediments, and/or with tell-tale erosional features within left-over sediments that indicate rapid drawdown, and/or flood sediments downstream – although evidence of outburst floods are often superimposed or eroded by later valley
f
processes. We have recognized that a deep gulley cut through a dam, or bypassing a
oo
dam, can also be formed by long-term gradual erosion as well – many features are not
pr
conclusively diagnostic for sudden failure or for slower erosion. Between ’stable’ and
e-
‘unstable’ exists a large number of cases where the cause of lake drainage (catastrophic or gradual) remains unknown.
Pr
Table 2 Terms characterizing landslide dam stability Casagli and Ermini (1999)
Artificially controlled Slow erosion
Existing
Includes all the lakes extinguished due to progressive filling Includes lakes existing at present
Jo
Filling
Includes all the cases of a catastrophic dam collapse. Includes all the cases where the remedial works drastically changed the natural evolution of the phenomenon. Includes all the lakes extinguished because of slow progressive erosion of the dam without catastrophic events.
ur n
Collapsed dam
al
Includes all the cases of partial blockage, channel deviation, erosion of the landslide toe, in which the formation of a complete dam is prevented.
Dam not formed*
Ermini and Casagli (2003) Unstable dam Stable dam:
Landslide dam that has undergone erosion or collapse leading to a catastrophic breach, with the subsequent release of the impounded lake waters. Landslide dam that has remained stable and has not encountered a catastrophic breach thus still impounding an existent or relict lake.
Korup (2004) Stable The landslide-dammed lake persists for over a decade. Tacconi Stefanelli et al. (2016) The landslide reached the riverbed, but only a partial damming is realized (no lake Not-formed* formed). Over times that can range from hours to centuries, the dam collapsed or was Formed-unstable breached by the river, including the artificially breached or removed landslide dams. The blockage is complete with the formation of a dam and a lake, which are still Formed-stable existing or disappeared for sediment filling. The dam could have been overtopped during its life, but no total failure or destructive flooding wave occurred. Note: *Even though these landslides did not create full river blockages, they are considered by the respective authors, and included herein, since they may still alter the ri ver flow and sediment delivery or form partial blockages, 9
Journal Pre-proof which could be of relevance in risk assessment and is of interest for inventory analyses (i.e. factors relevant to the full dam formation and failure statistics).
A potential history of multiple breaches or predictions of future stability (or longevity) is rarely taken into hazard analyses (Hermanns et al., 2013). For a landslide dam to remain stable, it needs to be able to hold the impounded water indefinitely and/or release it gradually so that no catastrophic flood may impact the downstream area. From a practical point of view, and for a first-order assessment of a dam’s potential
f
threat, we propose the use of ‘apparent’ stability (Table 3), which applies to any dam
oo
that still holds a lake, but which either has not been investigated yet or has been stable
pr
for a long enough time for any potential risks (e.g. due to external causes) to be forgotten (all red categories in Table 3). The only proven ‘stable’ dams are those in the
e-
category ‘infilled’ since these demonstrably maintained a lake without catastrophic
Pr
breaching. This classification scheme furthermore takes into consideration the potential for partial and episodic breaching. It should also prompt further investigations of the
al
landslide dam in question, and assist in communication between scientific disciplines
ur n
and stakeholders. Furthermore, it should minimize the risk of inferring long-term stability in situations that are not fully investigated.
Jo
Longevity in this study refers simply to how long a landslide dam has existed before its lake emptied, regardless of the cause (catastrophic breach or gradual erosion). Stability, on the other hand, relates to whether a dam is prone to sudden collapse—in the landslide dam literature, this is mainly applied to the momentary situation wherein dams have either failed or not at the time of data acquisition and publication. The time-scales before/until failure and those at which the effects of a landslide dam are noticeable in the landscape system are classed as follows in this paper. Short -term refers to the first days to months after formation, since this is the most crucial for hazard assessment. Most dams that have failed catastrophically did so within the first year after 10
Journal Pre-proof formation (Clague and Evans, 1994; Costa and Schuster, 1988; Fan et al., 2012c; Peng and Zhang, 2012); hence we consider up to one year as short-term. It is also the time of the most pronounced landscape impacts by the blockage itself, disrupting river flow, and the filling of the impoundment creating a new landslide-dammed lake. Medium-term encompasses years to decades, a time where mitigation measures might still be necessary, a landslide-dammed lake might still exist and catastrophic breaching could still be possible (e.g. through external causes like earthquakes or strong/prolonged
oo
f
rainfall). Long-term involves centuries to millennia, a time-scale in which the dam (or remnant) can still influence the landscape, but in which, from a human lifespan point of
pr
view, hazards are either diminished by erosion or, if a landslide-dammed lake still
e-
exists, vanished from the collective consciousness. Database analyses often, by necessity, deviate from these first-order timelines to either investigate processes at
al
Pr
higher time-resolution or because data availability determines the intervals for analysis.
Table 3 Classification scheme of apparent landslide dam stability
ur n
Infilled
Jo
Long-term stable
Long-term eroded dam
Short-term stable
Short-term eroded dam Dam breached
Impoundment infilled with sediment (i.e. retained landslide-dammed lake throughout its lifetime) and there is no evidence for any catastrophic breaching during its existence. Landslide dam that has remained stable and has not encountered a breach over centuries or longer and thus still impounds an existent or relict lake and no immediate change is expected. However, extreme weather / climatic conditions or other effects (e.g. landsliding into the landslide-dammed lake or unusually high inflow) might cause catastrophic failure. Entirely/partially eroded, i.e. did not retain the impoundment lake waters, but released it gradually over a hundred to thousands of years. Existing landslide-dammed lake with no sign of partial breaching, however, there is a high threat for catastrophic failure in the near future (days to months) based on the material of the dam and/or geomorphic parameters. Entirely/partially eroded, i.e. did not retain the landslide-dammed lake waters, but released it gradually in days to months. Breached (or partially) with evidence of catastrophic outburst flood(s) – includes internal and external causes for breaching (e.g. seepage erosion or erosion caused by overtopping)
Note: blue categories are dams where no remediation is needed; red categories advice hazard assessment and monitoring. 11
Journal Pre-proof 2.2
Geomorphometric landslide dam classification
The dams discussed in this paper are mainly formed by slides, falls, and flows according to the Varnes (1978) classification as modified by Hungr et al. (2014). A landslide dam, therefore, is the result of a landslide of any form whose runout and interaction with local topography allows its deposits to form a barrier that can hold back water above ‘normal’ river level. The shape and size of the landslide dam in relation to the size of the blocked valley is
oo
f
one of the most widely accepted morphological classifications of landslide dams and was proposed by Costa and Schuster (1988). It is based on 225 world-wide case
pr
studies (Table 4) and defines six classes. Hermanns et al. (2011b) modified it and
e-
added four further classes, also listed in Table 4; however, these refer to very specific
Pr
and rare cases at river confluences and drainage divides. Table 4 Geomorphometric landslide dam classifications.
Description
Type (Costa and Schuster, 1988)
Type (Hermanns et al., 2011b)
Small dams relative to the valley size that does not reach the opposite slope
I
II c
Large dams, but still small compared to the valley dimensions. This is one of the most widespread types , and they are frequently formed by rotational or translational landslides
II
II a
Very large dams that fill the valley from side to side. The collapsed materials are distributed both upstream and downstream from the release area, at times with a Tshaped deposit. Casagli and Ermini (1999) introduced a distinction in this class, separating the phenomena in which the dam also affects the tributary valleys of the main one (III a) from those in which these are not involved (III b).
III
IV a
Jo
ur n
al
Sketch
12
Journal Pre-proof
IV
II b
Several dams are formed by multiple tongues of a singular landslide. They are not very common for landslides but may form by glaciers (glacier tongues).
V
III b
Landslides with sub-channel rupture surfaces.
VI
II d
-
I
-
III a
Type of a landslide dam affecting water divides; with the landslide deposit depositing directly in the drainage divide.
-
Va
Single landslide dam having different forms distinguished between a rockslide crossing the valley partially or entirely, a rockslide involving the valley floor, and rockslides which, after impacting the valley bottom, spread in an up- and down valley direction (almost the same as III of Costa and Schuster).
-
II e
More complex multi-river geometries, focusing on the morphological relations of the dam in relation to the river valley
-
IV b-e, VII a-b, VIII, IX
pr
oo
f
Those very rare dams are formed from the contemporary movement of two landslides detached from opposite sides of the same valley.
Pr
e-
Supra-landslide lakes on landslide deposits are frequent on larger rockslide deposits , but can occur on all types of landslides. They are not connected to located within the channel of the dammed river.
Jo
ur n
al
Sub-type of plan view distribution of multiple landslide dams in a single valley in a line formed by a single landslide. Another type is III b.
-
13
Journal Pre-proof Supra-landslide lakes (Type-I of the Hermanns et al. (2011b) classification (Table 4)) are very frequent phenomena and their deposits often help to date prehistoric events. However, those water bodies are, with exceptions (i.e. Los Rojos supra-landslide lake— a large pond on top of a landslide deposit in North Patagonia, Hermanns et al., 2011a), comparatively small in size and rarely represent any threat to society. Hereafter, stability considerations for the small supra-landslide lakes are not considered.
Compositions and origins of river-damming landslides
oo
f
2.3
Dam geometry and composition strongly depend on the type of landslide forming the
pr
blockage. In the following, we briefly outline the types, materials and origins of
e-
landslides forming dams. Whereas all the river-damming landslides reviewed belong to
Pr
the slide and flow movement categories in the classifications by (Varnes, 1978) and Hungr et al. (2014), further attributes such as size and saturation with water are taken
al
into account herein since they play important roles in debris spreading and hence in the
ur n
final dam plan geometry, thickness, and erodibility. We used elements of the classifications of Cruden and Varnes (1996) and Hungr et al. (2014), but added volcanic
Jo
categories, because volcanic landslides and their unique materials (e.g. pyroclastics or wide variety of components and grain sizes) are important enough for the formation of landslide dams to include separate mention. Therefore, the following adapts classifications to include landslide size, degree of saturation, and origins.In the prevalent literature, the landslides forming dams cluster into: (1) large rockslides and rock avalanches (volumes >10 6 m3) composed of crushed rock with broad grain size distributions (Figure 1a), (2) volcanic mass movements (volcanic debris avalanches and pyroclastic flows; Figure 1b-c), (3) debris flows (saturated flows, often emplaced in pulses, typically thinner than slides/flows in sediment) (Figure 1d), (4) slides and flows 14
Journal Pre-proof in unconsolidated sediments (various sizes, unsaturated) (Figure 1e), (5) debris avalanches (unsaturated mixtures of (variably) rock, soil, sediment, vegetation), and (6) complex landslides, where one type of landslide transforms into another during
Jo
ur n
al
Pr
e-
pr
oo
f
emplacement.
Figure 1 (a) Aerial view of the Usoi landslide dam and Lake Sarez (photo: Mergili, M. (2019): The world in images. Digital media (https://www.mergili.at/worldimages)). (b) Sentinel 2 image (RGB composite image created with 843 bands, 10 m in resolution, from GloVis, USGS) showing the distribution of the debris avalanche deposit originated from the 1888 collapse of the Bandai volcano (Japan). The deposit blocked several branches of the Nagase River, forming three main lakes. (c) View from the southwest toward El 15
Journal Pre-proof Chichón Volcano and the Tuspac Valley. On 4 April, block-and-ash flows rushed along the Agua Tibia and Tuspac gullies, coalescing in the Magdalena River to form a dam 25– 30 m in maximum height. The lake breakout occurred on 26 May 1982, originating hot lahars up to 35 km of distance. (Picture courtesy of R. Mota taken on 24 April 1982). Notice the steaming pyroclastic flow deposits at the surface of the hot lake. (d) The dam formed by a debris flow at Brenot Creek in British Columbia. (e) Landslide dam at Wrigley Creek sediment type --- matrix-supported diamicton having sub-rounded to angular clasts at Wrigley Creek in Northwest Territories, Canada occurred probably after 1940 (Jermyn and Geertsema, 2015). Photos in (d) and (e) taken by Marten Geertsema.
2.3.1 Dams formed by large rockslides and rock avalanches
oo
f
Large, long-runout rock avalanches (RA) and rockslides (RS) are specific types of extremely rapid (m/s) landslides which arise from sudden failure of high rock slopes
pr
with failure volumes exceeding ~10 6 m3 (e.g. Davies and McSaveney, 2011; Davies and
e-
McSaveney, 2012; Erismann and Abele, 2013; Heim, 1932; Hewitt et al., 2008; Scheidegger, 1973). For rock avalanches below failure volumes of ~106 m3, the
Pr
characteristic long runout, flow-like motion, and intense fragmentation are commonly
al
absent. Both rock avalanches and rockslides are sourced in (relatively) intact bedrock
ur n
but differ in their spreading behaviour. The former break up into highly fragmented flowtype movements (“Sturzstrom”, Heim, 1932) and often spread radially (where unconfined) into sheets of dispersed surface features (ridges, lobes and hummocks),
Jo
whereas the latter slide, often in a roughly linear trajectory, as more coherent ‘blocks’ of highly fragmented debris (e.g. Erismann and Abele, 2013; Gruber et al., 2009; Hutchinson, 2006; Pollet et al., 2005). RA/RS dams have been reported with thicknesses up to 1,000 m (Hewitt, 2009b) and causing impoundments on the order of cubic kilometres (e.g. the 17 km3 Lake Sarez behind the Usoi dam; Alford and Schuster, 2000; Ischuk, 2011; Schuster and Alford, 2004) over tens of kilometres of valley length. The landslide dams formed by these events are often some of the most long-lived and the largest and are the most 16
Journal Pre-proof implicated in causing long-term disruptions to landscapes over a long time- and spacescales. 2.3.2 Dams formed by volcanic mass movements A comprehensive review on identification, stability and secondary effects of volcanic dams is given by Capra (2007). In their study, granulometric characteristics of formation and breach of volcanic dams were reviewed explicitly, and the stability of volcanic dams was analysed from a textural point of view. Volcanic eruptions, as well as lateral
oo
f
collapses, can produce gravity-driven granular flows, mobilizing large volumes (>106 m3) of debris. The most common events in causing river blockages are (1) volcanic
pr
debris avalanches (VDA) (Figure 1b), (2) pyroclastic density currents (e.g. Figure 1c),
e-
and (3) lahars.
Pr
Emplacement of VDAs from volcano lateral collapse commonly involves large volumes (>1 km3 ) and rapid emplacement (few minutes). A notable example is the 1980 Mt. St.
al
Helens eruption and lateral collapse with the emplacement of a 2.3 km3 VDA that
ur n
caused the formation of five main landslide dams along the North Fork Toutle River and raised the level of the Spirit Lake by 60 m (Meyer et al., 1986). These landslide dams
Jo
can be up to few hundred meters high (e.g. the 150-m high dam of Mt Spurr, Alaska, USA), impounding a volume of water in the order of few km 3 (e.g. 1 km3 lake volume at Volcán de Colima, México) (Table 5). Table 5 Examples of landslide dams in volcanic environments. Volcano
date
landslide type
18 May 1980
volcanic debris avalanche
Lake name Mt. St. Helens, USA Spirit Lake Elk Rock Lake Castle Lake Coldwater Lake Jackson Creek Lake Bandai, Japan
15 July 1888
volcanic debris 17
dam height [m]
lake volume [km 3]
time to failure
69 9 37 71 4.5
0.33 0.0003 0.024 0.83 0.0025
---unfailed--63 days ---unfailed-----unfailed--644 days
ref.
1
2
Journal Pre-proof
Iriga, Philippines Buhi Lake Nevado del Huila, Colombia Nevado de Colima, Mexico Volcán de Colima, Mexico
1628 AD 46,000-200,000 yrs BP 18,500 yrs BP 3,600 yrs BP
Parinacota, Chile
8,000 yrs BP
Hualca Hualca, Peru
Late Pleistocene
Mt. Bakening, Kamchatka Bezymyannoe Lake Verkhneavacha Lake Mt. Hood, USA Mt. Spurr, USA
8,000-8,500 yrs BP
1991 1982
lahar lahar lahar lahar and pyroclastic flows pyroclastic flow
---unfailed-----unfailed-----unfailed---
150
1.2
---unfailed---
3
30
0.1224
---unfailed---
4
60
0.5
> 3 days
5
80
1.0
> 1 year
6
60
0.4
> 11 days
7
40
0.4
---unfailed---
8
-
-
-
9
---unfailed-----unfailed--12 minutes
10
10 60 20
0.45 0.1-0.01
24
0.075
< 1 month
12
55
0.04
2 months
13
e-
Pinatubo, Philippine Mapanuepe Lake Chichón, Mexico
25 Dec. 1980 Late Holocene 1953-1992
volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche volcanic debris avalanche
0.45 0.6 0.165
f
10,000-38,000 yrs BP
Mt. Spurr, USA
30 30 30
oo
avalanche
pr
Hiab ara Lake Onogana Lake Akimoto Lake
11 3
al
Pr
Referenc es: 1) Glick en and Nak amura (1988) and Nak amura and Glick en (1988); 2) S wans on et al. (1982); 3) Waythomas and Nye (2001); 4) Lagmay et al. (2000); 5) Pulgarín et al. (2004); 6) Capra and Macıas (2002) ; 7) Cortés (2002) and Cortés et al. (2010); 8) Clavero et al. (2004); 9) Gómez-Avalos et al. (2004); 10) Melek estsev et al. (1999); 11) Gallino and Pierson (1985); 12) Umbal and Rodolfo (1996); 13) Macías et al. (2004).
ur n
Pyroclastic flows and lahars can also form landslide dams, but those are generally short-lived. All the known cases failed a few days to months after their formation,
Jo
possibly because of their smaller size and higher erodibility (Carey et al., 1996; Gallino and Pierson, 1985; Macías et al., 2004; Schuster and Crandell, 1984; Umbal and Rodolfo, 1996; Waythomas and Kaufman, 1991; Waythomas and Nye, 2001). 2.3.3 Dams formed by debris flows Hungr et al. (2001) define a debris flow as ‘very rapid to the extremely rapid flow of saturated non-plastic debris in a steep channel’. Debris flow volumes typically range from several m3 to ~107 m3, whereby lahars and combined rock avalanche -debris flows can reach volumes up to ~10 9 m3 (Fort, 1987; Fort and Peulvast, 1995; Griswold and 18
Journal Pre-proof Iverson, 2008; Iverson, 1997; Rickenmann, 2005). Debris flows often grow during runout through entrainment of water and sediment from the runout path (Frank et al., 2015). The rheology can vary substantially within a single debris flow event, with surgelike behaviour common in a single ‘event’, and the resulting deposits are poorly-sorted mixtures of grain sizes ranging from clay-sized to large boulders (Hungr et al., 2001; Iverson, 1997). The fluid nature promotes spreading and thinning once unconstrained. As a result, debris flow dams, where not topographically confined, form quite low dams
oo
f
that often fan out from tributary valleys into the main channel network. Debris-flow dams (Figure 1d) are often part of a geohazard cascade (Gallino and
pr
Pierson, 1985; Wei et al., 2018; Yin et al., 2016) or process chain, related to the need
e-
for a debris source that can be transformed into a slurry-like flow. Multiple debris flow dams can become chronic problems (Fan et al., 2019a) until the debris sources are
Pr
exhausted, with a clear correlation to the temporal patterns of rainfall that both trigger
al
debris flows and increase river flow to dammed reaches.
ur n
2.3.4 Dams formed by slides/flows in unconsolidated sediments Landslides which are neither dominated by granular rock texture (2.3.1) nor are fully
Jo
water-saturated (2.3.3), yet can be of large volume, are those formed in unconsolidated sediments (e.g. till; Figure 1e). Examples include the glacial and associated water-lain sediments, common in former glaciated terrains (Geertsema, 1998; R. L. Hermanns et al., 2012; Reginald Leonhard Hermanns et al., 2012; Locat et al., 2017) which often produce landslide dams (Clague and Evans, 1994; Geertsema et al., 2018; Geertsema and Clague, 2006; Geertsema et al., 2006a; Geertsema and Cruden, 2008; Geertsema et al., 2006b; Geertsema et al., 2009; Geertsema et al., 2006c; Lefebvre et al., 1991; Locat et al., 2017; Miller et al., 2018; Miller and Cruden, 2002). One of the most infamous of these is the 2014 Oso landslide in Washington State, USA, which ran over 19
Journal Pre-proof a rural residential area and killed more tha n 40 people (Iverson et al., 2015; Wartman et al., 2016). Rupture surfaces can daylight far above stream levels (Evans et al., 1996; Fletcher et al., 2002; Geertsema, 1998), but they may also occur near or below river channels (Cruden et al., 1993; 1997; Miller et al., 2018; Miller and Cruden, 2002). Those that are located below the channels will form Type VI dams (Costa and Schuster, 1988), which can be impervious. For these latter dams, the channel alluvium may be elevated almost
oo
f
vertically (Miller et al., 2018; Miller and Cruden, 2002). Upon overtopping of the Type VI dams, an incision into the sediment comprising the dams can be rapid; however, where
pr
the post-landslide stream encounters the pre-landslide alluvium, incision can be
e-
considerably delayed.
