Ground-based LIDAR data on permafrost-related rockfall activity in the Mont Blanc Massif

Abstract

It is hypothesized that climatic warming since 1980 increases rock wall instability in high mountains due to permafrost degradation. This is supported by the observation of ice in several rockfall scars. Due to a lack of systematic observations, magnitude and frequency of high mountain rock failures remain poorly known. As part of the French-Italian PERMAdataROC project, we apply ground-based LIDAR to monitor instability on representative permafrost-affected rock walls (3000 to 4650 m a.s.l.) in the Mont Blanc massif. Initial results indicate that rockfall activity probably relates to different conditions at the 3 reported sites. The Piliers de Frêney and Grand Pilier d’Angle, both above 4000 m, are virtually stable (0 m³ of rockfalls) and indicate conservation of permafrost at high altitudes even on south-facing rock walls. With a probably critical state of permafrost, Tour Ronde E-Face and Arête Freshfield NE-Face (3460–3792 m) released ca. 1000 m³ of rockfall from 2005–2007. On Les Drus (2700–3700 m a.s.l.), 560 m³ of rockfalls were observed; we argue that these occur due to slope readjustment to the 2005 rock avalanche and are not directly linked to permafrost degradation

Full text

Ninth International
Conference on Permafrost
Edited by Douglas L. Kane and Kenneth M. Hinkel
Volume 1
Proceedings of the Ninth International Conference on Permafrost
University of Alaska Fairbanks
June 29–July 3, 2008
Institute of Northern Engineering
University of Alaska Fairbanks
2008

Ninth International Conference on Permafrost
Edited by Douglas L. Kane and Kenneth M. Hinkel
© 2008 Institute of Northern Engineering
University of Alaska Fairbanks
All rights reserved.
Printed in the United States of America
Elmer E. Rasmuson Library Cataloging in Publication Data
International Conference on Permafrost (9th : 2008 : Fairbanks, Alaska)
Ninth International Conference on Permafrost /
edited by Douglas L. Kane and Kenneth M. Hinkel.
— Fairbanks, Alaska : Institute of Northern Engineering,
University of Alaska Fairbanks, 2008.
2 v., : ill., maps ; cm.
Includes bibliographical references and index.
June 29–July 3, 2008
1. Permafrost–Congresses. 2. Frozen ground–Congresses.
I. Title. II. Kane, Douglas L. II. Hinkel, Kenneth M.
GB641.I6 2008
ISBN 978-0-9800179-2-2 (v.1)
ISBN 978-0-9800179-3-9 (v.2)
Cover Photo: Low-Centered Polygons, North Slope, Alaska
© 2007 Steven Kazlowski / AlaskaStock.com
Production Editors: Thomas Alton and Fran Pedersen
UAF is an Afrmative Action / Equal Opportunity employer and educational institution.
Kane, D.L. & Hinkel, K.M. (eds). 2008. Ninth International Conference on Permafrost. Institute of Northern Engineering,
University of Alaska Fairbanks (2 Vols.), 2140 pp.

349
Ground-Based LIDAR Data on Permafrost-Related Rockfall Activity
in the Mont Blanc Massif
Philip Deline, Stéphane Jaillet, Antoine Rabatel, Ludovic Ravanel
EDYTEM Lab, Université de Savoie, CNRS, Le Bourget-du-Lac, France
Abstract
It is hypothesized that climatic warming since 1980 increases rock wall instability in high mountains due to permafrost
degradation. This is supported by the observation of ice in several rockfall scars. Due to a lack of systematic
observations, magnitude and frequency of high mountain rock failures remain poorly known. As part of the French-
Italian PERMAdataROC project, we apply ground-based LIDAR to monitor instability on representative permafrost-
affected rock walls (3000 to 4650 m a.s.l.) in the Mont Blanc massif. Initial results indicate that rockfall activity
probably relates to different conditions at the 3 reported sites. The Piliers de Frêney and Grand Pilier d’Angle, both
above 4000 m, are virtually stable (0 m³ of rockfalls) and indicate conservation of permafrost at high altitudes even
on south-facing rock walls. With a probably critical state of permafrost, Tour Ronde E-Face and Arête Fresheld NE-
Face (3460–3792 m) released ca. 1000 m³ of rockfall from 2005–2007. On Les Drus (2700–3700 m a.s.l.), 560 m³ of
rockfalls were observed; we argue that these occur due to slope readjustment to the 2005 rock avalanche and are not
directly linked to permafrost degradation.
