Ignorer kommandoer på båndet
Gå til hovedindhold
Naviger op
Log på
> Forsiden > Publikationer > Geology of Greenland Survey Bulletin > Vol. 191 Geol. Greenl. Surv. Bull. > Review of Greenland Activities 2001, pp 73-83

GEOLOGY OF GREENLAND SURVEY BULLETIN 191

 
Tsunami-generating rock fall and landslide on the south coast of Nuussuaq, central West Greenland

HTML-udgave af pdf-filen:   gsb191p73-83.pdf (856 Kb)

Spørgsmål og yderligere oplysninger, venligst kontakt artiklens forfatter(ere)

background image
During the afternoon of 21 November 2000 the village
of Saqqaq in central West Greenland was hit by a series
of giant waves. Ten small boats were destroyed, but
luckily neither humans nor dogs were killed. The
following day a police inspection by helicopter revealed
that the giant waves were caused by a major landslide
at Paatuut, c. 40 km north-west of Saqqaq on the south
coast of Nuussuaq (Figs 1, 2). The landslide deposits
were dark grey-brown in colour, in marked contrast to
the snow-covered slopes, and protruded as a lobe into
the Vaigat strait. Along the adjacent coastlines the
snow had been washed off up to altitudes about 50 m
a.s.l. and severe damage had been caused at the aban-
doned coal-mining town Qullissat on the opposite side
of Vaigat.
Landslides in Greenland
Landslides in Greenland are influenced by permafrost,
glacial ice, high topographic relief and repeated freezing
and thawing (Pedersen et al. 1989). In most parts of
Greenland where gneisses and granites dominate, slides
are rare. By contrast, the Nuussuaq Basin comprises
weakly consolidated sedimentary rocks overlain by a
thick pile of dense volcanic rocks. This stratigraphical
succession is favourable to the generation of slides.
Consequently, large parts of Disko, Nuussuaq and
Svartenhuk Halvø are strongly affected by landslides,
especially along the coasts (Fig. 2).
Tsunamis
In coastal areas landslides may reach the sea and gener-
ate swells which may travel considerable distances. The
swells present a much larger risk than do the landslides
because they affect much larger areas. A wave (swell)
may not be noticeable in open waters off the coast, but
grows rapidly as it approaches the shore. Swells of this
type are known as tsunamis and are frequently gener-
ated by submarine earthquakes. The giant waves which
hit Saqqaq, Qullissat and the shores of the Vaigat in
November 2000, are referred to below as a tsunami,
although they were initiated by a landslide.
73
Tsunami-generating rock fall and landslide on the south
coast of Nuussuaq, central West Greenland
Stig A. Schack Pedersen, Lotte Melchior Larsen,Trine Dahl-Jensen, Hans F. Jepsen, Gunver Krarup
Pedersen,Tove Nielsen, Asger Ken Pedersen, Frants von Platen-Hallermund and Willy Weng
1400 m a.s.l.
900 m a.s.l.
350 m a.s.l.
Head of slide
Bypass zone
Vaigat
Toe of slide
Sh
Fig. 1. The 21 November 2000 landslide
at Paatuut, viewed from the sea a few
days later. The slide lacks snow cover; cf.
Fig. 4. Note the black shale (Sh)
cropping out in the lower part of the
coastal escarpment. Photo: Christoffer
Schander, Arktisk Station, Qeqertarsuaq.
Geology of Greenland Survey Bulletin 191, 73­83 (2002) © GEUS, 2002
GSB191-Indhold 13/12/02 11:31 Side 73
background image
74
72
°
71
°
51
°
53
°
70
°
69
°
55
°
51
°
Svartenhuk
Halvø
Uummannaq
Illorsuit
Nuugaatsiaq
Niaqornat
Qaarsut
Asuk
Nuusap
Qaqqarsua
Ukkusissat
Disko
Disko Bugt
Ilulissat
Saqqaq
Fig. 3
Qeqertaq
Qasigiannguit
Qullissat
Paatuut
Aasiaat
Nuussuaq
Inland Ice
Qeqertarsuaq
Kangerluk
Vaigat
Sermersuaq
Ujarasussuk
Ataata Kuua
Greenland
·
Nuuk
·
·
Upernavik
Summit
500 km
Areas where landslides
may generate tsunamis
Areas with high risk of landslides
Palaeogene volcanic rocks
Cretaceous sandstones and shales
Precambrian basement
50 km
Fig. 2. Geological map of the
Nuussuaq Basin with the landslide-
prone areas indicated. Frame shows
location of Fig. 3, a detailed map of
the Paatuut area.