Pr
2.3.5 Debris avalanches and complex landslides Debris avalanches can occur anywhere on a slope and are not confined to an
al
established channel; which debris flows commonly are Hungr et al. (2014). They may
ur n
initiate directly on the slope or be triggered by the impact from a rockfall or rockslide if liquefiable material exists downslope from these other landslides (Hungr et al., 2014).
Jo
Or they may be a consequence of erosion and entrainment of soil and sediment into another landslide type. Where one landslide transforms into or triggers another during emplacement, the term ‘complex landslide’ (Cruden and Varnes (1996) applies. The latter include ‘rockslide-debris avalanches’, ‘rotational slide-earth flow’, ‘rock avalanchedebris flow’, and many more combinations.
20
Journal Pre-proof 2.4
Dam failure susceptibility
The material and sedimentological structures of the dams are not taken into account in geomorphometric
classification
schemes,
even
though
the
resulting
variable
geotechnical properties play an important role in dam stability (Casagli et al., 2003; Dunning et al., 2005). Therefore, Fan et al. (2017) proposed a three-fold classification scheme based on material and internal geological structure from an analysis of 32 landslide dams resulting from the 2008 Wenchuan earthquake (Table 6).
Jo
ur n
al
Pr
e-
pr
oo
f
Table 6 Landslide dam classification based on dam composition and internal structures (modified from Fan et al., 2017).
The classifications by Costa and Schuster (1988) and Fan et al. (2017) can be combined to obtain a preliminary indication of dam stability by considering the topographic setting and the sedimentological features of the valley-landslide system. 21
Journal Pre-proof Since dam types I to III of Costa and Schuster (1988) taken together constitute between 84% and 96% of known dams (Figure 2), most landslide dams can be assigned to
Pr
e-
pr
oo
f
these classes.
ur n
al
Figure 2 Landslide dam classification (Type I to VI) according to Costa and Schuster (1988) (% of landslide dams in each class) for previous inventories: moving from the outer ring to inner ring: (a) Worldwide - 225 cases Costa and Schuster (1988); (b) 19 cases by Ermini (2000); (c) Italy - 300 cases Tacconi Stefanelli et al. (2015); (d) Cordillera Blanca, Peru - 51 cases Tacconi Stefanelli et al. (2018).
Table 7 is an example of how, by combining two classifications methods , it could be
Jo
possible to provide a preliminary assessment of landslide-dam stability a very short time (days) after their formation that is largely independent of landslide type. For a rapid initial assessment of a dam’s failure probability, a three-class scale can be used: green for apparently stable, orange for stable in the short-term but unstable in the medium- to long-term, and red for potential instability in the short-term (immediate mitigation advised). Stability of landslide dams is also affected by many other factors, which cannot be directly considered in the initial, rapid assessment, 2-dimensional matrix solution in Table 7. Other factors such as catchment area, dam height and volume, lake volume, discharge, and erosive power can positively or negatively be assigned 22
Journal Pre-proof accordingly within such a stability matrix. For example, dams with small catchment areas/inflow, and large heights tend to be more stable; whereas dams with a large volume of impoundment, in a river of high discharge and high erosive power may be less stable and vice versa. Landslide triggering mechanisms are not considered in this assessment. Apparently stable dams below very active catchments or those situated in active
seismic
zones
(especially
during
an
aftershock
sequence
post-dam
emplacement) need to be considered more likely to breach. We emphasize that Table 7
oo
f
is for a rapid assessment of landslide dams within a large number of cases, but it cannot satisfy all conditions—at this stage, it is not a conclusive suggestion, rather a
pr
first attempt at the design of more comprehensive hazard levels.
Pr
e-
Table 7 Preliminary landslide dams stability classification matrix combining Costa and Schuster (1988) and Fan et al. (2017) methods. Green is for basically stable, orange for the stable in short time (but unstable in the medium- to long-term), red for unstable in short time. Fan et al. (2017)
I
Type
Sub-type B
Sub-type C
Small dam blocking valley composed by loose fine materials, formed by debris flows, avalanches, and slides with long long-runout
Small dam blocking valley formed by large boulders and blocks, originated by rock avalanches or rock falls
Small dam blocking valley composed by intact rock strata topped by coarse materials, formed by deepseated rock slides or avalanches
Large dam able to block the valley composed by loose fine materials, formed by debris flows, avalanches, and slides with long long-runout
Large dam able to block the valley formed by large boulders and blocks, originated by rock avalanches or rock falls
Large dam able to block the valley composed by intact rock strata topped by coarse materials, formed by deep-seated rock slides
Very large dam that fills the valley composed by loose fine materials, formed by debris flows, avalanches, and slides with long long-runout
Very large dam that fills the valley formed by large boulders and blocks, originated by rock avalanches or rock falls
Very large dam that fills the valley composed by intact rock strata topped by coarse materials, formed by deep-seated rock slides
al
Type
Sub-type A
ur n
Type
Jo
Schuster (1988)
II
Type III
23
Journal Pre-proof
3 Worldwide landslide dam distribution and database compilation 3.1
Problems of current inventories
Current datasets and inventories of landslide dams encompass dams formed by the full variety of landslides, i.e. from small, slow-moving landslides to water-saturated debris flows and to large (106 – 109 m3 ) catastrophic failures of entire rock or volcano slopes. Material types range from soil and debris through to variably broken rock, and can include any mixture of those materials plus vegetation and/or water /ice. The landslide
oo
f
deposits show the variability of their morphometric characteristics (e.g. plan area,
pr
height, width, volume), and differ in geotechnical parameters (e.g. permeability, grain size distributions, water content, armo uring, erodibility). Many of these parameters are
e-
difficult to obtain before the dam is cut by erosion or a breach. A comprehensive
Pr
database allows for sensible transferability of characteristics from those that have
al
breached, to those that have just formed--within error bounds. Landslide (dam) inventories are: (1) typically geomorphological inventory maps
ur n
containing little geotechnical information, (2) event-based, i.e. including all the slope movement types that occurred after a particular triggering event (e.g. earthquake,
Jo
rainstorm), or (3) historical and regional, consisting of all landslides (dams) in a specific area going back in time tens to even tens of thousands of years (Malamud et al., 2004). The available datasets furthermore differ in two critical aspects: (1) quality/resolution of data sampling and (2) age of the events (Tacconi Stefanelli et al., 2015). The morphometric information for each inventory is generally extracted from a wide range of remote sensing data (satellite imagery, aerial photos, and digital terrain models), possibly with different spatial resolution and with limited ground-truthing. This often leads to measurement errors of morphometric parameters , and comparisons between 24
Journal Pre-proof datasets (and even individual events within a given dataset) need to consider this level of inaccuracy. Many cases in the historical inventories are not ‘recent’ events and, generally, the farther the case lies in the past, the less reliable the data is due to deposit modifications over time (erosion, vegetation, consecutive landslides, and anthropogenic modifications; even more so for smaller sizes). However, the estimate of ‘stability’ is likely better when based on long-term evidence than for rapidly collected event inventories. The direct consequence of this is a general underestimation of the
oo
f
morphometric parameters of the landslide dam body (i.e. height, length, width and volume), difficulties in landslide-dam reconstruction for any back-analyses of stability,
pr
or, biased preservation of the largest events that resist long-term removal (Sanhueza-
e-
Pino et al., 2011). Therefore, we take the first attempt to compile the existing
(see supplementary material).
Worldwide landslide dam database
al
3.2
Pr
inventories and datasets and create templates and standards for further data collections
ur n
We use the term ‘inventory’ for regional, world-wide or event-based catalogues; ‘dataset’ for smaller collections, and ‘database’ for the catalogue compiled within this
Jo
research since the goal of our data collection efforts is to make the resulting database accessible, searchable and open to new records in the future. Table 8 contains the published landslide dam inventories from which our database was compiled. They are historical, geological and event-based inventories. Those in bold are the most comprehensive in terms of the required key information, rather than simply number of events, for each geographical region (see Section 4.3 and supplementary material for the required key information). Table 8 Summary of reviewed landslide dam inventories and datasets. In bold letters, the most important/recent ones. Blue shaded rows are the reviewed datasets from which 25
Journal Pre-proof individual records were taken to the new database. In brackets of the blue shaded rows are the numbers of landslide dams selected for the new database compilation. Region, Country
N. of case s
Information on the inventory
Historical or event-based
Source or reference
1
world
187
Conclude the failure and lasting time.
Historical
Schuster (1986)
2
world
13
Most of these dams have been caused by relatively high-velocity debris flows originating from tributary valleys to briefly block rivers in main valleys. Generally, dams formed in this manner are neither high, nor are they particularly resistant to erosion.
Historical
Costa and Schuster (1988)
3
world
463
Including location, triggering mechanism, landslide type , size, failure time and mechanism, etc.
Historical
Costa and Schuster (1991)
Historical
[~100]
f
No.
world
18
Using a model to predict peak discharge of floods caused by the failure of natural landslide dams and artificial dams.
5
world
350
Most of the collected cases are from the western USA, Japan, and Italy. A common feature is a high frequency of damming events along or near the active margin areas of tectonic plates .
Historical
Ermini and Casagli (2003)
6
world
18
Between years 1840-2010, dams having impoundment volume exceeding 100 million m 3.
Historical
Evans et al. (2011a)
7
world
1239 [290]
Based on records of 52 landslide dam cases with breaching information in the database, empirical models for estimating the breaching parameters of landslide dams are developed.
Historical / Eventbased
Peng and Zhang (2012)
8
Argentina (Patagonia)
41
Overview of prehistoric rockslide dams, with geomorphic parameters extracted from GIS analysis.
Geological
Hermanns et al. (2011a)
9
Argentina (Central Andes)
20
Overview of prehistoric rockslide dams, with geomorphic parameters extracted from GIS analysis.
Geological
Hermanns et al. (2011b)
10
Canada
38
16 existing and 22 landslide dams.
historical
Historical
Clague and Evans (1994)
11
Canada
8
Geomorphic parameters and stability indices for recent landslide dams in glaciolacustrine sediments.
Historical
Miller et al. (2018)
12
Central Asia
190
Include location, state of the dammed lake (existing, silted, breached), data on the size and type of the 190 river-damming landslides exceeding 1 million m3 in volume for which parameters of the dammed lakes were indicated. The actual number of past rockslide dams is much larger.
Prehistoric mainly and several historical events
Strom and Abdrakhmatov (2018)
13
China
147
Including name, location, date,
Historical
Chai et al. (1995)
Jo
ur n
al
Pr
e-
pr
oo
4
26
Walder and O'Connor (1997)
Journal Pre-proof landslide volume, blocked river, last period, triggering factors and impacts. China
33
All are caused by the Wenchuan earthquake. Risks of dam collapsing and estimated collapse mechanism are given in the table.
Event-based
Yin et al. (2009)
15
China
257
All triggered by M8.0 Wenchuan earthquake.
Event-based
Cui et al. (2009)
16
China
256
Because of limited access, relatively detailed data are available for only 32. The paper presents statistical analyses of the distribution, classification, characteristics, and hazard evaluation of these 32 dams.
Event-based
Xu et al. (2009)
17
China
828 [20]
All in the Longmenshan mountains, China, triggered by the M w 7.9 2008 Wenchuan earthquake. It is the largest of its kind attributable to a single regional-scale triggering event.
Eventbased
Fan et al. (2012c)
18
Karakoram Himalaya
322
In surve ys between 1993 and 2009, in the Karakoram and Hindu Raj Mountains, a total of 322 Hewitt (2004); Hewitt (2006); rock avalanches were identified; this total has been increased to more than 370 in recent work (Hewitt (2011). 65% dammed rivers of the Upper Indus (all are rock avalanches).
Geological
19
Italy, Apennine
70
Several geomorphic parameters have been measured, during field investigations and through air photo interpretation, or estimated, using historical and bibliographic information.
Event-based
Casagli and Ermini (1999)
20
Italy, Northern Apennines
68
Including grain size distribution of dammed lakes only.
Historical
Casagli et al. (2003)
21
Italy
300 [49]
Several geomorphic parameters have been measured, during field investigations and through air photo interpretation, or estimated, using historical and bibliographic information.
Historical
Tacconi Stefanelli et al. (2015)
22
Japan
43
This study utilizes 43 landslide dams to determine the significant variables, including logtransformed peak flow (or catchment area), and logtransformed dam height, width, and length in hierarchical order, which affect the stability of a landslide dam.
Historical
Dong et al. (2009)
23
New Zealand
21
Eleven small lakes were formed by landslides caused by the 1929 Buller earthquake; four others were formed by other historical earthquakes in New Zealand. At least nine other New Zealand lakes are also dammed by landslides and were probably formed by prehistoric earthquakes.
Historical / Event-based
Adams (1981)
Hewitt (2011)
Jo
ur n
al
Pr
e-
pr
oo
f
14
27
Journal Pre-proof New Zealand
232
Geomorphometric data on landslide dams, associated lakes, and contributing catchment characteristics
Historical
Korup (2004)
25
New Zealand
39
Geomorphic parameters, landslide types and features of very large landslides in Fiordland. Only landslides with volumes greater 6 3 than 10 m are listed. Comparison with 11 other giant landslides from around the world.
Historical / Event-based
Hancox and Perrin (2009)
26
New Zealand
240
More than half of all recorded landslide dams in New Zealand were formed by large rock-slope failures, with most of them occurring in the Southern Alps and adjacent Fiordland. Almost 30% of all recorded landslide dams in New Zealand were formed by rock avalanches
Historical
Korup (2011)
27
Peru
51 [7]
Geomorphic parameters have been measured, during field investigations and through air photo interpretation.
28
Venezuela
35
Overview of historic and prehistoric rockslide dams in the Andes of Venezuela.
29
Switzerland
35
List the natural lakes in Switzerland, which cover an area above approximately 0.5 km 2.
30
Taiwan, China
17
31
Turkey
32
33
oo
f
24
pr
Historical
Tacconi Stefanelli et al. (2018)
Ferrer (1999)
Historical
Evans et al. (2011a)
This paper gives the risk level, present condition, and its authority.
Eventbased
Chen et al. (2011)
3
Only one remains. All are induced by a rainfall event.
Eventbased
Duman (2009)
Nepal
*
The 25 April 2015 M w 7.8 Gorkha earthquake triggered many shallow failures with only a few landslide dams that were either very short-lived or only partially blocked the river course.
Eventbased
Dhital et al. (2016)
New Zealand
#
196 significant drainage blocking or drainage constricting landslides were identified after the 14 September 2017 M7.8 Kaikoura earthquake.
Eventbased
Massey et al. (2018)
ur n
Few
al
Pr
e-
Geological
Jo
196 [4]
Note: * Inventories reporting the landslide dams are not available. #Only 4 large landslide dams with > 106 m3 landslide volume were reported.
The inventories in Table 8 total >1,800 entries. In addition, another dataset from Strom and Abdrakhmatov (2018) has 1,000 pre-historic rockslides from the Central Asian region 549 of which had formed dams, though parameters were measured for 190 cases only. As the numbers of known landslide dams increases, we have found uncertainty enlarges and statistical representativeness of the dataset diminishes 28
Journal Pre-proof because of missing information. Therefore, to minimize uncertainties, our analysis focuses only on a sub-set taken from the entire database (>1 ,800 dams), comprising 410 large (i.e. more than 106 m3 in landslide volume) and recent (i.e. time after 1900) landslide dams with detailed information. From hazard perspectives, more attention needs to be paid to these recent, large dams. Evans et al. (2011a) also considered rockslide dams with volume in excess of 106 m3 . These can be potentially dangerous to the downstream population considering the potentially large volume of water present in
oo
f
the impoundment. The distribution of the existing landslide dam inventories and the 410
Jo
ur n
al
Pr
e-
pr
landslide dams is shown in Figure 3.
29
Journal Pre-proof
f o
l a n
o r p
e
r P
r u o
J
Figure 3 Worldwide distributions of existing landslide inventories and recent large landslide dams in different regions. The yellow stars are landslide dam locations; the circles show the order sourced from Table 7, and the numbers inside brackets are the total number of landslide dams. The background is modified from Amante and Eakins (2009) . 30
Journal Pre-proof The distribution in Figure 3 seems to suggest that most landslide dams are located the ice-free areas of Asia, and many are in North America, Europe, Australia, and South America, and that next to none have been reported on the African continent. However, strong biases are introduced by (a) researchers’ regional preferences and citizenship, (b) single events with many landslides (such as the 2008 Wenchuan earthquake) skewing the distribution, and (c) our bias towards large landslide dams. However, geologically this distribution makes sense . In Asia, affected by the plate collision
oo
f
between the Indian Ocean and the Eurasian Plates, the Himalayas and surrounding areas suffer from large earthquakes due to intense tectonic movements (Blöthe et al.,
pr
2015; Gong et al., 2018; Kirby et al., 2000; Kirby et al., 2003). Earthquake-induced
e-
landslide dams are distributed along collision areas and subduction zones, such as the
and Nepal; Japan; Central Asia.
Pr
Tibetan Plateau bordering Western China; Himalayan ranges bordering India, Pakistan
al
Taking China as an example, the historical landslide dams are predominantly found in
ur n
Sichuan Province, where the destructive 2008 Wenchuan earthquake occurred (Shen et al., 2009; Yi et al., 2008; Yin et al., 2009). Landslide dams are widely distributed in
Jo
Japan and are well documented by Swanson et al. (1986). Most of them were triggered by earthquakes and rainfall, while some cases triggered by snowmelt are also reported (Costa and Schuster, 1991; Dong et al., 2009; Koi et al., 2008; Tabata et al., 2002; Wang et al., 2019). In North America, landslide dams are mainly concentrated in the west of Canada and the United States, close to the Pacific Ocean (Clague and Evans, 1994). With the influence of marine climate, rainfall is the most prominent triggering factor here. In Europe, landslide dams are abundant in western Italy, Switzerland, and Norway. The European Alps, Norway, and the Apennines in Italy are characterized by wide geo31
Journal Pre-proof tectonic and climatic variability where landslide dams and related flooding are frequent (Bonnard, 2006; Casagli and Ermini, 1999; Guzzetti and Tonelli, 2004; Salvati et al., 2010; Tacconi Stefanelli et al., 2015). In New Zealand, located in a deforming plate boundary zone (Adams, 1980; Kingma, 1959) where steady tectonic deformation is occurring (King, 2000), many historical landslide dams are oft-assumed triggered by earthquakes. For around 59 % of the 232 landslide dams examined by Korup (2004), the trigger mechanism remains unexplained, whereas 29 % are believed to have
oo
f
formed by earthquakes and another 11 % are tentatively believed co-seismic. The very strong rainfall typical for the Southern Alps of New Zealand could account for some of
pr
the unknown triggers (Henderson and Thompson, 1999). In South America, influenced
e-
by tectonics, volcanism, and climate change (Carey et al., 2012; Schauwecker et al., 2014; Worni et al., 2014), landslide dams are prominent along the Andes in Peru,
Pr
Argentina, and Venezuela (Ferrer, 1999; Hermanns et al., 2011a; Tacconi Stefanelli et
al
al., 2018). The Andes range, a part of the continental divide, is tectonically active
plates.
