Geomorphology 26 Ž1998. 139–150
Geoecology and mass movement in the Manaslu-Ganesh and
Langtang-Jugal Himals, Nepal
R.A. Marston
b
a,)
, M.M. Miller b, L.P. Devkota
c
a
UniÕersity of Wyoming, Laramie, WY 82071, USA
Foundation for Glacier and EnÕironmental Research and UniÕersity of Idaho, Moscow, ID 83843, USA
c
Central Department of Meteorology, P.O. Box 127, Lalitpur, Nepal
Received 15 May 1996; revised 10 December 1997; accepted 1 March 1998
Abstract
This study describes and explains the spatial distribution of mass movement in the central Nepal Himalaya. Judgments
were formulated on the origin and rates of mass movement using field evidence, topographic maps, geologic maps, and
SPOT imagery. Mass movement scars were mapped in the field during a 240-km traverse of the Langtang-Jugal Himal and a
300-km traverse of the Manaslu-Ganesh Himal. Chi-square analyses revealed that the frequency of slope failures varies with
slope aspect, and position aboverbelow the Main Central Thrust ŽMCT.. Human disturbance did not account for a
statistically significant increase in mass movement, except in sites occupied by mid-slope roads and where excessively steep
slopes, marginal for agriculture or grazing, have been deforested. q 1998 Elsevier Science B.V. All rights reserved.
Keywords: geoecology; mass movement; Nepal; Main Central Thrust
1. Introduction
The geomorphic development of hillslopes in the
middle mountain, high mountain, and high Himalaya
physiographic regions of central Nepal has been
dominated by ancient and modern mass movement
coupled with dramatic incision by major rivers.
Travel through these regions is difficult without being impressed by the extent of mass movement.
Considerable debate rages among Himalayan scientists over the relative effect of human activities on
the magnitude and frequency of mass movement.
)
Corresponding author.
Excellent summaries of this debate are provided by
Carson Ž1985., Ives and Messerli Ž1989., Bruijnzeel
and Bremmer Ž1989. and in papers presented at the
1995 Workshop on Landslide Hazard Management
and Control in the Hindu Kush Himalaya in Kathmandu Že.g., Mool, 1995.. Very few engineering
studies of slope stability have been reported, primarily because of the difficulty in acquiring adequate
field data of the detail needed. In any case, engineering slope stability analyses often do not lead to a
general understanding of controls on mass movement
because of the difficulty in extrapolating from one
field site to the next without an equivalent amount of
detailed field data. Thus, we remain at a stage where
0169-555Xr98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 9 - 5 5 5 X Ž 9 8 . 0 0 0 5 5 - 5
140
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
inventories of mass movement are useful ŽMool,
1995..
The consequences of mass movement in the Himalaya are well documented. One set of consequences involves the loss of productive land for
forestry, cultivation, or range use. This dimension of
the hazard from mass movement is acknowledged by
mountain villagers as a serious problem, but in many
cases, supernatural causes are blamed. Johnson et al.
Ž1982. have described the range of responses by
mountain villagers to mass movement events, including preventive maintenance and repair for reuse at a
lower land use intensity. A second consequence of
mass movement involves the sedimentation impact
on stream channels ŽFig. 1.. Aggradation promotes
stream undercutting of slopes which triggers yet
more mass movement and aggradation—a positive
feedback not easily remedied by engineering works.
In some cases, temporary dams are created across
channels that cause catastrophic flooding upon failure in the manner described by Costa and Schuster
Ž1988.. Damage to hydropower and irrigation projects is a major impact of this sedimentation ŽHMGr
WESC, 1987.. Mass movement, and the channel
shifting and flooding related to it, are considered by
some to be another manifestation of the destruction
of life support systems on the Ganges River Plain by
the actions of subsistence farmers in the mountains
ŽIves and Messerli, 1989..