GSB191-Indhold 13/12/02 11:31 Side 74
background image
Landslides may move at different speeds, and slowly
moving slides or debris flows are not considered danger-
ous. To initiate a tsunami the landslide must almost
instantly displace huge volumes of water, and the likeli-
hood of such an event increases with the steepness of
the coast. The risk of tsunamis following landslides is
consequently particularly high along the steep south
coast of Nuussuaq where erosion is rapid and large
volumes of debris accumulate at high altitudes (Fig. 2).
Furthermore, tsunamis generated along the south coast
of Nuussuaq are confined by the narrow Vaigat strait
and are reflected from the opposite coast.
Rock falls and landslides in the
Disko­Nuussuaq area
Old slide areas are recognisable from their geomor-
phological expression, and sometimes also by brick-
red colours produced by self-combustion in slipped
carbon-rich shales. Such burnt lithologies are promi-
nent at Paatuut and Ataata Kuua (Fig. 3) and indicate
a high frequency of slides. Many slides are indicated
on the Survey's 1:100 000 geological maps of the
region. Most of these slides must be less than 10 000
years old, as they post-date the last glaciation. In the
period 10 000­3000 B.P. relative sea level was falling
in the Disko Bugt area, and marine terraces as well as
raised beaches were formed. Since 2800 B.P. sea level
has been rising (Long et al. 1999). Along that part of
the south coast of Nuussuaq where landslides are
frequent, marine terraces are generally not seen; this
indicates that the landslides present are probably
younger than 3000 B.P.
The Paatuut area is characterised by large landslides
(Fig. 3). Between Paatuut and Ataata Kuua in particular,
large lobes of rock debris superpose the glacial morpho-
logical surface. Huge solitary blocks of volcanic rocks on
the slopes facing Vaigat are evidence of occasional rock
falls, some of which may have generated tsunamis.
The first published account of a landslide in Green-
land is that of Steenstrup (1900), who recorded an
event that occurred in 1870 near Ujarasussuk on the
north coast of Disko. A mudflow with large floating
blocks of basalt built a lobe into the sea and hindered
passage on foot along the coast. This landslide does
not appear to have caused a tsunami or any damage
to houses or inhabitants in Ujarasussuk.
Landslides are a recurring phenomenon near Nia-
qornat in northern Nuussuaq. In most cases sponta-
neous combustion in shales is reported about one year
after the slide (Rosenkrantz 1967; Henderson 1969).
The village Niaqornat was hit by falling rocks in 1978,
and west of the village the risk of future slides is high
(Pulvertaft 1979).
75
Landslide
15.12.1952
Landslide
Paatuut
21.11.2000
Vaigat
52
°
45
53
°
70
°
15
70
°
18
70
°
21
Tupaasat
Ataata
Kuua
Landslide
Escarpment at head of slide
Basalt scree below steep basalt cliff
Alluvial fan
Undifferentiated Quaternary cover
Palaeogene basalt formations
Atane Formation with Quaternary cover
Atane Formation, sandstones interlayered
with shales and coal seams
5 km
Contour interval 100 m
Fig. 3. Map of part of the south coast of
Nuussuaq showing the many large and
small landslides in the area, including the
new slide at Paatuut. The contours are
drawn from vertical aerial photographs at
scale 1:150 000 taken in 1985 and, for the
Paatuut slide, from oblique photographs
taken in July 2001.
GSB191-Indhold 13/12/02 11:31 Side 75
background image
On the south coast of Nuussuaq, west of Tupaasat
(Fig. 3), a large landslide occurred on 15 December
1952. The slide was initiated 700 m a.s.l. and had a
width of 1 km where it reached the sea. It generated a
tsunami which caused some damage in Qullissat; areas
50­100 m from the shoreline were flooded, the elec-
tricity plant was damaged and a boat (22 ft cutter) with
a crew of three was washed ashore with the loss of
one man.