Recommendations for data collection
Jo
3.3
ur n
through the collision of the Nazca, South American, and a part of Antarctic lithospheric
To ensure databases are as complete as possible, it is important to standardize how key variables are collected for each landslide dam. Taking how dam height is measured as a simple example: dam surfaces are typically highly irregular, and it is not always clear whether the maximum or the minimum (where water may flow) elevation is reported in the literature. Ideally, all the information should be recorded (for example see Table 2 of Korup (2002), but in practice, few papers record all, either due to unavailable data or a lack of a well-established recording protocol. We recommend:
32
Journal Pre-proof (1)
Creating the inventory using geographic information system (GIS), which can show the accurate location linked with an attribute table to compile all the dam parameters, and it can also facilitate further geospatial and geomorphological analysis;
(2)
Creating metadata in order to ensure correct and proper use and interpretation of the inventory by its users; Using standard terminologies and providing clear definitions of the terms used;
(4)
Creating an Open Access portal in which collated records can be updated as
oo
f
(3)
new data are available, and wherein new events can be verified and added.
pr
We also create a data collection template (including over 60 diverse geological,
e-
geomorphological, geotechnical, and climate factors) for anyone to use and upload the
Pr
new collection of landslide dam inventories in the planned open-access portal. The template is updated after the versions of Casagli and Ermini (1999), Korup (2002) and
al
Tacconi Stefanelli et al. (2015), and is available as supplementary information. We
ur n
further provide an example of how to use the template by filling in information of a recent landslide dam event in eastern Tibet, China (Fan et al., 2019b) (see
Jo
supplementary information).
4 Formation and Stability 4.1
Controlling factors
In Figure 4, we compare the statistics of the triggers and types of river-damming landslides, movement and material, and the longevity of landslide dams based on the inventory from Costa and Schuster (1991) and the new landslide dam database from this study. About 100 cases from Costa and Schuster (1991) are included in the new database. The intention of this comparison is to visualise whether different sampling 33
Journal Pre-proof strategies with thresholds of size and age (as used in the new database) influence the results. Rainfall was the trigger for 42% and 44% of landslide dams, and the earthquake was the second major trigger for 29% and 20%, respectively in Costa and Schuster (1991) and in the new database (Figure 4a). Even though only large landslide dams are included in the datasets, the overall statistics do not differ much from Costa and Schuster (1991). The categories listed here are based on the classifications provided in each of the
oo
f
inventories used to compile our new database. As discussed in sections 2.3 and 3.3, terminologies are not always clear and classifications sometimes exclude specific
pr
phenomena or they were updated after an inventory was published. For example,
e-
‘complex’ is listed as a movement type in Varnes (1978) and as a modifier (e.g.
Pr
complex rock slide – debris flow) in Cruden and Varnes (1996), but is excluded in Hungr et al. (2014). Careful checking of each entry in the new database and updating
al
the classification schemes for landslides forming dams are two of the overarching
Jo
ur n
research goals of our ongoing database compilation and research.
34
Journal Pre-proof Figure 4 (a) Triggering mechanisms and (b) landslide, movement and material types of landslide dams compared between Costa and Schuster (1991) and our new database
In the inventories and datasets shown in Figure 5, of those landslide dams that did fail, about 80% did so within the first year after formation (Costa and Schuster, 1991; Fan et al., 2012c; Peng and Zhang, 2012). If a landslide dam survives the first year, it is statistically likely to survive for several decades or even centuries as it has found internal stability until some external influence, such as earthquakes, modifies that equilibrium. However, it is always unknown how many will remain stable and which
oo
f
ones will experience partial or complete breaching despite being classified as stable at the time of inclusion in any inventory. In most of the (historical) cases, the failure cause
pr
is unknown, and it is not possible to differentiate external (mainly earthquakes) or
e-
internal influences. The Kummersee rockslide dam in the Italian Alps that had blocked
Pr
the valley up to 150 m high is an example of multiple breaching episodes. It was formed in 1404 and experienced six partial failures until its final, catastrophic breach in 1774
al
(Ermini and Casagli, 2003). Other examples are the 1914 breach of the rockslide dam
ur n
in the Rio Barrancas Valley in Argentina that had existed at least 425 years before breaching (Hermanns et al., 2011a), the 1963 breach of the Issyk rockslide dam near
Jo
Almaty City in Kazakhstan (failure triggered by a debris flow that was produced by the breach of the Jarsay glacial lake upstream), and the 1966 breach of the Yashinkul rockslide dam in Kyrgyzstan (Strom and Abdrakhmatov, 2018). Both dams in Central Asia had existed for centuries, maybe even millennia before their catastrophic failure.
35
pr
oo
f
Journal Pre-proof
e-
Figure 5 Age of landslide dams at failure from different inventories; data modified from Fan et al. (2012c).
Pr
Understandably, these concerns drive much of the research, one of the major objectives of which is to understand the stability of the dam, to estimate when or if it is
al
likely to fail catastrophically. However, although the known population of dams can be
ur n
separated into those that have failed and those that have not, those in the latter category - may still fail in the future (e.g. Nash et al., 2008). Thus, while some large
Jo
dams seem unlikely ever to fail, based on either empirical or other investigations, the possibility of a catastrophe cannot be ignored completely. Most of the reviewed inventories and datasets (Table 7) are heterogeneous, as are the methods and biases of different scientific disciplines when it comes to identifying the factors controlling landslide dam stability. Key factors usually recorded include dam dimension, composition, and internal structures (post-erosion or breach), the river hydraulic gradient, inflow rate, and, as a proxy for inflow, the size of the upstream catchment area (e.g. Capra, 2011; Casagli and Ermini, 1999; Casagli et al., 2003; 36
Journal Pre-proof Dunning and Armitage, 2011; Dunning et al., 2007; Strom and Abdrakhmatov, 2018). Some data are readily available in the field or from remote sensing images, such as the dam dimensions, whereas other geotechnical parameters of the dam material and its internal structures, which control the stability, e.g. dam permeability, are rarely available (Casagli et al., 2003; Fan, 2013; Wang et al., 2016). As a result, some parameters affecting stability can be directly assessed from the landslide dam, whereas others must be inferred based on similarity to other landslides or landslide dams.
oo
f
In Table 9, we outline the characteristics that past works have deemed most important for landslide dam stability in each of their inventories. The apparent differences in
pr
emphasis highlight the need for continued, focused research, detailed observations,
e-
and comprehensive data collection to enable statistically significant analyses that go
Pr
beyond individual or regional case studies (cf. 3.1).
Table 9 Key stability-controlling variables used to infer (in) stability.
Worldwide and regional inventories of various landslide types
ur n
Landslide dam volume
Notes on the landslides that the referenced studies are based on
al
Factors identified as key in dam stability
Catchment; taken as proportional to river discharge
Landslides in unconsolidated sediments
References Dal Sasso et al. (2014); Ermini and Casagli (2003); Korup (2004); Tacconi Stefanelli et al. (2016) Geertsema and Clague (2008); Geertsema et al. (2006a); Miller et al. (2018) Capra (2006); Capra (2007); Capra (2011); Davies and McSaveney (2011)
Geometry, which influences the hydrologic gradient and location of the highest landslide dam point
Long-runout rockslides and rock avalanches
Korchevskiy et al. (2011); Strom (2006); Strom and Abdrakhmatov (2018)
Jo
Sedimentology/texture
Volcanic debris avalanches (in comparison to lahars and pyroclastic flows); long-runout rockslides and rock avalanches
4.2
Geomorphic indices
Geomorphic indices have been developed by various researchers to quantitatively assess the formation probability of a dam and predict its state once formed by employing morphometric parameters (Figure 6); most of which are easy and rapid to 37
Journal Pre-proof collect. The stability thresholds are then determined by analysing the geomorphic parameters in a bi- or multivariate way and graphically delineating domains in which ‘stable’ or ‘unstable’ (i.e. with existing or silted-up lakes or without lakes) dams exist
Jo
ur n
al
Pr
e-
pr
oo
f
exclusively (often with a large area of uncertainty between the two domains).
Figure 6 Cartoon illustrating the various geometric parameters of a dam and catchment
4.2.1 Dam formation indices Whether a landslide can form a complete dam when it reaches a river valley depends on several factors, including landslide velocity and deposit parameters (volume, 38
Journal Pre-proof thickness, and shape; i.e. resulting dam geometry), topography (valley width and shape), and hydrology (stream discharge and transport capacity) (Fan et al., 2014a; Fan et al., 2014b). Four indices are published that investigate dam formation probability (Table 10). The Annual Constriction Ratio (ACR) relates the width of the valley (Wv) with landslide velocity (v, here in m/a): ACR = v/Wv and was initially intended to estimate channel response (e.g. erosion of constricting material, channel diversion, lake
oo
f
formation) to delivery of material by slope processes (Swanson et al., 1986). Applied to an inventory of different landslide types from slow (1.5 m/a) to extremely rapid ones (>
pr
3m/s) in Japan, an ACR > 100 was assumed to be required to cause a lake to form
e-
(Swanson et al., 1986). Using an inventory of landslides in Italy, Tacconi Stefanelli et al.
Pr
(2016) redefined the annual constriction ratio into ACR = log (W v/v), with v expressed in m/s this time. They found a threshold of 4.26 below which all landslides formed dams.
al
The index took into account landslide velocity as an empirically evaluated parameter
ur n
(Cruden and Varnes, 1996) that is not always possible to measure, but for which calculated velocities based on runup heights (Chow, 1959; Jibson et al., 2006) and
Jo
seismic signal analyses delineate a range of values (Coe et al., 2016; Dufresne et al., 2019; Guthrie et al., 2012; Jiskoot, 2011). The Dimensionless Morpho-Invasion Index (DMI; Dal Sasso et al., 2014) is a derivative and more complex formulation of the ACR, relating landslide velocity (v), the density of the landslide material (ρL) and the landslide volume (VL) to the valley width (WV), the density of water (ρw), the gravitational acceleration (g), the hydraulic level (h), which is the elevation (m) of water surface from the bottom of river bed, and the landslide width (WL) in the relationship DMI = (2·ρL·v2 ·VL)/(ρw·g·h2·Wv·WL).
39
Journal Pre-proof Ermini (2003a) formulated the Dimensionless Constriction Index (DCI) to include the erodibility of the dam material in the stability estimates, taking into account landslide velocity (v), landslide width and height (WL and HL), cumulative grain size distribution (D30), the water discharge with a return period of 5 years (QP), and the width of the valley (WV) in DCI = (v·WL·HL·D 30) / (QP·WV). Correlating landslide volume (VL) with the width of the valley (W V), Tacconi Stefanelli et al. (2016) furthermore designed the Morphological Obstruction Index, MOI = log
oo
f
(VL/WV).
Dimensionless MorphoInvasion Index Dimensionless Constriction Index
> 100
log (WV/v)
< 4.26
(2·ρL·v2 ·VL) /
(QP·WV)
Morphological Obstruction
<1
> 0.002
log (VL/WV)
Region Japan Italy
Reference Swanson et al. (1986) Tacconi Stefanelli et al. (2016)
Italy
Dal Sasso et al. (2014)
Italy
Ermini (2003a)
<3
Italy
< 3.08
Peru
> 4.6
Jo
Index
> 6.88
>1
(ρw·g·h 2·Wv·WL)
(v·WL·HL·D30)/
e-
Ratio
v/ WV
No dam forms
Pr
Constriction
Dam forms
al
Annual
Equation
ur n
Index
pr
Table 10 Indices formulated to assess the probability of a landslide to block a river and form a dam.
Tacconi Stefanelli et al. (2016) Tacconi Stefanelli et al. (2018)
4.2.2 Dam stability indices Once a dam has formed, its stability depends on morphometric (size and shape of the dam and valley system), hydrogeological (pertaining to the river, the catchment and impoundment areas, and to climatic systems), and engineering geological (i.e. dam geotechnical properties) factors. The parameters used in the indices all interlink, and their relative weighting and significance regarding their ease of measure requires further investigation. 40
Journal Pre-proof With landslide dam volume (VD) in the numerator, the Blockage Index, BI = log(VD/AC) takes into account the catchment area, AC (Canuti et al., 1998; Casagli and Ermini, 1999), whilst the Impoundment Index (Casagli and Ermini, 1999) considers the volume of the impoundment, i.e. the landslide-dammed lake (VL), in II = log (VD/VL). Introducing dam height (HD) to the Blockage Index, Ermini and Casagli (2003) designed the Dimensionless Blockage Index, DBI = log (AC*H D/VD). Based on dam height (HD), Korup (2004) defined the Backstow Index IS = log (HD3/VL)
oo
f
and the Basin Index IA = log (H D2/AC). Note that in IS both parameters are in m3 , whereas in IA the dam height is in m and the catchment area in km2 . In the Relief Index
pr
IR = log (HD/HR) (IR; Korup, 2004), the dam height is correlated with the relief upstream
e-
of blockage, HR = Emax – Emin – HD; where Emax and Emin are the highest and lowest
Pr
altitudes in the catchment respectively, where the lowest altitude is corrected for landslide dam height. In IR (HD and HR), the units are meters.
al
Tacconi Stefanelli et al. (2016) proposed to consider dam destabilization by the river
ur n
using a simplified expression of stream power (AC*S) per unit channel length (where S is the local longitudinal slope of the channel bed), in their Hydromorphological Dam
Jo
Stability Index, HDSI = log (VD/AC*S). Table 11 categorizes all the above-mentioned geomorphic stability indices and their thresholds for apparent stability. Table 11 Indices drafted to categorize stable and unstable dams. In bold are the indices used for the analysis shown in Fig. 7. Index
Blockage Index (BI)
Equation
Stable
Uncertain
Unstable
Region
>5
4-5
3-4
Italy
>5
4-5
<4
Italy
log (VD/AC)
41
Reference Canuti et al. (1998) Casagli and
Journal Pre-proof Ermini (1999)
>0
Na
<0
>1
Na
<1
< 2.75
2.75-3.08
> 3.08
<3
3-5
>5
< 2.43
2.43-3.98
> 3.98
log (H D3/VL)
>0
0-(-3)
Basin Index (BI)
log (H D2/AC)
>3
Na
Relief Index (RI)
log (H D/HR)
> -1
Impoundment Index (II)
Italy
Casagli and Ermini (1999)
log (VD/VL) New Zealand
world-wide
Korup (2004)
Ermini and Casagli (2003)
Index (HDSI)
log (VD/AC*S)
ur n
Dam Stability
<3
New Zealand
Italy
New Zealand New Zealand New
Na
< -1
5.74-7.44
< 5.74
Italy
5.26-8.07
< 5.26
Peru
Zealand
al
> 7.44 Hydromorphic
oo
Index (IS)
< -3
pr
Backstow
e-
Index (DBI)
log (AC*HD/VD)
Pr
Blockage
f
Dimensionless
> 8.07
Korup (2004)
Tacconi Stefanelli et al. (2016)
Korup (2004)
Korup (2004)
Korup (2004)
Tacconi Stefanelli et al. (2016) Tacconi Stefanelli et al. (2018)
Jo
Owing to the empirical nature of HDSI, the thresholds vary with the dataset used and depend on regional differences (as demonstrated in studies in Italy and Peru by Tacconi Stefanelli et al. (2016); Tacconi Stefanelli et al. (2018)), and perhaps are also influenced by data acquisition methods. The thresholds will continue to be modified as more data becomes available. In order to evaluate a specific dam’s stability, however, these thresholds should be narrowed down for each distinct landslide type. For example, Miller et al. (2018) and Dufresne et al. (2018) tested the indices on a set of landslides in glaciolacustrine sediments (Peace River region, Canada), and large rock avalanches and rockslides (Oetz Valley, Austria), respectively. The calculations of dam 42
Journal Pre-proof stability remained largely inconclusive in both cases. However, the two studies are based on very limited cases and need to be reproduced with larger datasets (such as the one compiled herein) since some dams fall into “wrong” domains. The same issues of transferability were noted for dams formed by large-scale rockslides in Central Asia (Strom, 2010). As a first step, we chose a set of four very different landslides but of comparable volume, and in a similar geomorphic setting to contrast the different stability indices.
oo
f
The (1) Tangjiashan landslide dam (Sichuan, China) was formed by a rockslide triggered by the Mw 7.9 2008 Wenchuan earthquake. An artificial spillway was excavated
pr
to mitigate the dam-breach flood (Fan et al., 2012a). The (2) Signatico landslide (Italy)
e-
is complex debris and earth flow that was reactivated more than once over time. In
Pr
1945, a large debris flow formed a landslide-dammed lake in Parma Creek, which survived for several years (Tacconi Stefanelli et al., 2015). (3) Between 28 March and 4
al
April 1982, El Chichón erupted, producing pyroclastic deposits which dammed the
ur n
Magdalena River (Mexico). On 26 May, the landslide dam was breached releasing sediments and hot water downstream (Macías et al., 2004). (4) On 22 March 2014,
Jo
after a long period of heavy rainfall, the Oso landslide involving glacial and lacustrine sediments dammed the North Fork of the Stillaguamish River (Washington, USA) (Iverson et al., 2015).
All these four landslide dams have volumes of about 107 m3 and, apart from the Oso landslide, have had a landslide-dammed lake some km in length. Almost all the indices correctly classified the formation of the four landslide dams, with the exception of the Annual Constriction Ratio (with the formulation of Swanson et al., 1986) and, for the Signatico dam, the Dimensionless Morpho-Invasion Index (Table 12). In the stability calculations, nearly all the cases were classified as unstable except for the 43
Journal Pre-proof Dimensionless Blockage Index, which classified the Tangjiashan landslide as stable (Table 13). However, the Tangjiashan landslide dam was artificially breached, and its stability without human intervention remains unknown, though qualified to be potentially hazardous considering dam height, dam composition and a maximum capacity of the impoundment (Xu et al., 2009). Table 12 Index results for the prediction of the dam formation for the Tangjiashan, Signatico, El Chichón, and Oso landslides.
Constriction Ratio
1.68
Morphological Obstruction
Formed Formed
5.78
Formed Formed
4.85
2.38E-02 1.62
Formed
Formed
1.96E-06
12.80
Formed
n.a.
2.00E-
Not
Formed
02
Formed
Formed
1.70
Formed
2
Uncertain
2
Uncertain
3
Formed
Not
Formed
1
3.92 Formed
59.91
Formed
208.30
Formed
4
n.a.
-
n.a.