Keywords: LIDAR; Mont Blanc massif; PERMAdataROC project; permafrost; rockfall; terrestrial laser scanning.
Introduction
Recently, large rock and rock/ice avalanches have occurred
in high mountain areas worldwide (e.g., Kolka-Karmadon,
Caucasus 2002, Huggel et al. 2005). In the Alps, Brenva
Glacier (1997), Punta Thurwieser (2004), the west face of
Les Drus (2005), and the east face of Monte Rosa (2006,
2007) are the most recent examples (Deline 2001, Noetzli
et al. 2003, Fischer et al. 2006, Ravanel 2006). In addition,
innumerable smaller rockfalls detached from steep rock
walls during the hot summer of 2003 (Gruber et al. 2004).
The hypothesis that the increase of high mountain rock
wall instability relates to permafrost changes gains force
(Haeberli et al. 1997, Gruber & Haeberli 2007) from the
fact that (1) ice was observed in many starting zones; (2) the
mean annual air temperature in the Alps has increased more
than 1°C during the 20th Century; and (3) the warming trend
has accelerated since 1980.
However, frequency and volume of instability events in
high mountains are still poorly known because of the lack of
systematic observations, and ongoing permafrost changes in
rock walls remain poorly understood due to the difculties
in carrying on in situ measurements. So far, permafrost
studies are mainly based on modeling, with a few existing
instrumented sites.
The PERMAdataROC project
The PERMAdataROC project aims at studying the relation
between permafrost degradation and high mountain rock wall
instability in two west Alpine areas, the Mont Blanc massif
and the Matterhorn, based on the interface of three research
assignments (Ravanel & Deline 2006).
The rst assignment deals with the collection, maintenance
and analysis of recent rockfall/rock avalanches in the Mont
Blanc massif in a data base, based on (1) systematic survey
of slope instability events (localisation, exposition, time,
meteorological conditions, snow conditions, estimated
volume, path) carried out by local, trained people (mountain
guides, rescue people, hut keepers) in collaboration with
the researchers; (2) digitalisation of the events in a GIS;
and (3) analysis of the topographical, geological and
climatic parameters of the affected rock walls. This data
base is complemented by past events that are documented
by newspapers, hut, and guide books, as well as previous
studies and guide interviews.
The second research assignment deals with measuring
and the thermal regime in rock walls. The instrumentation
(thermistors at 5, 10, 30, and 55 cm depth) and measurement
of relevant properties (albedo, irradiation, thermical
conductivity) of rock wall supercial layer and surface at
the selected study sites, combined with high altitude climatic
data recorded by a movable automatic weather station, will
contribute to validate the models of temperature distribution
and variations.
The third research assignment is provided by the
monitoring of the instability of representative rock walls,
by (1) frequently repeated surveys with long-range ground-
based LIDAR (LIght Detection And Ranging) and terrestrial
photogrammetry, and (2) the installation of a geophone
network in one of the study sites to determine the frequency
and volume of rockfalls, considering variable parameters
(altitude, aspect, slope angle, lithology, fracturing, shadow
effect, height drop).
Italian (CNR-IRPI Torino, ARPA Valle d’Aosta and FMS
Courmayeur) and French (EDYTEM Lab) partners are
involved in the PERMAdataROC project, with collaboration
with the GGG of University of Zürich.
Here we present initial results of the monitoring of the

350 Ni N t h iN t e r N a t i o N a l Co N f e r e N C e o N Pe r m a f r o s t
instability of rock walls at three sites in the Mont Blanc
massif since 2005.