This study evaluates the spatial distribution of
mass movement in the central Nepal Himalaya. An
attempt was made to formulate some judgments on
the origin and rates of mass movement. On this later
point, it is dangerous to try isolating triggering
mechanisms. The processes that trigger mass movement operate on different time scales. For instance,
contrast the long-term effect of progressive weathering along the soil–bedrock contact with the seasonal
effect of fluctuating water tables and the sporadic
effect of earthquakes. Moreover, it is difficult to
identify the triggering mechanism of mass movement
events from hundreds or thousands of years ago. As
Fig. 1. Streamside mass movement and associated channel aggradation common to the central Nepal Himalaya. Located along unnamed
tributary to the Malemchi Khola near Talamarang.
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
Fig. 2. Route through the Langtang-Jugal Himal and Manaslu-Ganesh Himal. All scars from mass movement within the viewshed along this route were surveyed.
141
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
142
an alternative, it is more useful to focus on the
intrinsic and more static elements of those landscapes prone to mass movement.
2. Methods
Scars from mass movements were mapped in the
field during a 240-km traverse of the Langtang-Jugal
Himal and a 300-km traverse of the 1987 ManasluGanesh Himal ŽFig. 2.. Each scar from mass movement was mapped within the ‘ viewshed’ as long as
the location and degree of vegetation disturbance
could be evaluated. The surveys included portions of
the middle mountain, high mountain and high Himalaya physiographic regions as described in
HMGrWESC Ž1987. and in the work of Ives and
Messerli Ž1989.. The middle mountain and high
mountain physiographic regions are divided by the
Main Central Thrust ŽMCT.. The MCT is a major
lithologic, metamorphic, and structural discontinuity.
Below the MCT occur low grade, argillaceous and
calcareous metasediments. Above the MCT occur
high grade arenaceous metasediments, but with an
inverse metamorphic gradient. The MCT is the only
indisputable thrust fault within the eastern Himalaya
ŽSchelling, 1987.. In the field, kyanite is often found
immediately above the MCT, and serves as an aid to
mapping.
Table 1
Frequency of 272 recorded slope failures by slope aspect, position aboverbelow the MCT, and whether natural or human-caused
Aspect
Aboverbelow
MCT
Cause
Observed no.
Žfor chi-square calculations.
Expected no.
Žfor chi-square calculations.
N 14.9%
below 4.1%
natural 3.0%
human 1.1%
natural 7.0%
human 3.8%
natural 5.4%
human 4.1%
natural 3.5%
human 0.2%
natural 1.5%
human 1.5%
natural 5.2%
human 2.0%
natural 1.1%
human 4.5%
natural 5.8%
human 4.0%
natural 1.8%
human 1.8%
natural 8.0%
human 2.3%
natural 4.8%
human 2.0%
natural 2.6%
human 1.1%
natural 1.6%
human 0.9%
natural 5.3%
human 1.4%
natural 2.4%
human 3.2%
natural 6.3%
human 3.5%
5
1
3
1
8
2
5
2
16
3
11
6
17
3
10
3
31
11
25
9
15
4
10
6
13
2
9
5
17
5
9
5
8
3
19
10
14
4
9
1
5
4
14
5
3
12
16
11
5
5
22
6
4
5
7
3
4
3
14
4
6
9
17
10
above 10.8%
NE 10.5%
below 6.8%
above 3.7%
E 10.2%
below 3.0%
above 7.2%
SE 15.4%
below 5.6%
above 9.8%
S 13.9%
below 3.6%
above 10.3%
SW 10.5%
below 6.8%
above 3.7%
W 9.2%
below 2.5%
above 6.7%
NW 15.4%
below 5.6%
above 9.8%
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
143
Fig. 3. Debris slide of 75,000 m3 in undisturbed forest of the Middle Mountains below the Ganesh Himal. For scale, the waterfall at top
right is 80 m high.