-
5
Pr
81.07
Ref
Not
4.62
5.30 Formed
Dimensionless
Invasion Index
-
al
Dimensionless Constriction
Formed
Formed
Index
Morpho-
Not
1.67E-06
Oso
f
Not
El Chichón
oo
2.08E-02
Signatico
pr
Annual
Tangjiashan
e-
Index
Index
et al. 2014; 5: Ermini 2003
ur n
References, 1: Swanson et al. 1986; 2: Tacconi Stefanelli et al. 2016; 3: Tacconi Stefanelli et al. 2018; 4: Dal Sasso
Index
Jo
Table 13 Index results for the prediction of the stability of the dam for the Tangjiashan, Signatico, El Chichón, and Oso landslides. Tangjiashan
Signatico
Unstable Blockage Index
Impoundment Index
Dimensionless Blockage Index
0.53
El Chichón
Unstable -1.33
Oso
Unstable -1.14
Ref. Unstable
1
-2.32
Unstable
Unstable
Unstable
Unstable
2
Unstable
Unstable
Unstable
-
2
-1.29
0.06
-0.81
n.a.
Unstable
Unstable
Unstable
-
3
Stable
Uncertain
Uncertain
Unstable
4
Uncertain
3
1.39
2.81 Stable
2.84 Stable 44
3.62 Stable
Journal Pre-proof Stable
Uncertain
Uncertain
Uncertain
5
Backstow Index
-2.62
Unstable
-2.47
Uncertain
-2.51
Uncertain
n.a.
-
3
Basin Index
-2.72
Unstable
-5.22
Unstable
-4.53
Unstable
-6.48
Unstable
3
Relief Index
n.a.
-
n.a.
-
n.a.
-
n.a.
-
3
Unstable
5
Unstable
6
Hydromorphic Dam Stability Index
Unstable 2.44
Unstable 1.54
Unstable
Unstable 0.90
Unstable
0.17 Unstable
oo
f
References, 1: Canuti et al. (1998); 2: Casagli and Ermini (1999); 3: Korup (2004); 4: Ermini and Casagli (2003); 5: Tacconi Stefanelli et al. (2016); 6: Tacconi Stefanelli et al. (2018).
In addition to the four cases above, the Impoundment Index, Blockage Index, Backstow
pr
Index, and the Dimensionless Blockage Index were calculated for cases with sufficient
e-
details from our new database (Figure 7). While none of the indices clearly distinguish stable and unstable dams precise enough to make well-informed quantitative risk
Pr
decisions, the overall performance of the DBI is found better at identifying some of the
al
stable dams which lasted between one and >20 years. However, the DBI also identifies
ur n
some landslide dams that lasted one-day-to-one-month and one-month-to-one-year as stable. The Impoundment Index identifies some cases as stable. The other two indices
Jo
mostly identify the landslide dams as either unstable or uncertain. Info rmation to calculate other indices, i.e. Relief Index and Hydromorphic Dam Stability Index are not available from the database, a common issue which hampers evaluation. The basis on which these indices were developed does not include the time a dam has remained stable. The Impoundment Index and Dimensionless Blockage Index show increases in identifying more dams as stable the longer these had existed (with some fluctuation, particularly concerning the insignificant number of dams in the 10-to-20years category). Other factors, such as climatic conditions or seismicity also do not factor into these indices. These observations suggest that geomorphic stability indices 45
Journal Pre-proof may be of help to classify past landslide dams and tune parameters, but they are not reliable to make decisions for present cases , especially during an emergency. It is therefore time to develop geotechnical indices or stability criteria that consider the internal dam structure and dam composition (information that is hard to gather in the required risk management timeframes) for rapid assessment and prioritization of
Jo
ur n
al
Pr
e-
pr
oo
f
resources in an event (e.g. earthquake trigger) scenario.
Figure 7 Geomorphic stability indices, categorizing stable and unstable landslide dams verified against longevity for available cases from our new worldwide database. Key to read figure: numbers after longevity denote the number of data-available cases. Read left to right, i.e. for Dimensionless Blockage Index, for > 20 years longevity, 58 % cases are predicted stable, 17% cases are uncertain and 25% cases are falsely predicted as unstable.
46
Journal Pre-proof 4.2.3 Multivariate statistics on predicting the failure probability of a landslide dam Data-driven models employing geomorphometric parameters of landslide dams and lakes have demonstrated to be a useful approach for assessing landslide dam stability (e.g., Chen and Chang, 2016; Dong et al., 2011a; Dong et al., 2009; Hua and Shaoyu, 2015; Korup, 2004). Logistic regression (LR), support vector machine (SVM), and Fisher’s discriminant analysis (FDA) are typically used in deriving landslide
oo
f
susceptibility maps and are also applied for landslide dams. These multivariate techniques use a dependent variable which is categorical (e.g., stable dam or unstable
pr
dam), and often have a large number of the independent variables (geomorphometric
e-
parameters), which can be either categorical, numerical (most of them obtained from a digital elevation model, not always available immediately) or both (Dong et al., 2011b).
Pr
A list of geomorphometric parameters used in the data-driven models and their
al
definitions are provided in Table S2 (supplementary material).
ur n
The data should be divided into training data (consisting at least half, usually 70% of the total cases) and testing data (the remaining). The training cases with selected variables
Jo
are used to train the model, and the test dataset is for validating the model. The performance evaluations of prediction rates are evaluated using ROC (receiver operating characteristic) diagram and AUC (area under the curve) values. Alternatively, kappa values, accuracy or a confusion matrix which shows true positive, false positive, true negative, and false negative are often used as evaluation criteria. Hua and Shaoyu (2015) compared different data-driven methods using 121 landslide dams documented in Ermini and Casagli (2003) and Tabata et al. (2002). They show that SVM has about 10% higher predictive performance than LR and about 23% higher performance than DBI. Nevertheless, these approaches should always be used with caution since their 47
Journal Pre-proof performance in regions beyond the geographical location of inventories can produce failure probability maps that contain several false positives and false negatives (Dong et al., 2011a).
5 Landscape implications of river-blocking landslides Here we examine the effects of river-blocking landslides on broader landscape processes across a spectrum of spatial and temporal scales. Following the scheme of Hewitt (2002) (Figure 8), the impacts of landslide dams on the landscape can be
oo
f
categorized temporally as short- (days to months), medium- (years to decades), and long-term (centuries to millennia), and spatially as upstream- and downstream-effects.
pr
Long-term landslide dams naturally bring more important implications to the broader
e-
landscape and thus it is to those we devote lengthier discussion. Hewitt (2002) outlines
Jo
ur n
al
final removal (Table 14; Figure 8).
Pr
five stages that trace landscape responses from the time of dam emplacement to its
Figure 8. Timescales of landslide dam and lake processes outlining stages of response to river-blocking landslides (after Hewitt, 2002).
48
Journal Pre-proof Individual cases may vary widely and display some or all of these stages, and, the evidence of some stages may be absent from the geomorphic and geological record due to censuring by subsequent erosional and depositional processes. When compared to basin sedimentation, landslide dam impacts are local and unlikely to be preserved in the stratigraphic record (Miall, 2013; Reading and Levell, 1996). However, over the Himalaya at least 177 (+/- 31) km3 of valley fill (15-20% of all valley fill) is associated with mass wasting, mostly as 57 large landslide deposits (Blöthe and Korup, 2013).
oo
f
Impacts are significant over human timescales, and this framework provides insights to hazard cascades associated with landslide-dam formation, failure, and beyond. In the
pr
following, we review the stages, link them to the indices and inventories discussed
e-
above, and suggest future research directions aimed at filling data and knowledge
Pr
gaps.
al
Table 14. Stages in the evolution of a landslide-interrupted river, identifying characteristic features and how they combine to underpin a land systems approach described below (modified after Hewitt, 2002; Korup, 2005a; Korup, 2005b). MAIN DEVELOPMENT
Stage 1:
Stage 2:
Landslide emplacement and creation of a dam
Landslide debris lobes, hummocks, pressure ridges, furrows, ‘brandung’; rudaceous or fragmented deposits
Aggradation and constructional landforms upstream of dam, submergence/inundation of upstream areas, lake sedimentation, reduction of discharge, changes in flow regime, decrease in upstream gradient.
Axial, fluvio-lacustrine flats, multi-channel (braided, anastomosing) streams; delta(s) and fan deltas, cross -bedded scour and fill, scattered swamp deposits; redistribution of fine-grained sediment by valley winds and dust storms; dune fields, loess, stream piracy
Jo
Landslide dam formation
TYPICAL LANDFORMS AND DEPOSITS
ur n
STAGE
Lake impoundment
Valley side and trib utaries; talus, colluvium, sediment fans, debris flow cones, disturbed confluence environments. On-landslide; knickpoints, megaclasts, ponds, loess.
49
Journal Pre-proof
Stage 4: Superimposed interruption
Stage 5:
Strath terraces superimposed (epigenetic) gorges, pothole swarms ‘defended’ landforms, outwash fan
Channels and straths superimposed in bedrock; minor depositional Remnants in sheltered locations, remnant irregularities in river thalwegs
Exhumed, pre-landslide valley floor sediments, epigenetic gorges Migrating knickpoints in bedrock
Short-term impacts of landslide dams
e-
5.1
Incision into the pre-landslide floor, an exhumation of buried valley fill & channels, increasingly scattered, smaller remnants of impoundment complex, increased flood frequency
Reworked and mixed sediments
pr
‘Shadow’ interruption
Spillway across the dam, segmented fans, trenched valley fill, discontinuous erosion terraces; migrating knickpoints in valley fill, alluvial flats.
f
Erosional interruption
Overtopping or piping, outburst flood, excavation, trenching and removal of the impoundment complex, changes in channel morphology, downstream sedimentation
oo
Stage 3:
Pr
5.1.1 Dam formation, impoundment, and inundation The most immediate disruption is the blockage itself (Hammack and Wohl, 1996; Korup,
al
2005b; Ohmori, 1992). This occurs synchronously with emplacement of the landslide
sediments
ur n
over seconds to minutes (Stage 1, Figure 8). The landslide dam retains water and from upstream, consequently causing
water and sediment deficit
1997).
Jo
downstream, which may experience erosion due to the lack of sediment load (Kondolf,
Filling a landslide-formed impoundment may take from days to years, though in some cases just hours (Stage 2). The rate of filling is a function of the input (runoff) versus output (evaporation and seepage) (Nash, 2003). For instance, the breached Köfels rockslide dam (Austria) with an estimated lake volume of ~2.4 km3 would have taken 46 months to fill assuming an annual (modern-day average) inflow of 19.8 m3 /s (Dufresne et al., 2018). The Sarez Lake (Figure 1a) formed by the Usoi landslide in 1911 attained close to its present-day level in the 1940s (Ischuk, 2011) and it appears that inflow now 50
Journal Pre-proof almost equals the outflow through seepage to maintain an equilibrium condition (Delaney and Evans, 2011). Conversely, a large (~60 million m3) rockslide dam formed ~5,000 years BP at Contrada Monte, Italy, is not yet filled due to its small (6.4 km 2 ) catchment area (Nicoletti and Parise, 2002). Impoundments in deep, narrow valleys or in areas of high runoff fill faster. A well-known example is the Hunza lake (600 million m3) that formed behind the January 2010 Attabad landslide (45 million m 3 ) in Pakistan, filled within 5 months, inundating 240 houses, 5 villages and about 25 km of the
oo
f
Karakoram national highway (Schneider et al., 2013). 5.1.2 Dam failure and flooding
pr
Examination of previous dam failures shows that ~80-90% fail within the first year of
e-
formation (Costa and Schuster, 1988; Clague and Evans, 1994; Tacconi Stefanelli et
Pr
al., 2018) (see Figs. 5 and 9); an observation confirmed in the aftermath of the 2008 Wenchuan earthquake (Fan et al., 2012c). Specific cases vary widely; for example, in
al
the northern Apennines the fraction is just 58 % (Canuti et al. (1998). The compositions
ur n
of landslide dams in the Costa and Schuster (1991) inventory are not known in detail, but observations from Himalayan landslide dams reported by Weidinger (2011) suggest
Jo
a relationship exists between landslide type (cf. Fig. 9), constituent grain size and dam longevity: dams containing fine-grained sediment (e.g., debris-flow dams) failing shortly after formation, whereas large rock avalanche/rockslide dams with overall larger grain sizes (up to a few 10s meters in diameter, especially typical of the upper carapace) have the potential to last up to tens of thousands of years. Our new database (Section 3 and Figure 9) also suggests that complex and debris flow type landslide dams tend to last only for brief periods (i.e., 1-month-to-1-year and 1-to10-years) followed by unconsolidated sediments. For example, 80% of complex, 65% of debris flows, and ~60% of unconsolidated sediments type of landslide dams have a 51
Journal Pre-proof longevity of less than 1 year, whereas rock/debris avalanches and rockslides can last longer. The high longevity of volcanic dams might be biased due to the limited number of cases presented in Table 2, most of which lasted more than 100 years, and the typically larger volumes of volcanic debris avalanches (up to 100s km3, whereas landslides from non-volcanic sources rarely exceed 1 km3). ‘Complex landslides’ initiate as one landslide type (e.g. rockslide) and successively transform into one (e.g. debris avalanche) or more (e.g. debris flow, earth flow) other
oo
f
types during emplacement. The landslide type relevant for the river blockage is typically one of the later, distal ones that reached the valley floor . It therefore makes sense that
pr
their longevity (except those formed by rock avalanches that also convert from
e-
rockslides) is at the lower end together with debris flows and flows/slides in
Pr
unconsolidated sediments. It forms one end-member in this distribution by never exceeding 1-year longevity. At the other end are dams from volcanic sources, all of
al
which lasted longer than 1-month. The difference of accumulated percentage of the
ur n
other 4 types (rock avalanches and rockslides, debris avalanches, debris flows, flows/slides in unconsolidated sediments) decreases with increasing time lapsed after
Jo
the dam formed. That is, the influence of landslide type or material blocking the river on the longevity of a landslide dam apparently decreases beyond the first year (25% difference when longevity < 1year; 15% difference for longevity between 1-to-10 years; 10% difference for longevity between 10-to-20 year). Beyond 10 years, about 30-40% of the dams survived regardless of the landslide type of rock/debris avalanches, rockslides, debris flows, or slides/flows in unconsolidated sediments. Post-emplacement processes that may favourably influence dam stability over time are consolidation of the dam body, minerals precipitating into pore spaces, or fine sediments washing into the landslide-dammed lake sealing the upstream face of the 52
Journal Pre-proof dam (e.g. Weidinger, 2011; Hewitt, 2011). Another factor that is seldom investigated or reported, but would affect river incision through the dam, is the across-valley dam profile governed by the ‘along-way’ debris distribution (Korchevsky et al., 2011). However, more systematic analyses that consider all influencing factors (dam volume, type of transformed landslide in the ‘complex’ category, river mean and peak discharge, dam internal structure, post-emplacement processes, and dam and impoundment morphometric parameters, etc.) as well as the number of data points per category for
oo
f
statistical significance and biases in data collection are needed for more reliable
Jo
ur n
al
Pr
e-
pr
inferences.
Figure 9. The longevity of landslide dams relative to landslide type according to our database. Landslide dams for which material type is unknown are not shown. The sooner 53
Journal Pre-proof cumulative percentage marches towards 100 %, the lesser the longevity of the landslide dam.
In dam breach scenarios, the severity of the dam-breach flood depends upon several factors, including (i) volume of outburst flows; (ii) geotechnical properties of damming material (i.e., grain size distribution, internal structure, friction angle, cohesion etc.); (iii) sediment load in the landslide-dammed lake; and (iv) amount of loose, easily erodible materials in the channel pathway (Schuster (2000). Dam breaching leads to stored
f
sediments being partially rerouted downstream either gradually or catastrophically—
oo
outburst floods are capable of transporting millions of m3 of debris (King et al., 1989).
pr
Some of the largest known cases of catastrophic dam failures are reported from China
e-
(Li, 1994; Li et al., 1986). For example, the 1967 breaching of the Tanggudong landslide on the Yangtze River generated > 50 m high waves and the flood travelled
Pr
1,000 km downstream (Nash et al., 2008; Nash, 2003). Earthquake-induced landslide
al
dams and consequent dam-beach floods can cause severe damage. For example, the
ur n
Dadu River flood after the 1786 earthquake (Dai et al., 2005), flooding after the 1920 Haiyuan earthquake in the loess plateau of northwest China (Havenith et al., 2015;
Jo
Huang, 2009), and the Minjiang River flood after the 1933 Diexi earthquake (Chai et al., 1995) all affected the downstream areas killing numerous people. Debris flows or river flooding triggered by breaching of large landslide dams can cause rare secondary blockages within the valley, but these are of short persistence. Examples include the 2004 catastrophic breach of the Tsatichhu landslide dam, Bhutan (Dunning et al., 2006), and the Ram Creek Dam, New Zealand (Nash et al., 2008).
54
Journal Pre-proof 5.2
Medium-term impacts of landslide dams
5.2.1 River morphology effects Mountain valleys have relatively steep long-profiles and commonly irregular width where confined sections alternate with discontinuous floodplains. Landslide dams typically impose spatial variation on rates of the fluvial incision (Korup, 2005a). Upstream of landslide dams, sedimentation reflects the raised local base level (to lake height). For instance, Ouimet et al. (2007) observed lower river gradients and small
oo
f
alluviated tributaries upstream while downstream tributaries continued to incise. Huang and Fan (2013) showed that five years after the 2008 Wenchuan earthquake some river
pr
beds remained aggraded by > 10 m, which may be detrimental to hydropower schemes,
e-
and also increases the probability of floods in the future.
Pr
In landslide-prone areas, Type-I landslide dams (according to Costa and Schuster, 1988) frequently exert strong control on river planform. For example, the Obi -Khingow
al
River in the northern Pamir, Tajikistan, has been repeatedly blocked by large debris
ur n
flows, either partially or completely (Luknitskiy, 1955), causing the channel thalweg to be displaced laterally between alternate valley margins. Similarly, the Chingshui River,
Jo
Taiwan, has undergone five successive blockages in the past 140 years downstream of Chingshui, resulting in channel width expansion, decreased channel curvature, and development of laterally-mobile channels and anabranches (Chen et al., 2008). 5.2.2 Sediment flux effects The mechanisms and range of extreme sediment transport events in fluvial settings and the global recycling of sediments were recently reviewed by Korup (2012). The study underscores that sediment fluxes in mountain rivers are significantly modulated by large landslide dams and their associated outburst floods. That said, few documented cases 55
Journal Pre-proof compare pre- and post-event sediment load. Pratt-Sitaula et al. (2007) reported that in steep rivers subject to blocking landslides, bedload accounts for ∼35% of the sediment transport and suspended load to ~50%. Large landslide dams with sizeable impoundment capacity can act to buffer peak river discharge downstream. The effects on sediment flux vary upstream and downstream of the landslide dam, but remain essentially the product of climate, basin relief and hillslope angle, lithology, and seismo-tectonics, as well as the environmental changes
oo
f
that have occurred in the geological past (Leeder et al., 1998). Upstream of the barrier dam, fluvial deltas and lacustrine sediments are commonly observed (Montgomery et
pr
al., 2004). Downstream of the dam, abrupt peaks in sediment flux during outburst floods
e-
can pose a significant risk to settlements and infrastructure, especially those developed
Pr
on alluvial fans subject to channel avulsion and increased flood frequency (Korup, 2004). In the wet tropics of Papua New Guinea, the 1985 Bairaman landslide dam
Long-term impacts of landslide dams
ur n
5.3
al
failure flushed sediment at a rate of ~5×10 9 t km−2 yr−1 (King et al., 1989; Korup, 2012).