In the period 1984­1988 a mudflow at Paatuut was
active. Black shales derived from an altitude of 800 m
constituted the bulk of the flow. This section of the
coast appears to have been affected by a number of
flows ranging from dilute mudflows building alluvial
fans to denser mudflows. The last of these occurred in
1997 and is not reported to have generated a tsunami.
Evidence from geophysical investigations
The Vaigat strait is up to 650 m deep with a U-shaped
cross-section. This is interpreted as the result of erosion
during the Pleistocene by the Vaigat Ice Stream, which
drained from the Inland Ice into the Davis Strait.
During the summer of 2000, the Survey acquired a
large number of reflection seismic profiles in Vaigat
with a 600 m streamer and 40 l airgun array (Marcus-
sen et al. 2001). In the seismic profiles a number of
chaotic local accumulations of sediment are apparent,
which are interpreted to be old submarine slides or
subaqueous aggradations from subaerial landslides.
The thickness of such slide-deposited sediments is
generally 50­100 m, but may locally exceed 200 m.
The data do not allow a distinction between sediments
deposited by a single slide and the composite accu-
mulations formed by numerous slides.
The 21 November 2000 landslide
Stratigraphy
The stratigraphic succession at Paatuut comprises sili-
ciclastic sediments of the Cretaceous Atane Formation
and the Danian Quikavsak Formation, overlain by
hyaloclastite breccias and subaerial lava flows of the
Paleocene Vaigat and Maligât Formations (Pedersen &
Pulvertaft 1992; Pedersen et al. 1993; Dam & Sønder-
holm 1998; Dam & Nøhr-Hansen 2001).
The Atane Formation is exposed up to altitudes of
800 m. Lithologies consist of white to yellow, weakly
consolidated 1­10 m thick sandstones interbedded with
dark grey, 2­15 m thick shales and coal seams typically
0.5­1 m thick. In places the coal seams have under-
gone self-combustion after landslides, and the black
shales are converted into brick-red, orange or pink tiles.
The Quikavsak Formation constitutes palaeovalley fill
and is exposed locally at altitudes of 400­900 m. It
consists of whitish pebbly sand overlain by black
mudstones. The mudstone level around 900 m a.s.l.
corresponds to the level above which material was
removed during the landslide. This level is clearly seen
in the post-slide terrain model (see below).
The lowest volcanic rocks are 260 m thick hyalo-
clastite breccias of the Ordlingassoq Member of the
Vaigat Formation. At altitudes of 1160­1450 m these
are overlain by thin, greyish picritic subaerial lava
flows of the same member. The volcanic rocks above
1450 m a.s.l. are brownish subaerial lava flows of the
Maligât Formation. Dykes and up to 20 m thick sills cut
the succession, and the sills often form prominent ledges
on the sandstone slopes. Blocks of all the volcanic and
intrusive lithologies have been recognised amongst the
material in the landslide.
Slide morphology
The morphology and main features of the landslide
can be clearly seen in the photographs of Figs 1 and 4,
taken respectively a few days after the event and in
July 2001. The slide was initiated at altitudes of
1000­1400 m a.s.l., referred to as the head of the slide
(Fig. 1). The material transported by the slide compris-
es partly blocks of basalt that formed a steep scree
prior to the slide, and partly solid slabs of basalt
broken off from the cliff.
The central part of the slide is a series of gullies and
ridges in the Atane Formation; the material was trans-
ported downwards through these pre-existing gullies.
They were not deepened significantly due to erosion
during the slide, and this area is classified as a bypass
zone.
The lower part of the slide is a lobate accumulation
of material deposited on the former alluvial fan at the
base of the slope (Fig. 4). This aggradational toe of the
landslide has a width of c. 2 km and is bounded sea-
wards by a coastal escarpment (Fig. 4). The research
vessel Porsild collected a number of echo sounder
profiles off the coast of Paatuut in July 2001. These data
provide the first approach to a detailed map of the sea
floor (Pedersen et al. 2001, fig. 29) and reveal a rapid
increase in water depth and very uneven sea-floor
topography. Earlier seismic investigations have also
76
GSB191-Indhold 13/12/02 11:31 Side 76
background image
documented an uneven topography offshore Paatuut
(Marcussen et al. 2001), which is believed to reflect the
distribution of a succession of submarine slides.