Several characteristics at each site were recorded
for a total of 272 scars from mass movement ŽTable
1.. Slope aspect was recorded, but slope gradient
could not be measured accurately from a distance or
from topographic maps. The lithologic–structural
setting of each scar was noted, using direct observation where the scar could be inspected. Otherwise,
reference was made to the 1:200,000 scale French
geologic map ŽColchen et al., 1980. of the Annapurna–Manaslu-Ganesh region or the 1:1,000,000
scale map produced by the Nepal Department of
Mines and Geology ŽHMGrDMG, 1980.. More detailed geologic maps of this region are not available.
In terms of the origin of the scar, a simple classification as ‘natural’ or ‘human-caused’ was used, following the simple criteria of Laban Ž1979.. If the
scar was located in a forest or was undercut by a
river, it was placed in the natural class ŽFig. 3.. If the
scar was located in cleared or cultivated land, or was
directly attributed to road construction, it was classed
as human-caused. This classification has some inherent ambiguity because natural mass movement can
occur in disturbed areas, thereby possibly overstating
the extent of mass movement triggered by humans.
In the final analysis, however, results were not affected by this procedure. Meteorological variables
were ignored in this study; Caine and Mool Ž1982.
have noted that this is not a limiting factor. Poor
seismicity records prevented analysis of earthquake
activity as a control on the spatial distribution of
mass movement, although in some portions of the
Himalaya, this factor can be an important driving
force for mass movement ŽSarkar et al., 1995..
3. Results
3.1. The form and origin of scars from mass moÕement by physiographic region
The middle mountain physiographic region of the
Manaslu-Ganesh Himal is situated below the MCT,
so the bedrock is dominated by phyllites, quartzites,
and garnet mica schist. The combination of these
lithologic units with the hot–wet climate and dense
vegetation has led to deep weathering and a rounding
of slope breaks by soil creep. Slope gradients range
144
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
Fig. 4. Debris slide of 225,000 m3 on terraced hillslopes of Middle Mountains near the village of Linju. Note the breached dam and
lacustrine deposits behind the former dam.
Fig. 5. Debris slide of 900,000 m 3 in the Middle Mountains of the Helambu District, Langtang-Jugal Himal. The scale of this mass
movement scar prohibits any rehabilitation.
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
up to 308. Stream undercutting is locally important as
a trigger to debris slides along major rivers such as
the Marsyandi and Buri Gandaki and their tributaries
ŽFigs. 4 and 5..
The high mountain physiographic region of the
Manaslu-Ganesh Himal is underlain by a medium- to
coarse-textured augen gneiss. This lithology is structurally more competent, providing the framework for
the prevailing cuestaform topography. The cuestas
dip to the north, leading some to speculate that the
topography is the surface expression of thrust sheets.
Slope breaks are sharp between the dip slope and
145
scarp slopes. Irrigation drainage from the dip slopes,
sometimes discharged onto the scarp slope, triggers
debris slides ŽFig. 6.. The angle of dip increases
from 108 to more than 308 as one moves from south
to north. At higher elevations, the dip slopes become
excessively steep, dipping up to 458, and soils are
more shallow, with abundant evidence of ancient and
modern debris slides. Deforestation does accelerate
mass movement on steep slopes which are marginal
for agriculture ŽBishop, 1990..
The middle mountain and high mountain physiographic regions of the Langtang-Jugal Himal are
Fig. 6. Debris slide of 6000 m 3 in High Mountains of Manaslu-Ganesh Himal. This slide was caused by discharge of excess irrigation
drainage from terraced fields onto a scarp slope in the loess mantled, cuestaform slopes.
146
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
the largest slide in the world in crystalline rock has
been reported by Heuberger et al. Ž1984. and
Schramm et al., 1998, this issue. An estimated mass
of 10 km3 was displaced along a fault plane. The
sliding surface generated fused crystals. Quaternary
age glaciers in Langtang have removed or buried 60
to 75% of the deposits from this event.
3.2. Statistical analyses of scars from mass moÕement
The chi-square statistical procedure was used to
test several hypotheses regarding the spatial distribution of the 272 scars from mass movement. In each
test, the division between classes was normalized by
the percent of the study area sampled that occurred
in each class, a key procedure that was not followed
in all past studies. The first hypothesis could be
stated as follows.