For long-lived landslide dams, attention shifts from the dam itself to the modifications of
Jo
valley forms and deposits. Long-lived dams can control intermontane sedimentation, and the patterns and rates of denudation, notably by creating local base levels , and constraining sediment delivery (e.g. Korup, 2005a; Korup, 2005b) and peak discharge downstream. Hundreds of landslide dams have been identified as persisting for several millennia at least—in the European Alps (Abele, 1974; Heim, 1932), the Canadian Rockies (Cruden, 1985a; Cruden, 1985b; Eisbacher, 1979), the Southern Alps (Korup, 2002; Read et al., 1992; Whitehouse and Griffiths, 1983), the southern Andes (Hermanns et al., 2015; Hermanns et al., 2011a; Strecker and Marrett, 1999), in High Asia, including the Pamir, Tien Shan and Dzungaria (Abdrakhmatov and Strom, 2006; 56
Journal Pre-proof Strom, 2010; Strom, 2013; Strom, 2015; Strom and Abdrakhmatov, 2018), and the Nepal Himalaya (Weidinger, 2011). In the upper Indus basin, over 400 prehistoric landslide dams are now recognized, perhaps no more than one-third of those still awaiting identification (Evans et al., 2011a; Hewitt, 1988; Hewitt, 2011; Hewitt, 2014), yet their histories of longevity and failure/erosion remain unknown. Here, we outline what we may learn from paleo-dams by examining sediment archives recording their existence and their long-term impact on mountain rivers.
oo
f
5.3.1 Sedimentary archives of long-term landslide dams
Sedimentary archives of fluvial and lacustrine activity associated with numerous long -
pr
lived lakes can indicate high sediment loads during the Late Pleistocene and Holocene.
e-
For example, the thickness of lacustrine sediment (~15 ka to 30 ka BP) behind the Diexi
Pr
landslide dam on the Minjiang River, Sichuan, China, reaches 240 m (Wang and Han, 2014). Another example is the lake created by the Karakul rockslide dam on the Karasu
al
River, Tien Shan, which is now completely silted-up to a depth of 200 m and occupied
ur n
by the town of Karakul (Strom, 2010). The longevity of river damming can be assessed (at least to first approximation) based upon the sediment yield of the dammed river if
Jo
such data exist. Such an approach was used to estimate a ~2,200-3,100 lifespan for the 11-15 km3 sediment-filled lake created by the 8-12 km3 Beshkiol landslide on the Naryn River, central Tien Shan (Korup et al., 2006; Strom and Abdrakhmatov, 2018). Another important aspect of long-lived dams and their sedimentation record has been discussed in Korup and Montgomery (2008) and Korup et al. (2010). They pointed out that damming causes substantial hindering in river incision (impeded headward fluvial erosion), which protects the basins in the upper part of the river course. Thanks to sedimentary archives, we can also be alerted to cases in which a landslide dam and its lake remained stable for centuries to millennia before breaching 57
Journal Pre-proof catastrophically. The Casa de Los Loros dam o n the Las Conchas River, northwest Argentina, existed for ~7 ka before breaching 4.8 ± 0.5 ka BP, owing to a 225 × 106 m3 rock avalanche crashing into the lake 2 km upstream of the dam and causing an overtopping wave (Hermanns et al., 2004). Similar breaches caused by a landslide into a landslide-dammed lake are reported elsewhere (Glancy and Bell, 2000; Penna et al., 2008). The failure of the rockslide dam impounding Laguna Carri Lauquén, triggering an outburst flood over >900 km downstream is a good example of the impact of climate
oo
f
variability on dam stability. This dam existed for at least 425 yrs (Costa and Díaz, 2007; Hermanns et al., 2004) before breaching during heavy summer rains (González Díaz et
pr
al., 2001; Groeber, 1916); although the true cause may have been the recorded high
e-
temperatures of the preceding winter that promoted glacier melt and increased inflow to
Pr
the lake (Evans et al., 2011a; Evans et al., 2011b; Hermanns et al., 2004). Even once breached, landslide dams can impact the fluvial landscape for millennia as
al
they often persist in some remnant form, and they can add to or amplify the patchiness
ur n
that is typical of mountain rivers (Burchsted et al., 2014b; Burchsted and Daniels, 2014). The large Flims rockslide (8-12 km3 ), dated to around 8900 ± 700 yr B.P. (Ivy-Ochs et
Jo
al., 2009), is much-eroded, but still influences the upper Rhine River (von Poschinger et al., 2006), as do several landslides in the Ötztal Basin in the Alps (Dufresne et al., 2018; Hermanns et al., 2006), where knickpoints, extensive lake sediments, and outsized fans still dominate the valley some 3 ,000 years since the last blockage occurred. 5.3.2 Fragmentation of mountain river systems In many of the world’s high mountains, streams are repeatedly blocked, constricted, deflected or split into multiple channels by non-fluvial processes (Montgomery and Buffington, 1997; Wohl and Pearthree, 1991). Various studies find heterogeneity, 58
Journal Pre-proof discontinuities, or patchiness to be typical of mountain rivers (Benda et al., 2003; Benda et al., 2004; Burchsted et al., 2014a). However, some large landslide-related controls have been missed or misread as due to tectonic activity, climate change and, especially, glaciation (Fauqué et al., 2000; Fauqué and Tchilinguirian, 2002; Hewitt, 1999; Porter and Orombelli, 1981). Landslide dams have impacts comparable in scale and complexity to those identified with the ‘fragmented’ river systems concept (Hewitt, 2011); the ‘fragments’ being channel and valley sections above, below and between
oo
f
landslide dams. A majority of the obstructions may be one-time events that initiate local epicycles of
pr
deposition and erosion. And there has been a tendency to treat them as singular,
e-
extreme events, and their main geomorphic consequence as increases in sediment
Pr
yields (Gabet et al., 2008; Petley et al., 2007). However, the extent and significance of fragmented drainage in high mountains derive especially from multiple blockages along
landslides
occurring
upstream
and
downstream.
An
important
factor
ur n
other
al
the same streams. Individual cases endure long enough to affect and be affected by
predetermining landslide dam evolution is the presence of several dammed lakes along
Jo
the same river. Such multiple damming of a river predetermine the stability and longevity of the blockages at a large extent (Strom and Abdrakhmatov, 2018). If the upstream dam forms a large lake, it can transform flood runoff, accumulating summer water flow and releasing it gradually, thereby decreasing the erosional capacity supporting long-term stability of the downstream dam(s) that otherwise would be breached (see (Strom and Abdrakhmatov, 2018) for the Kulun rockslide dam and other cascading effects in the Central Asian region). On the other hand, breaching of an upstream dam may lead to the destruction of those downstream, as occurred during the
59
Journal Pre-proof 1963 Issyk Lake breach linked to the breach of the small Jarsay glacial lake upstream (Gerasimov, 1965; Strom and Abdrakhmatov, 2018). Landslide obstructions impose temporary local base levels of erosion that may survive for millennia. They are recurring aspects of some of the world’s highest mountain headwaters. In formerly glaciated highlands, they can be critical parts of, and controls over, Holocene conditions (eds. Evans et al., 2011b). Hewitt (2011) noted many landslide dam sites in the trans-Himalayan Upper Indus basin where streams mainly
oo
f
flow over thick valley fill. Bedrock channels comprise <2 % of the rivers in the Karakoram Himalaya. Large-scale storage of intermontane sediment and irregular
pr
channel slopes in the region mainly reflect drainage interruption by landslides. For
e-
example, the Yugo-Korphak (Y-K) rockslide-rock avalanche in the eastern Karakoram
Pr
(Figure 10a) and the Skardu Basin of the Indus River (Figure 10b) illustrate the large scale and scope of aggradation, as well as extensive reworking, trenching and removal
al
of valley fill. The extent of aggradation is unusually high for a region widely identified
ur n
with exceptional relief, high rates of tectonic uplift, and globally extreme erosion (Burbank, 2002; Cornwell et al., 2003; Shroder, 1994; Shroder Jr, 1998; Shroder Jr and
Jo
Bishop, 1998; Zeitler, 1985). There is a challenge for high mountain work that has emphasized steep channels incised in bedrock, or adjustments and evolution based on tectonics and climate change. These are surely important, but large landslide interruptions and associated deep valley-fills can counteract their influence (Korup and Montgomery, 2008; Korup et al., 2010). Sets of long-lived dams create opportunities to reconstruct (minimum) frequencies and spatial patterns of events, as well as landslide responses to changing climate and other conditions (Ho et al., 2017; Strom, 2015) if the inventory is large and complete enough.
60
al
Pr
e-
pr
oo
f
Journal Pre-proof
5.4
Jo
ur n
Figure 10. (a) The Yugo-Korphak (Y-K) rock avalanche dam on the Shyok River, eastern Karakoram (N 35.11, E 76.10), which blocked the Shyok (middle ground). Right panel: upstream view from where Y-K boulders mark the upper limits of the landslide dam. (b) The Indus River in the Skardu Basin, Baltistan, where the present river level is backed up behind the Katzarah rockslide (N 35.42, E 75.46). The original lake level of the landslide dam is represented by the terraces (50-70 m high) surrounding the basin.
Landslide dams and landscape evolution
Landslide dams can change the way fluvial systems function over geological timescales. Here, we outline how landslide dams cause fluvial knickpoint retreat and reorganisation of the river network. 5.4.1 Knickpoint development Knickpoints are local convexities or steepenings in river longitudinal profiles that occur as two main types: static and mobile. Static, or non-propagating knickpoints develop i n 61
Journal Pre-proof response to a resistant substrate, such as strong bedrock (Jansen, 2006) or where river-blocking landslide debris exceeds the sediment transport capacity of the river (Korup, 2006; Korup et al., 2010). Mobile knickpoints are typically the product of a change in climate-related river discharge or base-level fall due to tectonism (Crosby and Whipple (2006). These knickpoints behave as transient ‘incisional waves’ that propagate upstream at rates determined by discharge, bed slope, substrate erodibility, and sediment load (Gardner, 1983; Jansen et al., 2011; Yunus, 2016; Yunus et al.,
oo
f
2016). Most commonly, knickpoints are identified through visual analysis of ri ver profiles either
pr
mapped from topography or derived from a digital elevation model (DEM). Major
e-
knickpoints can be mapped quantitatively from an analysis of χ-space (Neely et al.,
Pr
2017), or channel slope versus drainage area (Foster and Kelsey, 2012; Zhang et al., 2011), and several application tools have recently appeared to facilitate this approach,
al
such as Stream Profiler tool (Whipple et al., 2007), Topotoolbox (Schwanghart and
ur n
Scherler, 2014; 2017)), KZPicker (Neely et al., 2017), GISknickfinder (Rengers, 2017), and LSDTopoTools (Gailleton et al., 2019).
Jo
Numerous examples of well-developed knickpoints associated with late Pleistocene, Holocene and recent landslide dams are reported from the Himalayas (Walsh et al., 2012), the Tibetan Mountains (Fan et al., 2019c), Central Asia (Korup, 2006; Strom and Abdrakhmatov, 2018) and the Southern Alps of New Zealand (Korup, 2006). In such landscapes, a downstream succession of multiple river-blocking landslides is rather common (Figure 11). Given that knickpoints can be generated by several different mechanisms, care is required for reliable interpretation. A normalised index of channel steepness known as 62
Journal Pre-proof Ksn has been used successfully to differentiate knickpoints at landslide dams from other types of knickpoints (Harrison et al., 2015; Korup, 2006). For example, detailed field surveys and DEM-based analysis of Ksn in the Nepal Himalayas indicate that steepness maxima and transitions to lower values reflect knickpoint formation associated with damming, subsequent alluviation, breaching, and incision (Walsh et al., 2012). The numbers and volumes of large landslides identified in some mountains, or parts of them, suggest they are a major factor in the unroofing of the orogen (Korup and Clague,
oo
f
2009; e.g. Malamud et al., 2004). And yet, the long-term effects of landslide dams would seem to retard fluvial incision and overall landscape denudation. One of the most
pr
critical effects of landslide dams concerns their role in potentially slowing down or
e-
stalling the upstream retreat of mobile knickpoints. Long-term blocking of mobile
Pr
knickpoints can halt the transmission of new base-level information throughout the landscape affecting long-profile development. Upstream of the landslide dam, river
al
incision abruptly declines and this altered coupling between rivers and hillslopes may
Jo
ur n
lead to a long-term reduction in the frequency of landsliding (Korup et al., 2010).
Figure 11 (a) Multiple knickpoints identified using ‘knickpointfinder’ (Schwanghart and Scherler, 2017) from the Minjiang River basin along 50 km stretch near Diexi, Sichuan, China—some of these are formed by massive rock avalanches. (b) Landslide boundary 63
Journal Pre-proof (white dashed lines) and knickpoint (red dot) location for a landslide dam induced by the 1933 Diexi earthquake. Image is a LiDAR DEM.
5.4.2 River piracy and network reorganisation River piracy, or stream capture, is the diversion of the headwaters of one stream into another, leading to a re-organisation of the stream network (Bishop, 1955). Piracy has been examined in terms of tectonic motion (Perucca et al., 2018; Valdiya, 1996), rapid base-level fall (Calvache and Viseras, 1997), and glacial environments (Carter et al.,
oo
f
2013; Shugar et al., 2017), but few studies have focused on river piracy caused by river-blocking landslides. River catchments are especially vulnerable to piracy when a
pr
landslide dam stands at a higher elevation than the lowest catchment drainage divide
e-
(Figure 12). Piracy has profound implications for sediment flux, hazard assessment,
Pr
and for sedimentary records of environmental change (Mather, 2003). Korup (2005b) recognized stream piracy as a generalized landslide impact in the alpine
al
river systems of southwest New Zealand. Cases of piracy have also been observed in
ur n
the Patagonian and the central Andes related to rock avalanches of volumes ranging 0.5-3 km3 (Fauqué et al., 2000; González Díaz et al., 2000; Hermanns et al., 2011a).
Jo
Yet more ancient cases are difficult to track due to their low preservation potential. However, investigation of lacustrine sediments tied to landslide debris provides the key to identifying stream captures from the past, but is not explored much. Abandonment of the Unaweep Canyon in the upper Colorado River, western USA, is thought to reflect stream piracy linked to a landslide dam (Aslan et al., 2014; Hood et al., 2014). Balco et al. (2013) and Oesleby (2005) suggested that the abandonment was caused by landslide-damming and the subsequent spillway in the western part of the Unaweep Canyon diverted the Gunnison River. 64
Journal Pre-proof
Figure 12 Schematic representation of river piracy caused by a landslide dam
oo
f
6 Conclusions and recommendations for future research Firstly, this review of landslide dams unified the wide range of terminologies used in the
pr
literature and recommended their use in future standardization for sensible cross-
e-
comparison of methods and data. In addition, we suggest the creation of a citable
Pr
encyclopaedia of landslide dam terminologies for cross -reference. Secondly, we reviewed the existing classification schema and proposed a preliminary landslide dams
al
stability classification matrix which can be used for emergency response. Further, we
ur n
compiled the existing landslide dam inventories and created a worldwide landslide dam database comprising 410 cases (since 1900 and with volumes >10 6 m3 ). Based on this
Jo
new database, various geomorphic indices for landslide dam formation and stability were reviewed and compared. None of these indices is found satisfactory; the challenge that remains is to develop a well-performing dam stability index with the least number of input parameters; mainly because the geotechnical properties of dam composition materials are often not available. Despite the growing number of publications on landslide dam case studies and event-based and regional landslide inventories, few landslide dams have complete data entries (e.g. triggers, failure mechanisms, accurate longevities, types of damming landslides, key parameters of dam stability, etc.), which hamper better understanding o f dam formation conditions and 65
Journal Pre-proof stability assessment. Finally, the short- to long- term impacts of the landslide dam were reviewed. The following are the research challenges and possible future research directions that we identify as most important: (1) Most landslide dam classifications are based on morphologic features, and neglect dam composition and internal geological structure. More well-studied cases formed by different types of landslides and new field measurement methods (e.g. geophysical methods) are encouraged, which will help to create a new classification that better
oo
f
serves landslide dam stability assessment.
pr
(2) Maintain and update the worldwide database with more data entries for as many case studies as possible to provide a sound statistical basis for any failure mechanism
e-
and longevity studies, and to produce adequate and representable input parameters for
Pr
models and hazard assessment. Climatic events and earthquake-triggered landslides
al
should be separated in order to examine their formation and failure.
ur n
(3) Landslide dam stability indices are so far based on heterogeneous datasets and need to be refined and tested for landslide dams of different materials, geo-tectonic and
Jo
climatic settings, and differentiated for events of different sizes. Physically-based and numerical models of dam formation, failure, and routing of water and sediment outbursts are highly required. (4) More work needs to be done on the influence of multiple dams, and whether they cluster spatially and temporally (and how dating methods can truly resolve a cluster versus a sequence), and how they, in turn, affect the stability of their downstream neighbours and the hazard cascades. (5) Catastrophic failures of landslide dams and resultant outburst floods increase sediment transport in mountain rivers. Quantifying this terrestrial sediment flux and 66
Journal Pre-proof linking it to the total sediment budget, and studies on the long-term orogenic processes warrant special attention. Future studies should be linked to approaches involving detachment-limited or transport-limited models that systematically estimate the deposit volumes and fractions to allow us to address key controls on incision rates and understand the sediment cascade in terms of temporal fluxes. (6) To detect ancient landslide dams in river profiles and differentiate knickpoints resulting from landslide-damming from those of other drivers (tectonic, climatic, base
oo
f
level, and lithological signals) is the key in developing a complete understanding of how large-scale landslide dams affect erosion and fluvial geomorphology. More research
pr
efforts are needed for developing a scalable landscape evolution model.
e-
(7) Further exploitation of lacustrine sediments will shed light on the reconstruction of
Pr
paleoenvironmental and paleoseismic events, as well as their signal in long -term
al
landscape evolution.
ur n
Acknowledgements
This research is financially supported by the National Science Fund for Outstanding
Jo
Young Scholars of China (Grant No. 41622206), the Funds for Creative Research Groups of China (Grant No. 41521002), and the Fund for International Cooperation (NSFC-RCUK_NERC), Resilience to Earthquake-induced landslide risk in China (grant No.
41661134010),
National Key R&D Program of China (No. 2017YFC1501002).
Sincere thanks to the editor, Prof. Chris Fielding, for the kind acceptance of our proposal and encouragement to submit this research work. Many thanks also go to editorial assistant Mr Tim Horscroft for the kind support. We thank the three anonymous reviewers whose comments improved the clarity and structure of this paper.
67
Journal Pre-proof
Competing interests Authors declare no competing interests.
Supplementary data The supplementary data for this manuscript is attached separately. A summary of the attached files is given below. 1. Supplementary_readme: A word file detailing the supplementary data and tables.
oo
f
2. Database: An excel file having details of the comprehensive database. 3. List_of_references_database: A word file listing the references cited in the
pr
database.