Head of slide (denudation zone)
The slide escarpment along which material slipped and
slid bounds the head of the slide (Figs 1, 4). The escarp-
ment comprises two steep (c. 60°) surfaces, which meet
at a high angle. Along the eastern escarpment surface
the slip was down-dip, and the vertical offset seen on
the denudated surfaces ranges from 121 m to 182 m;
this surface is interpreted as a pre-existing fault plane,
and the newly exposed rock has a pale colour (Fig. 4).
The western surface constitutes a side escarpment
which was clearly affected by oblique lateral move-
ment during the landslide event, as indicated by
groove mark striation on the cliff. Based on the
disrupted character of the landslide material and the
steepness of the slide escarpment, the process at the
head of the slide is interpreted as a rock fall.
A platform at an altitude of 800­900 m occurs at the
base of the escarpment. Prior to the landslide this plat-
form was characterised by accumulation of glacial
deposits. The platform may mark the boundary
between the shaly lithologies and the sandstones of
the Quikavsak Formation and the hyaloclastite brec-
cias of the Ordlingassoq Member. After the slide, the
platform appears as a barren landscape characterised
by huge blocks and chaotic disrupted material on
which frost mounds have subsequently developed
(Pedersen et al. 2001).
Central part of slide (bypass zone)
The central part of the slide is a bypass zone where the
slide material was transported through the gullies in
the Atane Formation. The gradient in the gullies varies
77
1480
1400
1000
700
355
145
1786
1695
Basalt formations
Atane Formation
Basalt formations
Atane Formation
Escarpment
Avalanche
Side ridge
Side ridge
Maximum
cross-section
Coastal escarpment
Lake in mud tongue
Trench
Fig. 4. Oblique photograph of the Paatuut landslide taken in July 2001 from a helicopter. The photograph was used in the multi-model
photogrammetric work; low-lying areas are concealed by coastal fog. The direction of main flow in the bypass zone is indicated by
an arrow, and the cross-section of the avalanche at its maximum is marked by a dotted contour. Some ridges and trenches in the
toe of the slide are indicated. The light, smeared patch above the escarpment is a dust cloud raised by continuing small rock falls
from the back wall. Elevations are in m a.s.l.
GSB191-Indhold 13/12/02 11:31 Side 77
background image
from 26° to 36°. Field work in 2001 showed that some
slide material was deposited about 200 m above the
floor of the gullies during transport through the bypass
zone. Groove marks with a gentle plunge towards the
coast were observed on the sides of the gullies. A
minor quantity of white sandstones of the Atane
Formation was removed from the top of a gully and
deposited in a separate lobe in the aggradation zone
(lower right corner of Fig. 4). This suggests that the
transport and deposition of the sandstones occurred as
a discrete event during a late phase of the slide.
The drape of mud, stones and boulders on the sides
of the gullies outlines the space taken up by the flow.
The maximum cross-sectional area through which the
avalanche passed is calculated to 70 000 m
2
. Seismic data
(see below) show that the main event lasted for approx-
imately 80 seconds; however, the maximum cross-
section would only have been sustained during a
considerably shorter period. This gives a minimum
velocity of the slide through the bypass zone of 60 km/h,
indicating that in the bypass zone the landslide had the
character of a rock avalanche. The seismic data give
considerably higher average velocities.
Toe of slide (aggradation zone)
Morphologically the lower part of the slide comprises
two lobes which merge downwards into one, with a
surface lined by trenches, ridges and flats (Figs 4, 5).
The floors of the trenches may be as much as 50 m
below the crests of the ridges. The alluvial fan upon
which the landslide was deposited had a dip of 6°­9°,
and the aggradation is up to 60 m (see later). The
ridges consist of clast-supported, apparently matrix-
free conglomerates and are interpreted as grain flow
deposits. The flats consist of clast-supported, matrix-
bearing conglomerates and matrix-supported conglome-
rates, indicating deposition from modified grain flows
and debris flows. In the flat areas frost mounds are
common, indicating the presence of fine-grained sedi-
ment at the base of the deposit.
On the western side of the landslide a tongue of
mud extends away from the main slide. Field observa-
tions (Pedersen et al. 2001) indicate that this tongue
formed due to gravity spreading of the avalanche as it
settled on the alluvial fan. In the tail end depression of
this tongue a small lake was formed, dammed by the
toe of the slide (Fig. 4).