H o : No difference in the frequency of mass movement exists among slope aspects.
Fig. 7. Debris slide of 5000 m3 in loess mantled slopes on ridge
above Trisuli River. Note Hindu chortens along ridgetop trail in
this Middle Mountain region of the Langtang-Jugal Himal.
mantled with loess probably derived from the Tibetan Plateau. Deep-seated slides and slumps are the
dominant form of mass movement in undisturbed
situations. Deforestation and poor control of terrace
drainage are more important here in triggering mass
movement than in the Manaslu-Ganesh Himal ŽFig.
7..
In the high Himalaya physiographic region of the
Manaslu-Ganesh and Langtang-Jugal Himals, frost
action generates huge talus cones and felsenmeer
Žsee Watanabe et al., 1998, this issue., especially
along fractures. Slopes at elevations above 3000 m,
oversteepened by glaciation, may have near vertical
slopes with local relief in excess of 2000 m. A
sheeting structure because of multiple joint sets was
identified in gneissic and granitic bedrock which
may contribute to massive block slides. What may be
Fig. 8 illustrates the difference between the observed frequency of mass movement and the ‘expected’ frequency of mass movement Ži.e., the frequency if mass movement was distributed between
Fig. 8. Expected and observed frequency of mass movement by
slope aspect.
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
slopes of different aspects proportional to area sampled in each aspect.. The data in Table 1 reveal that
mass movement on south-facing aspects was more
frequent than expected. This aspect is on the windward side of summer monsoon storms and receives
the most direct solar insolation. Therefore, soils may
be subject to numerous wet–dry cycles which can
contribute to mass movement. In addition, abandoned land on south-facing slopes is not as quick to
revegetate. The calculated chi-square value was
71.78, greater than the critical chi-square value of
24.32 for df s 7 Ži.e., eight different slope aspects. at
p - 0.001. Therefore, we rejected the first hypothesis and concluded that mass movement did vary with
slope aspect.
The second hypothesis regarding the spatial distribution of mass movement can be stated as follows.
H o : No difference in the frequency of mass movement exists above and below the MCT.
Fig. 9 illustrates that mass movement was more
frequent than expected below the MCT and less
frequent than expected above the MCT. The deeply
weathered gneiss above the MCT appears to be more
susceptible to piping and gullying than to mass
movement, confirming the findings of Brunsden et
al. Ž1981. from studies in eastern Nepal. The calculated chi-square value was 39.06, greater than the
critical chi-square value of 10.83 for df s 1 at p 0.001. Therefore, we rejected the second hypothesis
and concluded that mass wasting did vary with position above and below the MCT. No significant difference in mass movements could be discerned between the phyllites, shales, and schists below the
MCT.
Fig. 9. Expected and observed frequency of mass movement by
position aboverbelow the MCT.
147
Fig. 10. Expected and observed frequency of mass movement in
landscapes in a ‘natural’ condition and in landscapes disturbed by
human activities.
The third hypothesis can be stated as follows.
H o : No difference in the frequency of mass movements exists between disturbed and undisturbed
landscapes.
Fig. 10 illustrates that mass movement was more
frequent than expected in undisturbed Ž‘natural’. areas and less frequent than expected in disturbed
Ž‘human-caused’. areas. The calculated chi-square
value was 10.99, greater than the critical chi-square
value of 10.83 for df s 1 at p - 0.001. Therefore,
we rejected the third hypothesis and concluded that
mass movement did vary with the degree of disturbance, but opposite to the trend often reported for
other regions of the world. Does this finding mean
the vegetation is unimportant? It is necessary to
distinguish between shallow and deep-seated forms
of mass movement. On unvegetated slopes, mass
movements are smaller and more shallow. Larger,
deeper slides occur independent of vegetation cover
ŽFig. 11.. Moreover, human activities do account for
a disproportionate share of mass movement in some
settings, poor road construction, and trail disruption
of slopes being the most notable ŽFig. 12.. Clearcutting or poor drainage from terraced fields onto
loess-derived soils or steep slopes with shallow soils
Žespecially in the cuestaform topography of the high
Himalaya. also leads to accelerated mass movement.