Jo
ur n
al
Pr
e-
4. Data_collection_template: A word file for the data collection template.
68
Journal Pre-proof
References
Jo
ur n
al
Pr
e-
pr
oo
f
Abdrakhmatov, K. and Strom, A., 2006. Dissected rockslide and rock avalanche deposits; Tien Shan, Kyrgyzstan, Landslides from Massive Rock Slope Failure. Springer, pp. 551-570. Abele, G., 1974. Bergsturze in den Alpen. Ihre Verbreitung, Morphologie und Folgeerscheinungen. Wiss. Alpenvereinshefte, 25. Adams, J., 1980. Contemporary uplift and erosion of the Southern Alps, New Zealand. Geological Society of America Bulletin, 91(1_Part_II): 1-114. Adams, J., 1981. Earthquake-dammed lakes in New Zealand. Geology, 9(5): 215-219. Alford, D. and Schuster, R.L., 2000. Usoi Landslide Dam and Lake Sarez—An Assessment of Hazard and Risk in the Pamir Mountains. Tajikistan: United Nations Secretariat for International Strategy for Disaster Reduction, Geneva. Amante, C. and Eakins, B.W., 2009. ETOPO1 Global Relief Model converted to PanMap layer format. NOAA-National Geophysical Data Center: Boulder, CO, USA, 10. Aslan, A., Hood, W.C., Karlstrom, K.E., Kirby, E., Granger, D.E., Kelley, S., Crow, R., Donahue, M.S., Polyak, V. and Asmerom, Y., 2014. Abandonment of Unaweep Canyon (1.4–0.8 Ma), western Colorado: Effects of stream capture and anomalously rapid Pleistocene river incision. Geosphere, 10(3): 428-446. Balco, G., Soreghan, G.S., Sweet, D.E., Marra, K.R. and Bierman, P.R., 2013. Cosmogenic-nuclide burial ages for Pleistocene sedimentary fill in Unaweep Canyon, Colorado, USA. Quaternary Geochronology, 18: 149-157. Benda, L., Miller, D., Bigelow, P. and Andras, K., 2003. Effects of post-wildfire erosion on channel environments, Boise River, Idaho. Forest Ecology and Management, 178(1-2): 105-119. Benda, L.E.E., Andras, K., Miller, D. and Bigelow, P., 2004. Confluence effects in rivers: interactions of basin scale, network geometry, and disturbance regimes. Water Resources Research, 40(5). Bishop, A.W., 1955. The use of the slip circle in the stability analysis of slopes. Geotechnique, 5(1): 7-17. Blöthe, J.H. and Korup, O., 2013. Millennial lag times in the Himalayan sediment routing system. Earth and Planetary Science Letters, 382: 38-46. Blöthe, J.H., Korup, O. and Schwanghart, W., 2015. Large landslides lie low: Excess topography in the Himalaya-Karakoram ranges. Geology, 43(6): 523-526. Bonnard, C., 2006. Technical and human aspects of historic rockslide dammed lakes and landslide dam breaches. Italian Journal of Engineering Geology and Environment(Special Issue 1): 21-30. Burbank, D.W., 2002. Rates of erosion and their implications for exhumation. Mineralogical Magazine, 66(1): 25-52. Burchsted, D., Daniels, M. and Wohl, E.E., 2014a. Introduction to the special issue on discontinuity of fluvial systems. Burchsted, D., Daniels, M. and Wohl, E.E., 2014b. Introduction to the special issue on discontinuity of fluvial systems. Discontinuities in Fluvial Systems, 205: 1-4. Burchsted, D. and Daniels, M.D., 2014. Classification of the alterations of beaver dams to headwater streams in northeastern Connecticut, USA. Geomorphology, 205: 36-50. Calvache, M.L. and Viseras, C., 1997. Long‐term control mechanisms of stream piracy processes in Southeast Spain. Earth Surface Processes and Landforms: The Journal of the British Geomorphological Group, 22(2): 93-105. 69
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Canuti, P., Casagli, N. and Ermini, L., 1998. Inventory of landslide dams in the Northern Apennine as a model for induced flood hazard forecasting. Managing HydroGeological Disasters in a Vulnerable Environment. CNR-GNDCI and UNESCO IHP, Perugia: 189-202. Capra, L., 2006. Abrupt climatic changes as triggering mechanisms of massive volcanic collapses. Journal of Volcanology and Geothermal Research, 155(3-4): 329-333. Capra, L., 2007. Volcanic natural dams: identification, stability, and secondary effects. Natural Hazards, 43(1): 45-61. Capra, L., 2011. Volcanic natural dams associated with sector collapses: textural and sedimentological constraints on their stability, Natural and artificial rockslide dams. Springer, pp. 279-294. Capra, L. and Macıas, J.L., 2002. The cohesive Naranjo debris-flow deposit (10 km3):: A dam breakout flow derived from the Pleistocene debris-avalanche deposit of Nevado de Colima Volcano (México). Journal of Volcanology and Geothermal Research, 117(1-2): 213-235. Carey, M., Huggel, C., Bury, J., Portocarrero, C. and Haeberli, W., 2012. An integrated socio-environmental framework for glacier hazard management and climate change adaptation: lessons from Lake 513, Cordillera Blanca, Peru. Climatic Change, 112(3): 733-767. Carey, S., Sigurdsson, H., Mandeville, C. and Bronto, S., 1996. Pyroclastic flows and surges over water: an example from the 1883 Krakatau eruption. Bulletin of Volcanology, 57(7): 493-511. Carter, S.P., Fricker, H.A. and Siegfried, M.R., 2013. Evidence of rapid subglacial water piracy under Whillans Ice Stream, West Antarctica. Journal of Glaciology, 59(218): 1147-1162. Casagli, N. and Ermini, L., 1999. Geomorphic analysis of landslide dams in the Northern Apennine. Transactions of the Japanese Geomorphological Union, 20(3): 219-249. Casagli, N., Ermini, L. and Rosati, G., 2003. Determining grain size distribution of the material composing landslide dams in the Northern Apennines: sampling and processing methods. Engineering geology, 69(1-2): 83-97. Chai, H.J., Liu, H.C. and Zhang, Z.Y., 1995. The catalog of Chinese landslide dam events. Journal of Geological Hazards and Environment Preservation, 6(4): 1-9. Chen, C.-Y. and Chang, J.-M., 2016. Landslide dam formation susceptibility analysis based on geomorphic features. Landslides, 13(5): 1019-1033. Chen, S.C., Hsu, C.L., Wu, T.Y., Chou, H.T. and Cui, P., 2011. Landslide dams induced by typhoon Morakot and risk assessment, 5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment, Padua, Italy, pp. 653-660. Chen, X., Wu, J. and Hu, Q., 2008. Simulation of climate change impacts on streamflow in the Bosten Lake Basin using an artificial neural network model. Journal of Hydrologic Engineering, 13(3): 180-183. Chow, V.T., 1959. Open-channel hydraulics, Open-channel hydraulics. McGraw-Hill. Clague, J.J. and Evans, S.G., 1994. Formation and failure of natural dams in the Canadian Cordillera. Geological Survey of Canada Bulletin, 464 pp. Clavero, J.E., Sparks, R.S.J., Pringle, M.S., Polanco, E. and Gardeweg, M.C., 2004. Evolution and volcanic hazards of Taapaca volcanic complex, Central Andes of northern Chile. Journal of the Geological Society, 161(4): 603-618. Coe, J.A., Baum, R.L., Allstadt, K.E., Kochevar Jr, B.F., Schmitt, R.G., Morgan, M.L., White, J.L., Stratton, B.T., Hayashi, T.A. and Kean, J.W., 2016. Rock-avalanche dynamics revealed by large-scale field mapping and seismic signals at a highly 70
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
mobile avalanche in the West Salt Creek valley, western Colorado. Geosphere, 12(2): 607-631. Cornwell, K., Norsby, D. and Marston, R., 2003. Drainage, sediment transport, and denudation rates on the Nanga Parbat Himalaya, Pakistan. Geomorphology, 55(1-4): 25-43. Cortés, A., 2002. Depósitos de avalancha y flujos de escombros originados hace 3,600 años por el colapso del sector Suroeste del Volcán de Colima. Cortés, A., Macías, J.L., Capra, L. and Garduño-Monroy, V.H., 2010. Sector collapse of the SW flank of Volcán de Colima, México: the 3600 yr BP La Lumbre–Los Ganchos debris avalanche and associated debris flows. Journal of Volcanology and Geothermal Research, 197(1-4): 52-66. Costa, C.H. and Díaz, E.F.G., 2007. Age constraints and paleoseismic implication of rock avalanches in the northern Patagonian Andes, Argentina. Journal of South American Earth Sciences, 24(1): 48-57. Costa, J.E. and Schuster, R.L., 1988. The formation and failure of natural dams. Geological society of America bulletin, 100(7): 1054-1068. Costa, J.E. and Schuster, R.L., 1991. Documented historical landslide dams from around the world. 2331-1258, US Geological Survey. Crosby, B.T. and Whipple, K.X., 2006. Knickpoint initiation and distribution within fluvial networks: 236 waterfalls in the Waipaoa River, North Island, New Zealand. Geomorphology, 82(1-2): 16-38. Crozier, M.J. and Pillans, B.J., 1991. Geomorphic events and landform response in south-eastern Taranaki, New Zealand. CATENA, 18(5): 471-487. Cruden, D.M., 1985a. Destructive mass movements in high mountains: hazard and management, Geological Survey of Canada Paper 84-16: Book review. Canadian Geotechnical Journal, 22(3): 426-426. Cruden, D.M., 1985b. Rock slope movements in the Canadian Cordillera. Canadian Geotechnical Journal, 22(4): 528-540. Cruden, D.M., Keegan, T.R. and Thomson, S., 1993. The landslide dam on the Saddle River near Rycroft, Alberta. Canadian Geotechnical Journal, 30(6): 1003-1015. Cruden, D.M., Lu, Z.Y. and Thomson, S., 1997. The 1939 Montagneuse River landslide, Alberta. Canadian Geotechnical Journal, 34(5): 799-810. Cruden, D.M. and Varnes, D.J., 1996. Landslides: investigation and mitigation. Chapter 3-Landslide types and processes. Transportation research board special report(247). Cui, P., Zhu, Y.-y., Han, Y.-s., Chen, X.-q. and Zhuang, J.-q., 2009. The 12 May Wenchuan earthquake-induced landslide lakes: distribution and preliminary risk evaluation. Landslides, 6(3): 209-223. Dai, F.C., Lee, C.F., Deng, J.H. and Tham, L.G., 2005. The 1786 earthquake-triggered landslide dam and subsequent dam-break flood on the Dadu River, southwestern China. Geomorphology, 65(3-4): 205-221. Dal Sasso, S.F., Sole, A., Pascale, S., Sdao, F., Bateman Pinzón, A. and Medina, V., 2014. Assessment methodology for the prediction of landslide dam hazard. Natural Hazards and Earth System Sciences, 14(3): 557-567. Davies, T.R. and McSaveney, M.J., 2011. Rock-avalanche size and runout–implications for landslide dams, Natural and Artificial Rockslide Dams. Springer, pp. 441-462. Davies, T.R. and McSaveney, M.J., 2012. Mobility of long-runout rock avalanches. Landslides–types, mechanisms and modeling. Edited by JJ Clague and D. Stead. Cambridge University Press: 50-58.
71
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Davies, T.R.H. and Korup, O., 2010. Sediment cascades in active landscapes. Sediment Cascades: An Integrated Approach. John Wiley and Sons, Chichester: 89-115. Delaney, K.B. and Evans, S.G., 2011. Rockslide Dams in the Northwest Himalayas (Pakistan, India) and the Adjacent Pamir Mountains (Afghanistan, Tajikistan), Central Asia, Natural and Artificial Rockslide Dams. Springer, pp. 205-242. Delaney, K.B. and Evans, S.G., 2015. The 2000 Yigong landslide (Tibetan Plateau), rockslide-dammed lake and outburst flood: review, remote sensing analysis, and process modelling. Geomorphology, 246: 377-393. Dhital, M.R., Rijal, M.L. and Bajracharya, S.R., 2016. Landslide dams and their hazard after the 25 April 2015 Gorkha Earthquake in Central Nepal. Lowland technology international: the officisl journal of the International Association of Lowland Technology (IALT)/Institute of Lowland Technology, Saga University, 18(2): 129140. Dong, J.-J., Li, Y.-S., Kuo, C.-Y., Sung, R.-T., Li, M.-H., Lee, C.-T., Chen, C.-C. and Lee, W.-R., 2011a. The formation and breach of a short-lived landslide dam at Hsiaolin village, Taiwan—part I: post-event reconstruction of dam geometry. Engineering geology, 123(1-2): 40-59. Dong, J.-J., Tung, Y.-H., Chen, C.-C., Liao, J.-J. and Pan, Y.-W., 2009. Discriminant analysis of the geomorphic characteristics and stability of landslide dams. Geomorphology, 110(3-4): 162-171. Dong, J.-J., Tung, Y.-H., Chen, C.-C., Liao, J.-J. and Pan, Y.-W., 2011b. Logistic regression model for predicting the failure probability of a landslide dam. Engineering Geology, 117(1-2): 52-61. Dufresne, A., Ostermann, M. and Preusser, F., 2018. River-damming, late-Quaternary rockslides in the Ötz Valley region (Tyrol, Austria). Geomorphology, 310: 153167. Dufresne, A., Wolken, G.J., Hibert, C., Bessette-Kirton, E.K., Coe, J.A., Geertsema, M. and Ekström, G., 2019. The 2016 Lamplugh rock avalanche, Alaska: deposit structures and emplacement dynamics. Landslides, 16(12): 2301-2319. Duman, T.Y., 2009. The largest landslide dam in Turkey: Tortum landslide. Engineering Geology, 104(1-2): 66-79. Dunning, S.A. and Armitage, P.J., 2011. The grain-size distribution of rock-avalanche deposits: implications for natural dam stability, Natural and artificial rockslide dams. Springer, pp. 479-498. Dunning, S.A., Mitchell, W.A., Rosser, N.J. and Petley, D.N., 2007. The Hattian Bala rock avalanche and associated landslides triggered by the Kashmir Earthquake of 8 October 2005. Engineering Geology, 93(3-4): 130-144. Dunning, S.A., Petley, D.N., Rosser, N.J. and Strom, A.L., 2005. The morphology and sedimentology of valley confined rock-avalanche deposits and their effect on potential dam hazard, Proceedings of the International Conference on Landslide Risk Management, Edited by O. Hungr, R. Fell, R. Couture, and E. Eberhardt. Taylor & Francis, Balkema, London, pp. 691-701. Dunning, S.A., Rosser, N.J., Petley, D.N. and Massey, C.R., 2006. Formation and failure of the Tsatichhu landslide dam, Bhutan. Landslides, 3(2): 107-113. Eisbacher, G.H., 1979. Cliff collapse and rock avalanches (sturzstroms) in the Mackenzie Mountains, northwestern Canada. Canadian Geotechnical Journal, 16(2): 309-334. Erismann, T.H. and Abele, G., 2013. Dynamics of rockslides and rockfalls. Springer Science & Business Media. 72
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Ermini, L., 2000. Elaborazione di un modello per la precisione dell’evoluzione di sbarramenti fluviali causati da frane, unpublished. Unpublished PhD thesis, University of Florence, 159. Ermini, L., 2003a. Gli sbarramenti d’alveo da frane: Criteri speditivi per la stesura di scenari evolutivi derivanti dalla loro formazione. AIGA-IConvegno Nazionale: 355-367. Ermini, L., 2003b. Gli sbarramenti fluviali causati da frane: criteri speditivi per la stesura di scenari evolutivi derivanti dalla loro formazione, I Congresso Associazione Italiana Geologia Applicata, Chieti, pp. 355-367. Ermini, L. and Casagli, N., 2003. Prediction of the behaviour of landslide dams using a geomorphological dimensionless index. Earth Surface Processes and Landforms: The Journal of the British Geomorphological Research Group, 28(1): 31-47. Evans, S.G., Delaney, K.B., Hermanns, R.L., Strom, A. and Scarascia-Mugnozza, G., 2011a. The formation and behaviour of natural and artificial rockslide dams; implications for engineering performance and hazard management, Natural and artificial rockslide dams. Springer, pp. 1-75. Evans, S.G., Hermanns, R.L., Strom, A. and Scarascia-Mugnozza, G., 2011b. Natural and artificial rockslide dams, 133. Springer Science & Business Media. Evans, S.G., Hu, X.Q. and Enegren, E.G., 1996. The 1973 Attachie slide, Peace River Valley, near Fort St. John, British Columbia, Canada: a landslide with a highvelocity flowslide component in Pleistocene sediments, Proceedings, 7th International Symposium on Landslides, pp. 715-720. Fan, X., 2013. Understanding the causes and effects of earthquake-induced landslide dams. Fan, X., Huang, R., van Westen, C.J., Xu, Q., Havenith, H.-B. and Jetten, V., 2014a. A Conceptual Event-Tree Model for Coseismic Landslide Dam Hazard Assessment. In: K. Sassa, P. Canuti and Y. Yin (Editors), Landslide Science for a Safer Geoenvironment. Springer International Publishing, Cham, pp. 605-608. Fan, X., Rossiter, D.G., van Westen, C.J., Xu, Q. and Görüm, T., 2014b. Empirical prediction of coseismic landslide dam formation. Earth Surface Processes and Landforms, 39(14): 1913-1926. Fan, X., Scaringi, G., Korup, O., West, A.J., van Westen, C.J., Tanyas, H., Hovius, N., Hales, T.C., Jibson, R.W., Allstadt, K.E., Zhang, L., Evans, S.G., Xu, C., Li, G., Pei, X., Xu, Q. and Huang, R., 2019a. Earthquake-Induced Chains of Geologic Hazards: Patterns, Mechanisms, and Impacts. Reviews of Geophysics, 0(0). Fan, X., Tang, C.X., Van Westen, C.J. and Alkema, D., 2012a. Simulating dam-breach flood scenarios of the Tangjiashan landslide dam induced by the Wenchuan Earthquake. Natural hazards and earth system sciences, 12(10): 3031. Fan, X., van Westen, C.J., Korup, O., Gorum, T., Xu, Q., Dai, F., Huang, R. and Wang, G., 2012b. Transient water and sediment storage of the decaying landslide dams induced by the 2008 Wenchuan earthquake, China. Geomorphology, 171: 58-68. Fan, X., van Westen, C.J., Xu, Q., Gorum, T. and Dai, F., 2012c. Analysis of landslide dams induced by the 2008 Wenchuan earthquake. Journal of Asian Earth Sciences, 57: 25-37. Fan, X., van Westen, C.J., Xu, Q., Gorum, T., Dai, F., Wang, G. and Huang, R., 2013. Spatial Distribution of Landslide Dams Triggered by the 2008 Wenchuan Earthquake. In: C. Margottini, P. Canuti and K. Sassa (Editors), Landslide Science and Practice. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 279285. 73