Study Area
The Mont Blanc massif covers an area of approximately
350 km², 40% of which is glacierized. It reaches its highest
point at 4808 m a.s.l., but many of its granitic, fractured
faces, peaks, and aiguilles stand well above 3000 m. The
water divide between Rhône and Pô basins is a 35 km long
crest line which always exceeds 3300 m and is often higher
than 4000 m. Being one of the most active uplift spots in the
Alps (>1.5 mm.an-1), the massif is not only characterized by
very high, but also steep faces and rock walls.
Seven study sites from 3000 to 4650 m with different
aspects were selected in the Mont Blanc massif within the
PERMAdataROC project (Fig. 1): the west face of Les Drus,
a >70° rock wall between 2700 and 3700 m affected by a
series of rockfalls since 1950 with increasing magnitude until
June 2005, when collapse of the Pilier Bonatti generated a
rock avalanche of >250,000 m3 (Fig. 2) (Ravanel 2006); the
surveyed area at Les Drus is between 3000 and 3700 m; the
Piton Central of the Aiguille du Midi (3770–3842 m) which
towers above the cable car station, with all aspects; the east
face of the Tour Ronde, and the NE-facing Arête Fresh eld
which develops at the south (3460–3792 m), where rockfalls
have been active for several years; the west face of the
Aiguilles d’Entrèves (3490–3591 m); the Grand Flambeau
(3410–3561 m), close to the Helbronner cable car station, with
all aspects; the east-facing Piliers de Frêney (4000–4650 m)
and the south face of the Grand Pilier d’Angle (4050–4308
m), on the south side of Mont Blanc summit, and the NW
face of the Aiguille Blanche Nord de Peuterey (4000–4103
m). These high elevation sites are complemented by a low-
elevation control site (2200–2700 m) without permafrost, on
the SW-facing side of the Vallon du Miage.
Methods
Since June 2005, ground-based LIDAR measurements are
realised seasonally (summer/autumn) or annually at the eight
sites in the Mont Blanc massif, using helicopter or cable cars
Figure 1. Location map of GB-LIDAR surveyed sites in the Mont
Blanc massif within the framework of the PERMAdataROC project.
1: Drus; 2: Aiguille du Midi; 3: Tour Ronde-Arête Fresh eld; 4:
Aiguilles d’Entrèves; 5: Grand Flambeau; 6: Piliers du Frêney-
Grand Pilier d’Angle; 7: Aiguilles Blanches de Peuterey; 8: Vallon
du Miage. Largest glaciers are highlighted. This paper presents
initial results from sites 1, 3, and 6.
Figure 2. Upper part of the west face of Les Drus. The upper part
of the 2005 rock avalanche scar is delimited by the white line. To
survey the face, the ground-based LIDAR is set up on Les Flammes
de Pierre (crest on bottom right).

de l i N e e t a l . 351
for access. Data are processed for calculating high-resolution
(centimeter-scale) triangulated irregular networks (TIN).
Volumetric changes, extracted on the rock faces by compar-
ing the successive TINs, represent the fallen rocks between
the measurement periods (Ravanel & Deline 2006).
LIDAR survey
LIDAR measurements are performed using an Optech
ILRIS 3D ground-based LIDAR. This laserscanner works at
distances of up to 800 m if surface reectivity and visibility
are good. The angle of view is 40°×40°, and the sampling rate
reaches a maximum frequency of 2000 points per second. At
a distance of 100 m, the laser beam diameter is about 30 mm
(perpendicular shot), and the accuracy on a at surface is
about 3–5 mm. The LIDAR point to point distances on the
rock walls we are surveying range between 61 mm and 246
mm (in 2006 and 2007), on the closest and farthest areas,
respectively.
Data processing
Data processing (Rabatel et al., submitted) is realised using
InnovMetric PolyWorks software, with (1) alignment of
individual point clouds using the IMAlign module: they
are merged with a rototranslation matrix into a unique
local reference system, after cleaning individual scans
from outliers (Fig.3); and (2) creation of the TIN using the
IMMerge module.