Nevertheless, these data help refute the broad assumption that human activities greatly increase sediment production from mountain regions of Nepal, a
conclusion also reached by Ives and Messerli Ž1989.
and Stevens Ž1993.. The notion that deforestation
accelerates mass movement can be exemplified by
148
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
Fig. 11. Deep-seated slump of 2,000,000 m3 west of Trisuli Bazar in the Middle Mountains.
Fig. 12. Mass movement triggered by mid-slope road construction, near village of Tarkughat, Middle Mountains, Manaslu Himal.
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
studies of the effect of clearcutting around the Pacific Rim reviewed by Sidle et al. Ž1985.. They
found that long-term rates of mass movement in
clearcuts were 7.8 times greater than in forested
areas. Also, mass movement from individual storm
events was 17.1 times more frequent in clearcuts
than in forested terrain. The combination of terrain
variables in the Himalaya, however, are not found
elsewhere in the world. Our observations were that
terraces can serve to stabilize slopes, especially with
‘kari’ type terraces that include a berm on the outside edge of the terrace Ža bund. to control downslope water movement. This finding confirms the
observations of Kienholz et al. Ž1984..
149
Dr. David Wilson, Mission Director of US AID in
Nepal; Lew McFarlane, Charge d’Affairs in the
American Embassy in Kathmandu; and Major Robin
Marston and Dr. Lute Jerstad of Mountain Travel
Nepal. Support for the expedition was provided by
the Foundation for Glacier and Environmental Research, Pacific Science Center, Seattle, WA; the
University of Idaho via the Glaciological Institute
and American–Nepal Education Foundation; and the
Wyoming Water Resources Center and College of
Arts and Sciences at the University of Wyoming.
This manuscript was improved by the suggestions of
John R. Giardino ŽTexas A & M University., Dr.
Jean-Paul Bravard ŽUniversite´ Paris IV—Sorbonne.,
Dr. John F. Shroder, Jr., and anonymous reviewers.
4. Conclusions
References
This study has identified a few of the key terrain
and land use variables that can explain the spatial
distribution of scars from mass movements in the
central Nepal Himalaya. Some indication has been
provided of just how dominant mass movement is as
a modern geomorphic process in the evolution of
hillslopes in the region. The study demonstrates that
human activities do not account for a disproportionate share of mass movement, contrary to a large
number of references in the scientific literature and
in the popular media linking deforestation with mass
movement. Deforestation is occurring, although the
style and extent varies from one region of Nepal to
the next. At the same time, devastating mass movement is occurring, but the great leap in logic linking
these two phenomena cannot be supported by the
data in this study area.
Acknowledgements
Assisting in the liaison with the Nepal government and providing other help and advice scientifically and logistically were: Dr. Bidhur Upadhyay,
Head of the Department of Meteorology, Tribhuvan
University; Professor Suresh Chalise, Dean of Science at Tribhuvan University and currently a consultant to ICIMOD; Shailesh Chandra Singh, Executive
Director of the National Council of Science and
Technology in Nepal; Dr. Allen Bassett, geologist;
Bishop, B.C., 1990. Karnali under stress: livelihood strategies and
seasonal rhythms in a changing Nepal Himalaya. Geography
Research Paper Nos. 228–229. University of Chicago,
Chicago, IL, 460 pp.
Bruijnzeel, L.A., Bremmer, C.N., 1989. Highland–lowland interactions in the Ganges-Brahmaputra River basin: a review of
published literature. ICIMOD Occasional Paper No. 11, 136
pp.
Brunsden, D., Jones, D.K., Martin, R.P., Doornkamp, J.C., 1981.