The toe of the slide reached the coast and advanced
into the sea, where a steep (37°) escarpment terminates
the subaerial part of the slide. This coastal escarpment
provides the only exposed section through the slide
deposits (Fig. 1); it shows a predominance of matrix-
supported conglomerates at the base of the slide and
a clast-supported boulder-conglomerate at the top.
Locally the slide contains a significant proportion of
black shales, probably originating from the mudstone
level around 900 m a.s.l.
The coastal escarpment indicates the collapse of a
protruding tongue which originally formed the front of
the toe. The collapse is believed to have generated a
submarine slide and subsequently the tsunami. Volume
calculations (see below) suggest that about 30 million m
3
of material slid below sea level.
78
Fig. 5. Blocks and boulders in the central
part of the toe of the Paatuut landslide
are interpreted as grain flow deposits.
Arrow indicates person for scale.
Photograph taken July 2001.
GSB191-Indhold 13/12/02 11:31 Side 78
background image
Effects of the tsunami
The tsunami generated by the landslide affected not
only Saqqaq but, as noted above, also the abandoned
coal mining village of Qullissat 20 km across the fjord
where buildings at altitudes below 30 m a.s.l. were
almost completely destroyed. A line of timber debris
100­250 m from the coast marks the upper limit of the
tsunami. A hut on the shore at Asuk 20 km from
Paatuut (Fig. 2) was completely destroyed.
On the large alluvial fan at the outlet from Ataata
Kuua 7 km north-west of Paatuut, several icebergs
were stranded 300­700 m from the coast. Wreckage,
including fish crates, was scattered across the alluvial
fan up to 800 m inland and at altitudes up to 40 m
a.s.l., and in the summer of 2001 certain vegetation
species in the same area had withered, presumably
due to saltwater incursion.
Seismic activity
A seismometer at Qeqertarsuaq registered a distinct
seismic signal on 21 November 2000 at 15:05:10 local
time (Fig. 6). People in Saqqaq noted that the tsunami
arrived at 15:15, and the seismic signal is thus clearly
related either to the landslide itself or to an earthquake
which triggered the landslide. Seismic stations at
Upernavik c. 250 km north of Paatuut and at Summit,
the drilling camp at the centre of the Inland Ice c. 550
km to the east (Fig. 2), also registered a seismic event
on 21 November 2000 (Fig. 6). An analysis of the seis-
mic records suggests that a seismic event occurred at
15:04:52 at Paatuut, and that it was probably caused by
the release of the basaltic cliff at the head of the slide
(Pedersen et al. 2001). This event is seen as the arrival
of the primary wave (P) and secondary and surface
waves (S and Lg) at all stations (Fig. 6). The following
stronger signal reflects the surface waves generated by
the activity of the avalanche, estimated to have a
magnitude of 2.3 on the Richter scale. The constant
time difference between the surface waves from the
first and the second events at seismic stations at
increasing distances from Paatuut indicates that the two
seismic signals were generated at Paatuut, separated by
approximately one minute. The main slide event was
accomplished within approximately 80 seconds. With
an average slide distance of 3 km, this implies an aver-
age slide velocity of 140 km/h. Maximum velocities
would have been close to 200 km/h.
The analysis of the seismic signals does not support
the hypothesis that a deep-seated earthquake generated
79
Fig. 6. Seismic events related to the rock
fall and landslide at Paatuut 21 November
2000 recorded at Qeqertarsuaq, Upernavik
and Summit (Fig. 2). The red lines mark
the calculated arrival times of P (primary),
S (secondary), and Lg (surface) waves
from an initial release event at Paatuut.
The time is determined by interpreting
the P arrival at Qeqertarsuaq. The initial
release of the fall occurred at 18:04:52
GMT (15:04:52 local time) and generated
P, S and Lg waves. Subsequently, the
energy released by the sliding of 90
million m
3
of rocks down onto the
coastal slopes is visible as an additional
surface wave culminating approximately
one minute after the initial event. This
second and much larger event ­ estimated
at 2.3 on the Richter scale ­ is also
visible in seismograms from Nuuk,
Tasiilaq, Thule and possibly Alert in
Canada. E, N and Z are the east­west,
north­south and vertical components in
the ground movements.
main slide
main slide
frost crack
15:04:52 local time
7: SUM N
(Summit)
1: GDH Z
(Qeqertarsuaq)
2: GDH N
(Qeqertarsuaq)
3: GDH E
(Qeqertarsuaq)
4: UPN Z
(Upernavik)
5: UPN N
(Upernavik)
6: UPN E
(Upernavik)
sum P
sum Lg
sum S
main slide
gdh P gdh S
gdh Lg
1 min.
upn P upn S upn Lg
1 min.