The computation of the fallen rock volume in a rock wall
between two successive eld work campaigns is achieved
with the PolyWorks IMInspect module, which compares
two point clouds and quanties the thickness changes. A
reference plan is built, and the volume between the surface
topography and this plan is computed for each date.
Error estimation
The total uncertainty can be estimated by the quadratic
sum of the different independent errors in the processing.
(1) LIDAR error is 3–5 mm at 100 m (manufacturer data).
(2) TIN is interpolated from existing points of the global
point cloud (set of 3D images). To merge it into a unied
polygonal mesh, most parameter values are calculated using
input point cloud values. Due to the average mesh used,
the TIN construction error is ca. 7 cm. (3) To be compared,
diachronous TINs have to be very overlapped. But because
of very large TINs, there is a TIN overlapping error, which
is 5 cm as measured by Polyworks. This yields an overall
uncertainty of 9 cm, which is reduced by directly comparing
the point clouds.
Mask effect
Masks result from (1) the topography of the rock wall
(roofs, ledges, corners, spurs); (2) the common scarcity of
sites to set up the LIDAR (e.g., for the west face of Les Drus,
there is only one possible on Les Flammes de Pierre: Fig.
2); and (3) the snow cover, whose extension differs each
year. Masks could represent an important part of the surface
surveyed, and appear as holes in the TINs. This is particularly
the case if there are no multiple viewing angles (Les Drus),
or if the snow-ice cover is important (Peuterey).
The Polyworks IMEdit module allows to reduce their
extension. The hole area is rst selected. A tool allows the
holes to ll automatically using irregular triangles. Only the
maximum distance between the vertices of a triangle has to
be specied. The longer this distance, the greater resulting
size of the plugged hole.
Results
We present initial results from three of the seven high-
elevation selected sites in the Mont Blanc massif: (1) the
west face of Les Drus; (2) the east face of the Tour Ronde
and the NW side of the Arête Fresheld; and (3) the Piliers
de Frêney and the south face of the Grand Pilier d’Angle
(Table 1).
West face of Les Drus
Comparison of October 2005 and October 2006 TINs
reveals a detachment from the 2005 rock avalanche scar of
height rock elements of a volume ≥1 m³: ve boulders are
≤6 m³, and three are larger. At about 3600 m a.s.l., a notch
of 29×10×1.8 m (426 m3) is present on the 2006 TIN; the
rocks reached the small debris-covered glacier of Les Drus,
at the foot of the west face. Lower on the rock wall, two
elements of 19 and 84 m3 also collapsed in this one-year
period. In total, 546 m3 of rock were released in the surveyed
area between October 2005 and October 2006.
The third survey, carried out at the end of September 2007,
shows reduced rockfall activity over the period extending
from October 2006 to September 2007: only one small
rockfall occurred (22 m3), out of the 2005 scar.
Figure 3. Point clouds of the west face of Les Drus derived from
LIDAR surveys (left: October 2005; right: October 2006). The
height of the 2005 rock avalanche scar here represented is 500 m.

352 Ni N t h iN t e r N a t i o N a l Co N f e r e N C e o N Pe r m a f r o s t
East face of La Tour Ronde-NE face of Arête Fresh eld
At the Tour Ronde, the evolution of the surface topography
of the east face between July 2005 and July 2006 shows
two main rockfall events, with a volume of 382 m3 (Fig.
4) and 154 m3 (main scar sizes are 17.5×7.8×4.3 m and
15.1×9.3×1.4 m, respectively). The total volume of these
two rockfall reaches 536 m3.
The comparison of LIDAR measurements for the second
period, between July 2006 and October 2006, shows no
signi cant change during the 2006 summer.
During the third, and last, period of survey, no change was
observed on the east face of Tour Ronde. On the other hand,
at least two rockfalls occurred on the NE side of the Arête
Fresh eld, involving a set of large boulders (total volume:
448 m3) which detached from a small area where bedrock is
highly dislocated.