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Fan, X., Xu, Q., Alonso-Rodriguez, A., Siva Subramanian, S., Li, W., Zheng, G., Dong, X. and Huang, R., 2019b. Successive landsliding and damming of the Jinsha River in eastern Tibet, China: prime investigation, early warning, and emergency response. Landslides, 16(5): 1003-1020. Fan, X., Xu, Q., van Westen, C.J., Huang, R. and Tang, R., 2017. Characteristics and classification of landslide dams associated with the 2008 Wenchuan earthquake. Geoenvironmental Disasters, 4(1): 12. Fan, X., Yunus, A.P., Jansen, J.D., Dai, L., Strom, A. and Xu, Q., 2019c. Comment on ‘Gigantic rockslides induced by fluvial incision in the Diexi area along the eastern margin of the Tibetan Plateau’ by Zhao et al. (2019) Geomorphology 338, 27–42. Geomorphology: 106963. Fauqué, L., Cortés, J.M., Folguera, A. and Etcheverría, M., 2000. Avalanchas de rocas asociadas a neotectónica en el valle del río Mendoza, al sur de Uspallata. Revista de la Asociación Geológica Argentina, 55(4): 419-423. Fauqué, L. and Tchilinguirian, P., 2002. Villavil rockslides, Catamarca Province, Argentina. Catastrophic Landslides: Effects, Occurrence, and Mechanism: 30324. Ferrer, C., 1999. Represamientos y rupturas de embalses naturales (lagunas de obstrución) como efectos cosísmicos: Algunos ejemplos en los Andes venezolanos. Revista Geográfica Venezolana, 40(1): 109-121. Fletcher, L., Hungr, O. and Evans, S.G., 2002. Contrasting failure behaviour of two large landslides in clay and silt. Canadian Geotechnical Journal, 39(1): 46-62. Fort, M., 1987. Geomorphic and Hazards mapping in the dry, continental Himalaya: 1: 50,000 maps of Mustang District, Nepal. Mountain Research and Development: 222-238. Fort, M. and Peulvast, J.P., 1995. Catastrophic mass-movements and morphogenesis in the Peri-Tibetan Ranges: examples from West Kunlun, East Pamir and Ladakh. Steepland Geomorphology: 171-198. Foster, M.A. and Kelsey, H.M., 2012. Knickpoint and knickzone formation and propagation, South Fork Eel River, northern California. Geosphere, 8(2): 403416. Frank, F., McArdell, B.W., Huggel, C. and Vieli, A., 2015. The importance of entrainment and bulking on debris flow runout modeling: examples from the Swiss Alps. Natural Hazards and Earth System Sciences, 15(11): 2569-2583. Gabet, E.J., Burbank, D.W., Pratt-Sitaula, B., Putkonen, J. and Bookhagen, B., 2008. Modern erosion rates in the High Himalayas of Nepal. Earth and Planetary Science Letters, 267(3-4): 482-494. Gailleton, B., Mudd, S.M., Clubb, F.J., Peifer, D. and Hurst, M.D., 2019. A segmentation approach for the reproducible extraction and quantification of knickpoints from river long profiles. Earth Surface Dynamics, 7(1): 211-230. Gallino, G.L. and Pierson, T.C., 1985. Polallie Creek debris flow and subsequent dambreak flood of 1980, East Fork Hood River basin, Oregon, 2273. US Geological Survey. Gardner, T.W., 1983. Experimental study of knickpoint and longitudinal profile evolution in cohesive, homogeneous material. Geological Society of America Bulletin, 94(5): 664-672. Geertsema, M., 1998. Flowslides in waterlain muds of northwestern British Columbia, Canada, Proceedings of the 8th Congress of the International Association of Engineering Geology and the Environment, pp. 1913-1921. Geertsema, M., Blais-Stevens, A., Kwoll, E., Menounos, B., Venditti, J.G., Grenier, A. and Wiebe, K., 2018. Sensitive clay landslide detection and characterization in 74
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
and around Lakelse Lake, British Columbia, Canada. Sedimentary geology, 364: 217-227. Geertsema, M. and Clague, J.J., 2006. 1,000-year record of landslide dams at Halden Creek, northeastern British Columbia. Landslides, 3(3): 217-227. Geertsema, M. and Clague, J.J., 2008. Natural dams, temporary lakes, and outburst floods in western Canada, Proceedings of the 1st World Landslide Forum, pp. 211-214. Geertsema, M., Clague, J.J., Schwab, J.W. and Evans, S.G., 2006a. An overview of recent large catastrophic landslides in northern British Columbia, Canada. Engineering Geology, 83(1-3): 120-143. Geertsema, M. and Cruden, D.M., 2008. Travels in the Canadian Cordillera, Proceedings of 4th Canadian conference on geohazards, Quebec City, pp. 383390. Geertsema, M., Cruden, D.M. and Schwab, J.W., 2006b. A large rapid landslide in sensitive glaciomarine sediments at Mink Creek, northwestern British Columbia, Canada. Engineering Geology, 83(1-3): 36-63. Geertsema, M., Highland, L. and Vaugeouis, L., 2009. Environmental impact of landslides, Landslides–disaster risk reduction. Springer, pp. 589-607. Geertsema, M., Hungr, O., Schwab, J.W. and Evans, S.G., 2006c. A large rockslide– debris avalanche in cohesive soil at Pink Mountain, northeastern British Columbia, Canada. Engineering Geology, 83(1-3): 64-75. Gerasimov, V., 1965. The Issyk catastrophe in 1963 and its effects on geomorphology of the Issyk River Valley. Transactions of the All-union Geographic Society, 97(6): 541-547. Glancy, P.A. and Bell, J.W., 2000. Landslide-induced flooding at Ophir Creek, Washoe County, western Nevada, May 30, 1983. 2330-7102, US Geological Survey. Glicken, H. and Nakamura, Y., 1988. Restudy of the 1888 eruption of Bandai volcano, Japan, pp. 392-395. Gómez-Avalos, J.C., Macias, J.L., Siebe, C. and Ocola, A.L., 2004. Debris avalanche deposit of Hualca Hualca volcano and the formation of a volcanic dam in the Colca Valley, Arequipa e Peru. International Association of Volcanology and Chemistry of the Earth Interior General Assembly, Chile. Gong, W., Jiang, X., Zhou, H., Xing, J., Li, C. and Yang, K., 2018. Varied thermorheological structure, mechanical anisotropy and lithospheric deformation of the southeastern Tibetan Plateau. Journal of Asian Earth Sciences, 163: 108-130. González Díaz, E.F., Fauqué, L.E., Giaccardi, A.D. and Costa, C., 2000. Las lagunas de Varvar Co Campos y Varvar Co Tapia (N del Neuquén, Argentina): su relación con avalanchas de rocas. Revista de la Asociación Geológica Argentina, 55(3): 147-164. González Díaz, E.F., Giaccardi, A.D. and Costa, C.H., 2001. La avalancha de rocas del río Barrancas (Cerro Pelán), norte del Neuquén: su relación con la catástrofe del río Colorado (29/12/1914). Revista de la Asociación Geológica Argentina, 56(4): 466-480. Griswold, J.P. and Iverson, R.M., 2008. Mobility statistics and automated hazard mapping for debris flows and rock avalanches. US Department of the Interior, US Geological Survey. Groeber, P., 1916. Informe sobre las causas que han producido las crecientes del río Colorado (Territorios del Neuquén y La Pampa) en 1914. 1-29 p. Buenos Aires: Dirección General de Minas, Geología e Hidrogeología. Gruber, N., Gloor, M., Mikaloff Fletcher, S.E., Doney, S.C., Dutkiewicz, S., Follows, M.J., Gerber, M., Jacobson, A.R., Joos, F. and Lindsay, K., 2009. Oceanic 75
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
sources, sinks, and transport of atmospheric CO2. Global Biogeochemical Cycles, 23(1). Guthrie, R.H., Friele, P., Allstadt, K., Roberts, N., Evans, S.G., Delaney, K.B., Roche, D., Clague, J.J. and Jakob, M., 2012. The 6 August 2010 Mount Meager rock slide-debris flow, Coast Mountains, British Columbia: characteristics, dynamics, and implications for hazard and risk assessment. Natural Hazards and Earth System Sciences, 12(5): 1277-1294. Guzzetti, F. and Tonelli, G., 2004. Information system on hydrological and geomorphological catastrophes in Italy (SICI): a tool for managing landslide and flood hazards. Hammack, L. and Wohl, E., 1996. Debris-fan formation and rapid modification at Warm Springs rapid, Yampa River, Colorado. The Journal of Geology, 104(6): 729-740. Hancox, G.T. and Perrin, N.D., 2009. Green Lake Landslide and other giant and very large postglacial landslides in Fiordland, New Zealand. Quaternary Science Reviews, 28(11-12): 1020-1036. Harrison, L.M., Dunning, S.A., Woodward, J. and Davies, T.R.H., 2015. Post-rockavalanche dam outburst flood sedimentation in Ram Creek, Southern Alps, New Zealand. Geomorphology, 241: 135-144. Havenith, H.-B., Torgoev, I., Torgoev, A., Strom, A., Xu, Y. and Fernandez-Steeger, T., 2015. The Kambarata 2 blast-fill dam, Kyrgyz Republic:blast event, geophysical monitoring and dam structure modelling. Geoenvironmental Disasters, 2(1): 11. Heim, A., 1932. Landslides and human lives (Bergsturz und Menschenleben). Bi -Tech Publishers, Vancouver, BC: 196. Henderson, R.D. and Thompson, S.M., 1999. Extreme rainfalls in the Southern Alps of New Zealand. Journal of Hydrology (New Zealand): 309-330. Hermanns, R.L., 2013. Landslide dam. Encyclopedia of Natural Hazards: 602-606. Hermanns, R.L., Blikra, L.H., Naumann, M., Nilsen, B., Panthi, K.K., Stromeyer, D. and Longva, O., 2006. Examples of multiple rock-slope collapses from Köfels (Ötz valley, Austria) and western Norway. Engineering Geology, 83(1-3): 94-108. Hermanns, R.L., Dahle, H., Bjerke, P.L., Crosta, G.B., Anda, E., Blikra, L.H., Saintot, A. and Longva, O., 2013. Rockslide dams in møre og romsdal county, Norway, Landslide science and practice. Springer, pp. 3-12. Hermanns, R.L., Fauqué, L. and Wilson, C.G.J., 2015. 36Cl terrestrial cosmogenic nuclide dating suggests Late Pleistocene to Early Holocene mass movements on the south face of Aconcagua mountain and in the Las Cuevas–Horcones valleys, Central Andes, Argentina. Geological Society, London, Special Publications, 399(1): 345-368. Hermanns, R.L., Folguera, A., Penna, I., Fauqué, L. and Niedermann, S., 2011a. Landslide dams in the Central Andes of Argentina (northern Patagonia and the Argentine northwest), Natural and artificial rockslide dams. Springer, pp. 147176. Hermanns, R.L., Hansen, L., Sletten, K., Böhme, M., Bunkholt, H., Dehls, J.F., Eilertsen, R., Fischer, L., Lheureux, J.S. and Høgaas, F., 2012. Systematic geological mapping for landslide understanding in the Norwegian context. Landslide and engineered slopes: protecting society through improved understanding. Taylor & Francis Group, London: 265-271. Hermanns, R.L., Hewitt, K., Strom, A., Evans, S.G., Dunning, S.A. and ScarasciaMugnozza, G., 2011b. The classification of rockslide dams, Natural and artificial rockslide dams. Springer, pp. 581-593. Hermanns, R.L., Niedermann, S., Ivy-Ochs, S. and Kubik, P.W., 2004. Rock avalanching into a landslide-dammed lake causing multiple dam failure in Las 76
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Conchas valley (NW Argentina)—evidence from surface exposure dating and stratigraphic analyses. Landslides, 1(2): 113-122. Hermanns, R.L., Valderrama, P., Fauqué, L., Penna, I.M., Sepúlveda, S., Moreiras, S. and Zavala Carrión, B., 2012. LandsLIdes In tHe andes and tHe need to CommunICate on an InteRandean LeVeL on LandsLIde maPPIng and ReseaRCH. Revista de la Asociación Geológica Argentina, 69(3): 321-327. Hewitt, K., 1988. Catastrophic landslide deposits in the Karakoram Himalaya. Science, 242(4875): 64-67. Hewitt, K., 1999. Quaternary moraines vs catastrophic rock avalanches in the Karakoram Himalaya, northern Pakistan. Quaternary Research, 51(3): 220-237. Hewitt, K., 2002. Postglacial landform and sediment associations in a landslidefragmented river system: The TransHimalayan Indus streams, central Asia, Landscapes of transition. Springer, pp. 63-91. Hewitt, K., 2009a. Catastrophic rock slope failures and late Quaternary developments in the Nanga Parbat–Haramosh Massif, Upper Indus basin, northern Pakistan. Quaternary Science Reviews, 28(11-12): 1055-1069. Hewitt, K., 2009b. Rock avalanches that travel onto glaciers and related developments, Karakoram Himalaya, Inner Asia. Geomorphology, 103(1): 66-79. Hewitt, K., 2011. Rock avalanche dams on the Trans Himalayan upper Indus streams: a survey of late Quaternary events and hazard-related characteristics, Natural and artificial rockslide dams. Springer, pp. 177-204. Hewitt, K., 2014. Regions of risk: A geographical introduction to disasters. Routledge. Hewitt, K., Clague, J.J. and Orwin, J.F., 2008. Legacies of catastrophic rock slope failures in mountain landscapes. Earth-Science Reviews, 87(1-2): 1-38. Ho, K., Lacasse, S. and Picarelli, L., 2017. Slope Safety Preparedness for Impact of Climate Change. CRC Press. Hood, W.C., Aslan, A. and Betton, C., 2014. Aftermath of a stream capture: Cactus Park lake spillover and the origin of East Creek, Uncompahgre Plateau, western Colorado. Geosphere, 10(3): 447-461. Hua, B. and Shaoyu, W., 2015. An Improved Classification Algorithm Applied on Landslide Dam Disaster Events Detection. Huang, R., 2009. Some catastrophic landslides since the twentieth century in the southwest of China. Landslides, 6(1): 69-81. Huang, R. and Fan, X., 2013. The landslide story. Nature Geoscience, 6(5): 325. Hungr, O., Evans, S.G. and Hutchinson, I.N., 2001. A Review of the Classification of Landslides of the Flow Type. Environmental & Engineering Geoscience, 7(3): 221-238. Hungr, O., Leroueil, S. and Picarelli, L., 2014. The Varnes classification of landslide types, an update. Landslides, 11(2): 167-194. Hutchinson, J.N., 2006. Massive rock slope failure: perspectives and retrospectives on state-of-the-art, Landslides from Massive Rock Slope Failure. Springer, pp. 619662. Ischuk, A.R., 2011. Usoi rockslide dam and lake Sarez, Pamir mountains, Tajikistan, Natural and Artificial Rockslide Dams. Springer, pp. 423-440. Iverson, R.M., 1997. The physics of debris flows. Reviews of geophysics, 35(3): 245296. Iverson, R.M., George, D.L., Allstadt, K., Reid, M.E., Collins, B.D., Vallance, J.W., Schilling, S.P., Godt, J.W., Cannon, C.M. and Magirl, C.S., 2015. Landslide mobility and hazards: implications of the 2014 Oso disaster. Earth and Planetary Science Letters, 412: 197-208. 77
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Ivy-Ochs, S., Poschinger, A.V., Synal, H.A. and Maisch, M., 2009. Surface exposure dating of the Flims landslide, Graubünden, Switzerland. Geomorphology, 103(1): 104-112. Jansen, J.D., 2006. Flood magnitude–frequency and lithologic control on bedrock river incision in post-orogenic terrain. Geomorphology, 82(1-2): 39-57. Jansen, J.D., Fabel, D., Bishop, P., Xu, S., Schnabel, C. and Codilean, A.T., 2011. Does decreasing paraglacial sediment supply slow knickpoint retreat? Geology, 39(6): 543-546. Jermyn, C. and Geertsema, M., 2015. An Overview of Some Recent Large Landslide Types in Nahanni National Park, Northwest Territories, Canada, Engineering Geology for Society and Territory-Volume 1. Springer, pp. 315-320. Jibson, R.W., Harp, E.L., Schulz, W. and Keefer, D.K., 2006. Large rock avalanches triggered by the M 7.9 Denali Fault, Alaska, earthquake of 3 November 2002. Engineering geology, 83(1-3): 144-160. Jiskoot, H., 2011. Long‐runout rockslide on glacier at Tsar Mountain, Canadian Rocky Mountains: potential triggers, seismic and glaciological implications. Earth Surface Processes and Landforms, 36(2): 203-216. King, J., Loveday, I. and Schuster, R.L., 1989. The 1985 Bairaman landslide dam and resulting debris flow, Papua New Guinea. Quarterly Journal of Engineering Geology and Hydrogeology, 22(4): 257-270. King, P.R., 2000. Tectonic reconstructions of New Zealand: 40 Ma to the present. New Zealand Journal of Geology and Geophysics, 43(4): 611-638. Kingma, J.T., 1959. The tectonic history of New Zealand. New Zealand journal of geology and geophysics, 2(1): 1-55. Kirby, E., Whipple, K.X., Burchfiel, B.C., Tang, W., Berger, G., Sun, Z. and Chen, Z., 2000. Neotectonics of the Min Shan, China: Implications for mechanisms driving Quaternary deformation along the eastern margin of the Tibetan Plateau. Geological Society of America Bulletin, 112(3): 375-393. Kirby, E., Whipple, K.X., Tang, W. and Chen, Z., 2003. Distribution of active rock uplift along the eastern margin of the Tibetan Plateau: Inferences from bedrock channel longitudinal profiles. Journal of Geophysical Research: Solid Earth, 108(B4). Koi, T., Hotta, N., Ishigaki, I., Matuzaki, N., Uchiyama, Y. and Suzuki, M., 2008. Prolonged impact of earthquake-induced landslides on sediment yield in a mountain watershed: The Tanzawa region, Japan. Geomorphology, 101(4): 692702. Kondolf, G.M., 1997. PROFILE: hungry water: effects of dams and gravel mining on river channels. Environmental management, 21(4): 533-551. Korchevskiy, V.F., Kolichko, A.V., Strom, A.L., Pernik, L.M. and Abdrakhmatov, K.E., 2011. Utilisation of data derived from large-scale experiments and study of natural blockages for blast-fill dam design, Natural and artificial rockslide dams. Springer, pp. 617-637. Korup, O., 2002. Recent research on landslide dams-a literature review with special attention to New Zealand. Progress in Physical Geography, 26(2): 206-235. Korup, O., 2004. Geomorphometric characteristics of New Zealand landslide dams. Engineering Geology, 73(1-2): 13-35. Korup, O., 2005a. Geomorphic hazard assessment of landslide dams in South Westland, New Zealand: fundamental problems and approaches. Geomorphology, 66(1-4): 167-188. Korup, O., 2005b. Geomorphic imprint of landslides on alpine river systems, southwest New Zealand. Earth Surface Processes and Landforms, 30(7): 783-800. 78