1 min.
seconds
0
100
150
200
50
-50
GSB191-Indhold 13/12/02 11:31 Side 79
background image
the landslide. This is in line with the statistics provid-
ed by Keefer (1999), who noted that earthquakes with
magnitudes below 4 are unlikely to generate landslides.
Terrain model and volume calculations
Photogrammetrical mapping at the Survey has provid-
ed two topographical models of the terrain around
Paatuut before and after the landslide. The first model
was based on vertical aerial photographs taken in 1985
(scale 1:150 000), and the second model on oblique
colour stereo photographs taken with a small-frame
camera from a helicopter during the field work in July
2001. From these models (Fig. 7) the volume of
displaced material (denudation) has been calculated to
c. 90 million m
3
. The subaerial volume of the toe of the
slide (aggradation) amounts to c. 60 million m
3
, which
indicates that c. 30 million m
3
of rock material entered
the sea giving rise to the tsunami.
Discussion of landslide dynamics
The mechanism that triggered the landslide at Paatuut
is open to discussion. The seismic data suggest that the
event was initiated by instantaneous fracturing. The
appearance of the escarpment at the head of the slide
indicates that the escarpment may have formed along
a pre-existing fault or fracture zone. The disrupted
character of the slide and the number of still ongoing
small rock falls accompanied by clouds of dust, as
observed in July 2001, support the idea of a fractured
cliff. However, the question remains as to what initiat-
ed the release of the large volume involved in the slide.
The weather situation in the area prior to the release
of the slide may be significant. A week before the
event the weather was cold with temperatures below
­10°C. Three days before the event a föhn wind blow-
ing from the Inland Ice brought the air temperature up
to about +6°C, but on the day before, the weather
returned to colder conditions with temperatures down
to ­6°C. Free water produced during passage of the
föhn may have seeped along fractures and fissures,
subsequently freezing when it came into contact with
the surrounding permafrozen rocks of the cliff and air
temperature again dropped below zero. Frost action
and resulting increased pore pressure may have trig-
gered the rock fall.
The transformation of the rock fall into an enor-
mous avalanche also raises questions. From the
section at the coastal escarpment, black shale can be
seen to have been involved in the landslide, and the
shale may have acted as a lubricant facilitating the
passage of the rock fall. The high velocity of the
avalanche during passage through the gullies may
have been accentuated by the presence of snow on
the slopes. The data available suggest that a combina-
tion of grain flow processes and debris flow processes
prevailed during transport through the bypass zone.
80
2000
1500
1000
500
0
0
500
1000
1500
2000
2500
3000
3500
4000
4500 m
m a.s.l.
A
B
C
D
E
AB = slide escarpment
BC = platform
CD = gully in Atane Formation
DE = aggradation zone
area of aggradation
area of denudation
SW
NE
Surface trace 1985
Surface trace 2001
coast
bypass zone
Fig. 7. The relationship between denudation (removal) and aggradation (deposition) regimes in a vertical topographic profile along
the landslide area, based on aerial photographs from 1985 and 2001.
GSB191-Indhold 13/12/02 11:31 Side 80
background image
Grain flow is indicated by the presence of boulders
deposited on the shoulders of the gullies at an eleva-
tion of 350 m a.s.l. Debris flow is interpreted from the
matrix-supported conglomerate at the base of the
section at the coastal escarpment, and this is further
suggested by the large number of frost-mounds in the
flat areas, reflecting the fine-grained character of the
subsurface material.
Based on the considerations provided here and on
the photogrammetric studies, a four-stage block diagram
model of the landslide was constructed and is presented
in Fig. 8.
Risk of landslides in the
Disko­Nuussuaq­Svartenhuk area
Areas of low risk include most of Svartenhuk Halvø,
where the mountains are neither high nor steep, and
catastrophic landslides and tsunamis are unlikely to
occur. Similarly the south coast of Nuussuaq south-east
of Paatuut has a gentle slope, and the basaltic succes-
sion is thin and some distance from the coast. Land-
slides are known to occur in this area (Koch 1959), but
mostly as slowly moving coherent slumps or block
slides which are unlikely to initiate tsunamis.