The geomorphological character of part of the low Himalaya
of eastern Nepal. Z. Geomorphol. Suppl. Bd. 37, 25–72.
Caine, N., Mool, P.K., 1982. Landslides in the Kolpu Khola
drainage, middle mountains, Nepal. Mountain Res. Dev. 2,
157–173.
Carson, B., 1985. Erosion and sedimentation processes in the
Nepalese Himalaya. ICIMOD Occasional Paper No. 1. International Centre for Integrated Mountain Development: Kathmandu, Nepal, 39 pp.
Colchen, M., Le Fort, P., Pecher, A., 1980. Geologic research in
the Nepal’s Himalaya: Annapurna–Manaslu-Ganesh Himal.
CNRS, Anatole, France, 136 pp.
Costa, J.E., Schuster, R.L., 1988. The formation and failure of
natural dams. Geol. Soc. Am. Bull. 100, 1054–1068.
Heuberger, H., Masch, L., Preuss, E., Schrocker, A., 1984. Quaternary landslides and rock fusion in central Nepal and in the
Tyrolean Alps. Mountain Res. Dev. 4, 345–362.
HMGrDMG, 1980. Geological Map of Nepal. Department of
Mines and Geology, Kathmandu, Nepal, 1:1,000,000.
HMGrWESC, 1987. Erosion and sedimentation in the Nepal
Himalaya. Report No. 4r3r010587r1r1, His Majesty’s Government of Nepal, Water and Energy Commission Secretariat,
Ministry of Water Resources, Kathmandu, Nepal.
Ives, J.D., Messerli, B., 1989. The Himalayan Dilemma: Reconciling Development and Conservation. The United Nations University and Routledge, London, 295 pp.
150
R.A. Marston et al.r Geomorphology 26 (1998) 139–150
Johnson, K., Olson, E.A., Manandhar, S., 1982. Environmental
knowledge and response to natural hazards in mountainous
Nepal. Mountain Res. Dev. 2, 175–188.
Kienholz, H., Hafner, H., Schneider, G., 1984. Stability, instability, and conditional stability mountain ecosystem concepts
based on a field survey of the Kakani area in the middle hills
of Nepal. Mountain Res. Dev. 4, 55–62.
Laban, P., 1979. Landslide occurrence in Nepal. Phewa Tal
Project Report No. SPr13. Integrated Watershed Management
Project, ICIMOD, Kathmandu, Nepal.
Mool, P.K., 1995. GIS and remote sensing application in slope
instability study. Paper presented at the Regional Workshop on
Landslide Hazard Management and Control in the Hindu Kush
Himalayas, 12–14 July, Kathmandu, Nepal.
Sarkar, S., Kanungo, D.P., Mehrotra, G.S., 1995. Landslide hazard zonation: a case study in Garhwal Himalaya, India. Mountain Res. Dev. 15, 301–309.
Schelling, D., 1987. The geology of the Rolwaling-Lapchi Kang
Himalayas, east-central Nepal: preliminary findings. J. Nepal
Geol. Soc. 4, 1–19.
Schramm, J.-M., Weidinger, J.T., Ibetsberger, H.J., 1998. Petrologic and structural controls on geomorphology of prehistoric
Tsergo Ri slope failure, Langtang Himal, Nepal. Geomorphology 26, 107–121, this issue.
Sidle, R.C., Pearce, A.J., O’Loughlin, C.L., 1985. Hillslope stability and land use. Water Resour. Monogr. Ser. 11. American
Geophysical Union, Washington, DC, 140 pp.
Stevens, S.F. 1993. Claiming the High Ground: Sherpas, Subsistence, and Environmental Change in the Highest Himalaya.
University of California Press, Berkeley, CA, 537 pp.
Watanabe, T., Dali, L., Shiraiwa, T., 1998. Slope denudation and
the supply of debris to cones in Langtang Himal, Central
Nepal Himalaya. Geomorphology 26, 185–197, this issue.