Piliers de Frêney-south face of Grand Pilier d’Angle
Five successive LIDAR surveys performed since July
2005 display no change on the rock walls during a period of
27 months, including 3 summers. Mask effects are important,
due to the rough topography of the area (large pillars
separated by deep couloirs) and the unique site available to
set up the ground-based LIDAR. Thus, only a part of this
area is surveyed, but no signi cant rockfall was observed
over the 115,000 m2 it represents.
Discussion and Conclusions
Results indicate that rockfall activity probably relates to
different conditions at the three sites.
On the west face of Les Drus, the rockfalls which occurred
probably represent slope readjustment after the large rock
avalanche of 2005 and are not directly related to permafrost
degradation. For example, the detachment of 84 m3 between
October 2005 and October 2006 is due to the fall of an
individualized, hanging, and poorly-rooted slice. Moreover,
the small pieces of rock (<1 m3) identi ed between 2005
and 2006 were probably destabilized during the fall of the
largest element (426 m3). Lastly, no rockfall occurred into
the 2005 scar between 2006 and 2007: this suggests that the
mesoscale slope readjustment in the scar is now achieved.
Site Period of measurement
(d/m/y)
Surface of
surveyed area
by LIDAR
(m2)
Volume of
rockfalls (m3)
Mean rock wall retreat
rate in surveyed area
(mm a-1)
Extreme distance of
point to point on rock
wall
(mm)
Total 2 main
Drus 11/10/2005–11/10/2006 70,500 546 426 + 84 7.7 Not calculated
(W face) 12/10/2006–24/09/2007 22 - 0.3 71–208
Piliers de Frêney - 14/07/2005–10/10/2005
115,600
0 0 0.0 Not calculated
Grand Pilier 11/10/2005–30/06/2006 0 0 0.0 Not calculated
d’Angle (S face) 01/07/2006–13/10/2006 0 0 0.0 61–246
14/10/2006–12/10/2007 0 0 0.0 61–246
Tour Ronde (E fa- 13/07/2005–18/07/2006
67,400
536 382 + 154 8.4 Not calculated
ce) – Arête Fresh- 19/07/2006–12/10/2006 0 0 0.0 75–207
 eld (NE face) 13/10/2006–12/10/2007 448 448 6.6 75–207
Table 1. Rockfall data from LIDAR surveys (2005–2007).
Figure 4. East face of the Tour Ronde TIN (detail). The view
focuses on the area affected by the main rockfall between July 2005
(top box) and July 2006 (bottom box). Dimensions of the main scar
(visible in bottom box circle) are 17.5×7.8×4.3 m, with a volume of
382 m3 (image size: ca. 45 m × 45 m).

de l i N e e t a l . 353
But it is noteable that a new permafrost active layer forms in
the 2005 rock avalanche scar.
On the other hand, on the east face of La Tour Ronde and
on the NE side of the Arête Fresheld, the rockfalls (2005–
2006: 536 m3; 2006–2007: 448 m3) probably result from
permafrost degradation. This is suggested both by (1) the
high rock fall activity during recent years: the normal route
to the summit of Tour Ronde is not used when snow cover
has melted; and (2) the modeling of the surface temperature
of rock walls in alpine areas with similar weather conditions.
For instance, mean annual surface temperatures range
between -2°C and -4°C at Junfraujoch at 3500–3750 m a.s.l.
for E/NE aspect (Gruber et al. 2004: Fig. 2), without taking
into account the local topography of the rock faces.
Contrary to the two others sites, the area of the Piliers
de Frêney and the south face of the Grand Pilier d’Angle
show stable conditions. The results suggest that no or very
occasional degradation of the permafrost occurs at very
high altitude (>4000 m), including south-facing rock walls
like the Grand Pilier d’Angle. Several large grey scars at
the foot of the Piliers de Frêney and the nearby Piliers du
Brouillard and Mont Maudit, which clearly contrast to the
reddish surrounding granite, indicate that rockfalls and rock
avalanches have occurred at elevations greater than 4000
m; but preliminary studies (Böhlert et al. 2008) suggest
that these grey scars could have formed before 1,500 yr BP,
during previous Holocene cold periods.