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Korup, O., 2006. Rock-slope failure and the river long profile. Geology, 34(1): 45-48. Korup, O., 2011. Rockslide and rock avalanche dams in the Southern Alps, New Zealand, Natural and Artificial Rockslide Dams. Springer, pp. 123-145. Korup, O., 2012. Earth's portfolio of extreme sediment transport events. Earth-Science Reviews, 112(3-4): 115-125. Korup, O. and Clague, J.J., 2009. Natural hazards, extreme events, and mountain topography. Quaternary Science Reviews, 28(11-12): 977-990. Korup, O. and Montgomery, D.R., 2008. Tibetan plateau river incision inhibited by glacial stabilization of the Tsangpo gorge. Nature, 455(7214): 786. Korup, O., Montgomery, D.R. and Hewitt, K., 2010. Glacier and landslide feedbacks to topographic relief in the Himalayan syntaxes. Proceedings of the National Academy of Sciences, 107(12): 5317-5322. Korup, O., Strom, A.L. and Weidinger, J.T., 2006. Fluvial response to large rock-slope failures: Examples from the Himalayas, the Tien Shan, and the Southern Alps in New Zealand. Geomorphology, 78(1-2): 3-21. Korup, O. and Tweed, F., 2007. Ice, moraine, and landslide dams in mountainous terrain. Quaternary Science Reviews, 26(25-28): 3406-3422. Lagmay, A.M.F., de Vries, B.v.W., Kerle, N. and Pyle, D.M., 2000. Volcano instability induced by strike-slip faulting. Bulletin of Volcanology, 62(4-5): 331-346. Lee, C.F. and Dai, F.C., 2011. The 1786 Dadu River Landslide Dam, Sichuan, China, Natural and Artificial Rockslide Dams. Springer, pp. 369-388. Leeder, M.R., Harris, T. and Kirkby, M.J., 1998. Sediment supply and climate change: implications for basin stratigraphy. Basin research, 10(1): 7-18. Lefebvre, G., Rosenberg, P., Paquette, J. and Lavallée, J.G., 1991. The September 5, 1987, landslide on the La Grande River, James Bay, Quebec, Canada. Canadian Geotechnical Journal, 28(2): 263-275. Li, T., 1994. Landslide disasters and human responses in China. Mountain Research and Development: 341-346. Li, T., Schuster, R.L. and Wu, J., 1986. Landslide dams in south-central China, Landslide dams: processes, risk, and mitigation. ASCE, pp. 146-162. Locat, A., Demers, D., Locat, P. and Geertsema, M., 2017. Sensitive clay landslides in Canada, Proceedings of the 70th Canadian Geotechnical Conference, Ottawa, Paper. Luknitskiy, P.N., 1955. Travel over the Pamirs. Molodaya Gvardia Publishers, Moscow. Macías, J.L., Capra, L., Scott, K.M., Espíndola, J.M., García-Palomo, A. and Costa, J.E., 2004. The 26 May 1982 breakout flows derived from failure of a volcanic dam at El Chichón, Chiapas, Mexico. Geological Society of America Bulletin, 116(1-2): 233-246. Malamud, B.D., Turcotte, D.L., Guzzetti, F. and Reichenbach, P., 2004. Landslide inventories and their statistical properties. Earth Surface Processes and Landforms, 29(6): 687-711. Mason, K., 1929. Indus floods and Shyok glaciers. Himalayan Journal, 1: 10-29. Massey, C., Townsend, D., Dellow, G., Lukovic, B., Rosser, B., Archibald, G., Villeneuve, M., Davidson, J., Jones, K., Morgenstern, R. and et al., 2018. Kaikoura Earthquake Short-Term Project: landslide inventory and landslide dam assessments. GNS Science report; 2018/19, GNS Science, Lower Hutt (NZ). Mather, A., 2003. Book Review: Flood and megaflood processes and deposits: recent and ancient examples. The Holocene, 13(6): 954-955. Melekestsev, I.V., Dirksen, O.V. and Girina, O.A., 1999. A giant landslide-explosion circue and debris avalanche at Bakening volcano, Kamchatka. Journal of Volcanology and Seismology, 20(3): 265-279. 79
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Meyer, W., Sabol, M.A. and Schuster, R., 1986. LANDSLIDE DAMMED LAKES AT MOUNT ST. HELENS, WASHINGTON, Landslide Dams: Processes, Risk, and Mitigation. Proceedings of a Session in Conjunction with the ASCE Convention., pp. 21-41. Miall, A.D., 2013. Principles of sedimentary basin analysis. Springer Science & Business Media. Miller, B., Dufresne, A., Geertsema, M., Atkinson, N., Evensen, H. and Cruden, D., 2018. Longevity of dams from landslides with sub-channel rupture surfaces, Peace River region, Canada. Geoenvironmental Disasters, 5(1): 1. Miller, B.G. and Cruden, D.M., 2002. The Eureka River landslide and dam, Peace River Lowlands, Alberta. Canadian Geotechnical Journal, 39(4): 863-878. Montgomery, D.R. and Buffington, J.M., 1997. Channel-reach morphology in mountain drainage basins. Geological Society of America Bulletin, 109(5): 596-611. Montgomery, D.R., Hallet, B., Yuping, L., Finnegan, N., Anders, A., Gillespie, A. and Greenberg, H.M., 2004. Evidence for Holocene megafloods down the Tsangpo River gorge, southeastern Tibet. Quaternary Research, 62(2): 201-207. Nakamura, Y. and Glicken, H., 1988. Blast and debris-avalanche deposits of the 1888 eruption, Bandai volcano. Journal of Geography (Chigaku Zasshi), 97(4): 309316. Nash, T., Bell, D., Davies, T. and Nathan, S., 2008. Analysis of the formation and failure of Ram Creek landslide dam, South Island, New Zealand. New Zealand journal of geology and geophysics, 51(3): 187-193. Nash, T.R., 2003. Engineering geological assessment of selected landslide dams formed from the 1929 Murchison and 1968 Inangahua earthquakes. Neely, A.B., Bookhagen, B. and Burbank, D.W., 2017. An automated knickzone selection algorithm (KZ‐Picker) to analyze transient landscapes: Calibration and validation. Journal of Geophysical Research: Earth Surface, 122(6): 12361261. Nicoletti, P.G. and Parise, M., 2002. Seven landslide dams of old seismic origin in southeastern Sicily (Italy). Geomorphology, 46(3-4): 203-222. O'Connor, J.E. and Costa, J.E., 2004. The world's largest floods, past and present: their causes and magnitudes, 1254. US Geological Survey. Oesleby, T.W., 2005. Thick sediment fill in Unaweep Canyon: Implications for the history of the Uncompahgre Uplift, western Colorado, Geological Society of America Abstracts with Programs, pp. 16. Ohmori, H., 1992. Dynamics and erosion rate of the river running on a thick deposit supplied by a large landslide. Zeitschrift fur Geomorphologie Neue Folge, 36: 129-140. Ouimet, W.B., Whipple, K.X., Royden, L.H., Sun, Z. and Chen, Z., 2007. The influence of large landslides on river incision in a transient landscape: Eastern margin of the Tibetan Plateau (Sichuan, China). Geological Society of America Bulletin, 119(11-12): 1462-1476. Peng, M. and Zhang, L.M., 2012. Breaching parameters of landslide dams. Landslides, 9(1): 13-31. Penna, I.M., Derron, M.-H., Volpi, M. and Jaboyedoff, M., 2013. Analysis of past and future dam formation and failure in the Santa Cruz River (San Juan province, Argentina). Geomorphology, 186: 28-38. Penna, I.M., Hermanns, R.L. and Folguera, A., 2008. Remoción en masa y colapso catastrófico de diques naturales generados en el frente orogénico andino (36º38ºs): los casos Navarrete y Río Barrancas. Revista de la Asociación Geológica Argentina, 63(2): 172-180. 80
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Perucca, L.P., Espejo, K., Angillieri, M.Y.E., Rothis, M., Tejada, F. and Vargas, M., 2018. Neotectonic controls and stream piracy on the evolution of a river catchment: a case study in the Agua de la Peña River basin, Western Pampean Ranges, Argentina. Journal of Iberian Geology, 44(2): 207-224. Petley, D.N., Hearn, G.J., Hart, A., Rosser, N.J., Dunning, S.A., Oven, K. and Mitchell, W.A., 2007. Trends in landslide occurrence in Nepal. Natural hazards, 43(1): 2344. Pollet, N., Cojean, R., Couture, R., Schneider, J.-L., Strom, A.L., Voirin, C. and Wassmer, P., 2005. A slab-on-slab model for the Flims rockslide (Swiss Alps). Canadian geotechnical journal, 42(2): 587-600. Porter, S.C. and Orombelli, G., 1981. Alpine Rockfall Hazards: Recognition and dating of rockfall deposits in the western Italian Alps lead to an understanding of the potential hazards of giant rockfalls in mountainous regions. American Scientist, 69(1): 67-75. Pratt-Sitaula, B., Garde, M., Burbank, D.W., Oskin, M., Heimsath, A. and Gabet, E., 2007. Bedload-to-suspended load ratio and rapid bedrock incision from Himalayan landslide-dam lake record. Quaternary Research, 68(1): 111-120. Pulgarín, B., Capra, L., Cepeda, H. and Macías, J.L., 2004. Late Pleistocene deposits associated with a southern flank collapse of the Nevado del Huila volcanic complex (Colombia). Acta vulcanologica, 16(1/2): 1000-1022. Read, S.A.L., Beetham, R.D. and Riley, P.B., 1992. Lake Waikaremoana barrier–a large landslide dam in New Zealand, Proceedings of the Sixth International Symposium on Landslides, pp. 1481-1487. Reading, H.G. and Levell, B.K., 1996. Controls on the sedimentary rock record. Sedimentary environments: processes, facies and stratigraphy: 5-36. Rickenmann, D., 2005. Runout prediction methods, Debris-flow hazards and related phenomena. Springer, pp. 305-324. Ruegg, J.C., Kasser, M., Tarantola, A., Lepine, J.C. and Chouikrat, B., 1982. Deformations associated with the El Asnam earthquake of 10 October 1980: Geodetic determination of vertical and horizontal movements. Bulletin of the Seismological Society of America, 72(6A): 2227-2244. Salvati, P., Bianchi, C., Rossi, M. and Guzzetti, F., 2010. Societal landslide and flood risk in Italy. Natural Hazards and Earth System Sciences, 10(3): 465-483. Sanhueza-Pino, K., Korup, O., Hetzel, R., Munack, H., Weidinger, J.T., Dunning, S., Ormukov, C. and Kubik, P.W., 2011. Glacial advances constrained by 10 Be exposure dating of bedrock landslides, Kyrgyz Tien Shan. Quaternary Research, 76(3): 295-304. Schauwecker, S., Rohrer, M., Acuña, D., Cochachin, A., Dávila, L., Frey, H., Giráldez, C., Gómez, J., Huggel, C., Jacques-Coper, M., Loarte, E., Salzmann, N. and Vuille, M., 2014. Climate trends and glacier retreat in the Cordillera Blanca, Peru, revisited. Global and Planetary Change, 119: 85-97. Scheidegger, A.E., 1973. On the prediction of the reach and velocity of catastrophic landslides. Rock Mechanics and Rock Engineering, 5(4): 231-236. Schneider, J.F., Gruber, F.E. and Mergili, M., 2013. Recent cases and geomorphic evidence of landslide-dammed lakes and related hazards in the mountains of Central Asia, Landslide science and practice. Springer, pp. 57-64. Schuster, R.L., 1986. Landslide dams: processes, risk, and mitigation. ASCE. Schuster, R.L., 2000. Dams built on pre-existing landslides, ISRM International Symposium. International Society for Rock Mechanics and Rock Engineering. Schuster, R.L., 2006. IMPACTS OF LANDSLIDE DAMS ON MOUNTAIN VALLEY MORPHOLOGY. In: S.G. Evans, G.S. Mugnozza, A. Strom and R.L. Hermanns 81
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
(Editors), Landslides from Massive Rock Slope Failure. Springer Netherlands, Dordrecht, pp. 591-616. Schuster, R.L. and Alford, D., 2004. Usoi landslide dam and lake sarez, Pamir mountains, Tajikistan. Environmental & Engineering Geoscience, 10(2): 151-168. Schuster, R.L. and Costa, J.E., 1986. PERSPECTIVE ON LANDSLIDE DAMS, pp. 120. Schuster, R.L. and Crandell, D.R., 1984. Catastrophic debris avalanches from volcanoes, Fourth Int Symp Landslides Proc Toronto, pp. 567-572. Schwanghart, W. and Scherler, D., 2017. Bumps in river profiles: uncertainty assessment and smoothing using quantile regression techniques. Earth Surface Dynamics, 5(4). Shen, Z.-K., Sun, J., Zhang, P., Wan, Y., Wang, M., Bürgmann, R., Zeng, Y., Gan, W., Liao, H. and Wang, Q., 2009. Slip maxima at fault junctions and rupturing of barriers during the 2008 Wenchuan earthquake. Nature geoscience, 2(10): 718. Shroder, J.F., 1994. Book Review: Geology and tectonics of the Karakoram mountains. MP Searle. Wiley, New York, 1991. 358 pp., 218 illustrations, 1: 250,000 geology and structure map in color, slipcase. Price: $298. ISBN 0-471-92773-2. Geomorphology, 9: 79-81. Shroder Jr, J.F., 1998. Slope failure and denudation in the western Himalaya. Geomorphology, 26(1-3): 81-105. Shroder Jr, J.F. and Bishop, M.P., 1998. Mass movement in the Himalaya: new insights and research directions. Geomorphology, 26(1-3): 13-35. Shugar, D.H., Clague, J.J., Best, J.L., Schoof, C., Willis, M.J., Copland, L. and Roe, G.H., 2017. River piracy and drainage basin reorganization led by climate-driven glacier retreat. Nature Geoscience, 10(5): 370. Strecker, M.R. and Marrett, R., 1999. Kinematic evolution of fault ramps and its role in development of landslides and lakes in the northwestern Argentine Andes. Geology, 27(4): 307-310. Strom, A., 2006. Morphology and internal structure of rockslides and rock avalanches: grounds and constraints for their modelling, Landslides from Massive Rock Slope Failure. Springer, pp. 305-326. Strom, A., 2010. Landslide dams in Central Asia region. Journal of the Japan Landslide Society, 47(6): 309-324. Strom, A., 2013. Geological prerequisites for landslide dams’ disaster assessment and mitigation in Central Asia, Progress of Geo-Disaster Mitigation Technology in Asia. Springer, pp. 17-53. Strom, A., 2015. Natural river damming: climate-driven or seismically induced phenomena: basics for landslide and seismic hazard assessment, Engineering Geology for Society and Territory-Volume 2. Springer, pp. 33-41. Strom, A. and Abdrakhmatov, K., 2018. Rockslides and rock avalanches of Central Asia: distribution, morphology, and internal structure. Elsevier. Swanson, F.J., Fredriksen, R.L. and McCorison, F.M., 1982. Material transfer in a western Oregon forested watershed. Swanson, F.J., Oyagi, N. and Tominaga, M., 1986. Landslide dams in Japan, Landslide dams: processes, risk, and mitigation. ASCE, pp. 131-145. Tabata, S., Mizuyama, T. and Inoue, K., 2002. Landslide dams and disasters. Kokon Shoin: 205. Tacconi Stefanelli, C., Catani, F. and Casagli, N., 2015. Geomorphological investigations on landslide dams. Geoenvironmental Disasters, 2(1): 21. Tacconi Stefanelli, C., Segoni, S., Casagli, N. and Catani, F., 2016. Geomorphic indexing of landslide dams evolution. Engineering geology, 208: 1-10. 82
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Tacconi Stefanelli, C., Vilímek, V., Emmer, A. and Catani, F., 2018. Morphological analysis and features of the landslide dams in the Cordillera Blanca, Peru. Landslides, 15(3): 507-521. Umbal, J.V. and Rodolfo, K.S., 1996. The 1991 lahars of southwestern Mount Pinatubo and evolution of the lahar-dammed Mapanuepe Lake. Fire and mud: Eruptions and lahars of Mount Pinatubo, Philippines: Quezon, Philippines: 951-970. Valdiya, K.S., 1996. River piracy. Resonance, 1(5): 19-28. Varnes, D.J., 1978. Slope movement types and processes. Special report, 176: 11-33. von Poschinger, A., Wassmer, P. and Maisch, M., 2006. The Flims rockslide: history of interpretation and new insights, Landslides from massive rock slope failure. Springer, pp. 329-356. Walder, J.S. and O'Connor, J.E., 1997. Methods for predicting peak discharge of floods caused by failure of natural and constructed earthen dams. Water Resources Research, 33(10): 2337-2348. Walsh, L.S., Martin, A.J., Ojha, T.P. and Fedenczuk, T., 2012. Correlations of fluvial knickzones with landslide dams, lithologic contacts, and faults in the southwestern Annapurna Range, central Nepalese Himalaya. Journal of Geophysical Research: Earth Surface, 117(F1). Wang, F., Fan, X., Yunus, A.P., Siva Subramanian, S., Alonso-Rodriguez, A., Dai, L., Xu, Q. and Huang, R., 2019. Coseismic landslides triggered by the 2018 Hokkaido, Japan (M w 6.6), earthquake: spatial distribution, controlling factors, and possible failure mechanism. Landslides: 1-16. Wang, G., Furuya, G., Zhang, F., Doi, I., Watanabe, N., Wakai, A. and Marui, H., 2016. Layered internal structure and breaching risk assessment of the HigashiTakezawa landslide dam in Niigata, Japan. Geomorphology, 267: 48-58. Wang, X. and Han, M., 2014. Online sequential extreme learning machine with kernels for nonstationary time series prediction. Neurocomputing, 145: 90-97. Wartman, J., Montgomery, D.R., Anderson, S.A., Keaton, J.R., Benoît, J., dela Chapelle, J. and Gilbert, R., 2016. The 22 March 2014 Oso landslide, Washington, USA. Geomorphology, 253: 275-288. Waythomas, C.F. and Kaufman, D.S., 1991. Comment on:“Latest Pleistocene increase in wind intensity recorded in eolian sediments from central Alaska,” by N. Bigelow, JE Begét, and WR Powers. Quaternary Research, 36(3): 329-333. Waythomas, C.F. and Nye, C.J., 2001. Preliminary volcano-hazard assessment for Mount Spurr volcano, Alaska. 2331-1258, US Geological Survey. Wei, R., Zeng, Q., Davies, T., Yuan, G., Wang, K., Xue, X. and Yin, Q., 2018. Geohazard cascade and mechanism of large debris flows in Tianmo gully, SE Tibetan Plateau and implications to hazard monitoring. Engineering Geology, 233: 172-182. Weidinger, J.T., 2011. Stability and life span of landslide dams in the Himalayas (India, Nepal) and the Qin Ling Mountains (China), Natural and Artificial Rockslide Dams. Springer, pp. 243-277. Whipple, K.X., Wobus, C., Crosby, B., Kirby, E. and Sheehan, D., 2007. New tools for quantitative geomorphology: extraction and interpretation of stream profiles from digital topographic data. GSA short course, 506. Whitehouse, I.E. and Griffiths, G.A., 1983. Frequency and hazard of large rock avalanches in the central Southern Alps, New Zealand. Geology, 11(6): 331-334. Wohl, E.E. and Pearthree, P.P., 1991. Debris flows as geomorphic agents in the Huachuca Mountains of southeastern Arizona. Geomorphology, 4(3-4): 273-292.
83
Journal Pre-proof
Jo
ur n
al
Pr
e-
pr
oo
f
Worni, R., Huggel, C., Clague, J.J., Schaub, Y. and Stoffel, M., 2014. Coupling glacial lake impact, dam breach, and flood processes: A modeling perspective. Geomorphology, 224: 161-176. Xu, Q., Fan, X.-M., Huang, R.-Q. and Van Westen, C., 2009. Landslide dams triggered by the Wenchuan Earthquake, Sichuan Province, south west China. Bulletin of engineering geology and the environment, 68(3): 373-386. Yi, G.X., Wen, X.Z. and Su, Y.J., 2008. Study on the potential strong‐earthquake risk for the eastern boundary of the Sichuan‐Yunnan active faulted‐block, China. Chinese Journal of Geophysics, 51(6): 1151-1158. Yin, Y., Cheng, Y., Liang, J. and Wang, W., 2016. Heavy-rainfall-induced catastrophic rockslide-debris flow at Sanxicun, Dujiangyan, after the Wenchuan Ms 8.0 earthquake. Landslides, 13(1): 9-23. Yin, Y., Wang, F. and Sun, P., 2009. Landslide hazards triggered by the 2008 Wenchuan earthquake, Sichuan, China. Landslides, 6(2): 139-152. Yunus, A.P., 2016. Geomorphic and lithologic control on bedrock channels in drainage basins of the Western Arabian Peninsula. Arabian Journal of Geosciences, 9(2): 133. Yunus, A.P., Oguchi, T. and Hayakawa, Y.S., 2016. Remote identification of fluvial knickzones and their imprints on landscape morphology in the passive margins of Western Arabia. Journal of Arid Environments, 130: 14-29. Zeitler, P.K., 1985. Cooling history of the NW Himalaya, Pakistan. Tectonics, 4(1): 127151. Zhang, H., Zhang, P. and Fan, Q., 2011. Initiation and recession of the fluvial knickpoints: A case study from the Yalu River-Wangtian’e volcanic region, northeastern China. Science China Earth Sciences, 54(11): 1746.
84
Journal Pre-proof
Competing interests
Jo
ur n
al
Pr
e-
pr
oo
f
Authors declare no competing interests.
85