Areas of moderate risk are encountered on the north
coasts of Disko and Nuussuaq. Around Niaqornat slides
form at low altitudes and apparently cause little
damage, although houses should not be built closer to
the mountain or west of the present village (Pulvertaft
81
Collapse of slide foot
into the sea
triggers tsunami
Basalt
Sandstone
2
Avelanche and slide
Rock fall
3
4
1
Before slide
Mudstone
Fig. 8. The development of the landslide event at Paatuut on 21
November 2000, viewed from the west.
1, Cliff and valley slope prior to the landslide event. Note the
small crescent-shaped feature in the right side of the diagram which
is the head of a slide that was active in the last part of the 1980s.
2, Rock fall is triggered by frost fracturing and cliff cracking at
the head of the slide. A large amount of dust is released during
the cliff cracking.
3, Rock avalanche, grain flow and debris flow form the deposits
at the toe of the landslide, building out into the sea. Note the
lobe at the left side of the slide, which is mainly formed by a
debris flow squeezed out by gravity spreading from the main land-
slide aggradation area. In the tailing end depression of this lobe
a small lake was later formed, dammed by the toe of the slide.
4, One third of the landslide deposits, forming the tongue built
into the sea, subsequently collapses and continues as a subma-
rine slide, generating the tsunami. The steep coastal escarpment
constitutes the head of this slide.
GSB191-Indhold 13/12/02 11:32 Side 81
background image
1979). Along other sections of the north-facing coasts the
slides that occur are slow earth slides or rock glaciers.
Areas with a high risk of landslides accompanied by
tsunamis are found on the south coast of Nuussuaq
between Paatuut and Nuusap Qaqqarsua (Fig. 2).
Here, the very steep slopes and the active erosion of
the thick pile of hyaloclastite breccias generate thick,
steep, high-lying debris fans, which are unstable and
may easily slide.
Studies of aerial photographs show that at least 17
major landslides have occurred in the high-risk section
prior to 1985, but only three of these slides are known
to have reached the sea. In addition three landslides
occurred at Paatuut between 1988 and 2000. Four of
the twenty landslides generated within the last 3000
years reached the sea, and the minimum frequency of
tsunamis is thus 1­2 per 1000 years. However, two of
the four slides occurred within the last 50 years (1952
and 2000), and it may be that some older landslides
have not been recognised due to fluvial reworking.
The historical record indicates that catastrophic
landslides are not common. The town Qullissat was
founded in 1924 and, apart from the slide in 1952 which
caused some material damage and the loss of one life,
the town has not been damaged by tsunamis in
76 years. In the coastal cliff at Asuk, the tsunami in 2000
disturbed graves several hundred years old. Although
the graves were already somewhat damaged by ground
creep or perhaps previous tsunamis, it is obvious that a
tsunami like the one in 2000 is a rare event.
Prediction of future slides is not possible. However,
it is certain that the lithologies and the morphology of
the south coast of Nuussuaq from Paatuut and west-
wards will continue to generate landslides in the future.
Conclusions
The landslide at Paatuut on 21 November 2000 is
interpreted as a rock avalanche initiated by a rock fall
(Fig. 8). The fall was probably initiated by thawing
and freezing in fractures at altitudes of 1000­1400 m
a.s.l. The landslide removed 90 million m
3
of scree and
volcanic rocks which were transported to the coastal
area with an average velocity of about 140 km/h. It
built a lobe comprising c. 60 million m
3
, and c. 30
million m
3
continued seawards in a submarine slide
which initiated a tsunami and caused heavy damage in
the coastal areas of the Vaigat strait. The slide deposited
clast-supported conglomerates from modified grain
flows, and matrix-supported conglomerates from debris
flows. Despite the high risk of landslides on the south
coast of Nuussuaq the frequency of catastrophic land-
slides and tsunamis is not high. The seismic records
make the Paatuut 2000 landslide one of the best-docu-
mented events of its type.
Acknowledgements
This project was funded by the Bureau of Minerals and
Petroleum, Government of Greenland. Arktisk Station, Qeqertar-
suaq, is thanked for use of their research vessel Porsild and for
support during the field work.