Given that high-alpine rock walls are a poorly known
system, the introduced methodology shows a great potential
to reveal quantitative data on geomorphological processes in
permafrost-affected rock walls
Acknowledgments
The paper beneted from critical comments by the two
anonymous referees, one of them also improving the
English. Thanks to all partners of the PERMAdataROC
project. The European Union (FEDER), the Town Council of
Chamonix, and the Conseil Général de la Haute-Savoie are
acknowledged for funding. This paper is part of the Interreg
IIIA France-Italy # 196 PERMAdataROC project.
References
Deline, P. 2001. Recent Brenva rock avalanches (Valley of
Aosta): new chapter in an old story? Supplemento
Geograa Fisica e Dinamica Quaternaria 5: 55-63.
Böhlert, R., Gruber, S., Egli, M., Maisch, M., Brandová, D.,
Haeberli, W., Ivy-Ochs, S., Kubik, P.W. & Deline,
P. 2008. Comparison of exposures ages and spectral
properties of rock surfaces in steep, high Alpine rock
walls—a eld study at Aiguille du Midi, France.
Proceedings of the Ninth International Conference on
Permafrost, Fairbanks, Alaska, June 29–July3, 2008
(this proceedings).
Fischer, L., Kääb, A., Huggel, C. & Noetzli, J. 2006.
Geology, glacier retreat and permafrost degradation
as controlling factor of slope instabilities in a high-
mountain rock wall: the Monte Rosa east face. Natural
Hazards and Earth System Sciences 6: 761-772.
Gruber, S. & Haeberli, W. 2007. Permafrost in steep bedrock
slopes and its temperature-related destabilization
following climate change. Journal of Geophysical
Research 112:F02S18, doi: 10.1029/2006JF000547.
Gruber, S., Hoelzle, M. & Haeberli, W. 2004. Permafrost
thaw and destabilization of Alpine rock walls in the
hot summer of 2003. Geophysical Research Letters
31, L13504, doi: 10.1029/2004GL020051.
Haeberli, W., Wegmann, M. & Vonder Mühll, D. 1997.
Slope stability problems related to glacier shrinkage
and permafrost degradation in the Alps. Eclogae
Geologicae Helvetiae 90: 407-414.
Huggel, C. Zgraggen-Oswald, S., Haeberli, W., Kääb, A.,
Polkvoj, A., Galushkin, I. & Evans, S.G. 2006. The
2002 rock/ice avalanche at Kolka/Karmadon, Russian
Caucasus: assessment of extraordinary avalanche
formation and mobility, and application of QuickBird
satellite imagery. Natural Hazards and Earth System
Sciences 5: 173-187.
Noetzli, J., Hoelzle, M. & Haeberli, W. 2003. Mountain
permafrost and recent Alpine rock fall events: a
GIS-based approach to determine critical factors.
Proceedings of the 8thInternational Conference on
Permafrost, Zürich, Switzerland, August, 2003: 827-
832.
Rabatel, A., Deline, P., Jaillet, S. & Ravanel, L., submitted.
Rock falls in high-alpine rockwalls quantied by
terrestrial LIDAR measurements. A case study in the
Mont Blanc area. Geophysical Research Letters.
Ravanel, L. 2006. Contribution à l’étude des écroulements
dans les parois à permafrost de la haute montagne
alpine. L’exemple du Petit Dru (massif du Mont
Blanc) depuis la n du Petit Age Glaciaire). MSc
thesis, Université de Savoie, 116 pp.
Ravanel, L. & Deline, P. 2006. Nouvelles méthodes
d’étude de l’évolution des parois rocheuses de haute
montagne: application au cas des Drus. Actes du
Colloque Géologie et risques naturels: la gestion des
risques au Pays du Mont-Blanc, Sallanches, France,
November 18, 2006: 48-53.