References
Dam, G. & Nøhr-Hansen, H. 2001: Mantle plumes and sequence
stratigraphy: Late Maastrichtian ­ Early Paleocene of West
Greenland. Bulletin of the Geological Society of Denmark 48,
189­207.
Dam, G. & Sønderholm, M. 1998: Sedimentological evolution of
a fault-controlled Early Paleocene incised valley system,
Nuussuaq Basin, West Greenland. In: Shanley, K.W. & McCabe,
P.J. (eds): Relative role of eustasy, climate and tectonism in
continental rocks. Society of Economic Paleontologists and
Mineralogists Special Publication 59, 109­121.
Henderson, G. 1969: Oil and gas prospects in the Cretaceous­
Tertiary basin of West Greenland. Rapport Grønlands Geo-
logiske Undersøgelse 22, 63 pp.
Keefer, D.K. 1999: Earthquake-induced landslides and their effects
on alluvial fans. Journal of Sedimentary Research 69, 84­104.
Koch, B.E. 1959: Contribution to the stratigraphy of the non-
marine Tertiary deposits on the south coast of Nûgssuaq
Peninsula, Northwest Greenland. Bulletin Grønlands Geologiske
Undersøgelse 22, 100 pp. (also Meddelelser om Grønland 162(1)).
Long, A.J., Roberts, D.H. & Wright, R. 1999: Isolation basin stratig-
raphy and Holocene relative sea-level change on Arveprinsen
Ejland, Disko Bugt, West Greenland. Journal of Quaternary
Science 14(4), 323­345.
Marcussen, C., Chalmers, J.A., Andersen, H.L., Rasmussen, R. &
Dahl-Jensen, T. 2001: Acquisition of high-resolution multi-
channel seismic data in the offshore part of the Nuussuaq
Basin, central West Greenland. Geology of Greenland Survey
Bulletin 189, 34­40.
Pedersen, A.K., Larsen, L.M. & Dueholm, K.S. 1993: Geological
section along the south coast of Nuussuaq, central West
Greenland, 1:20 000, coloured sheet. Copenhagen: Geological
Survey of Greenland.
Pedersen, G.K. & Pulvertaft, T.C.R. 1992: The nonmarine Creta-
ceous of the West Greenland basin, onshore West Greenland.
Cretaceous Research 13, 263­272.
Pedersen, S.A.S., Foged, N. & Frederiksen, J. 1989: Extent and
economic significance of landslides in Denmark, Faroe Islands
and Greenland. In: Brabb, E.E. & Harrod, B.L. (eds): Landslides,
extent and economic significance, 153­156. Rotterdam: Balkema.
82
GSB191-Indhold 13/12/02 11:32 Side 82
background image
Pedersen, S.A.S., Dahl-Jensen, T., Jepsen, H., Larsen, L.M.,
Pedersen, G.K., Nielsen, T., Pedersen, A.K. & Weng, W. 2001:
Fjeldskred ved Paatuut. Danmarks og Grønlands Geologiske
Undersøgelse Rapport 2001/99, 47 pp.
Pulvertaft, T.C.R. 1979: Stenfald og fjeldskred ved Niakornat,
Umanak Kommune, Vest-Grønland. Rapport til Umanak
Kommune, 12 pp. Unpublished report, Grønlands Geologiske
Undersøgelse, København (in archives of Geological Survey
of Denmark and Greenland, Copenhagen, Denmark).
Rosenkrantz, A. 1967: Bjørnefælden ved Nûgssuaq og nord-
boernes Eysunes i geologisk belysning. Tidsskriftet Grønland
12, 377­384. Charlottenlund, Danmark: Det grønlandske
Selskab.
Steenstrup, K.J.V. 1900: Beretning om en Undersøgelsesrejse til
Øen Disko i Sommeren 1898. Meddelelser om Grønland 24(3),
249­306.
83
Authors' addresses
S.A.S.P., L.M.L., T.D.-J., H.F.J., T.N., F.v.P.-H. & W.W., Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copenhagen
K, Denmark. E-mail: sasp@geus.dk
G.K.P., Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark.
A.K.P., Geological Museum, University of Copenhagen, Øster Voldgade 5­7, DK-1350 Copenhagen K, Denmark.
GSB191-Indhold 13/12/02 11:32 Side 83
Review of Greenland Activities 2001, pp 73-83