LANDSLIDE SUSCEPTIBILITY ASSESSMENT IN

BRAHMAYANI WATERSHED, SINDHUPALCHOWK

A Dissertation Submitted to

CENTRAL DEPARTMENT OF ENVIRONMENTAL SCIENCE

Institute of Science and Technology
Tribhuvan University
Kirtipur, Kathmandu, Nepal

In Partial Fulfillment of the Requirements for the Award of Degree of M.Sc.

in Environmental Science

By
Sunil Acharya

August, 2024
 LANDSLIDE SUSCEPTIBILITY ASSESSMENT IN
BRAHMAYANI WATERSHED, SINDHUPALCHOWK

A Dissertation Submitted to

CENTRAL DEPARTMENT OF ENVIRONMENTAL SCIENCE

Institute of Science and Technology
Tribhuvan University
Kirtipur, Kathmandu, Nepal

In Partial Fulfillment of the Requirements for the Award of Degree of M.Sc.

in Environmental Science

By
Sunil Acharya
T.U. Registration No.: 5-2-37-2016-2079

T.U. Exam Roll No.: 1318

August, 2024
 DECLARATION

I hereby declare that the work presented in this dissertation is a genuine work done
originally by me and has not been submitted anywhere for the award of any degree. All the
sources of information have been specifically acknowledged by reference to author(s) or
institutions(s).

Sunil Acharya

August, 2024

ii
 Tel: 01-4332147
01-4332711
TRIBHUVAN UNIVERSITY
Kirtipur,
Central Department of Environmental Science
Kathmandu, Nepal

Date: August 2024

RECOMMENDATION
This is to certify that Mr. Sunil Acharya has completed this case dissertation work
entitled ―Landslide Susceptibility Assessment in Brahmayani Watershed,
Sindhupalchowk‖ as a partial fulfillment of the requirements of M.Sc. in Environment
Science under my supervision and guidance. To my knowledge, this research has not been
submitted for any other degree anywhere else.

I therefore, recommend the dissertation for acceptance and approval.

…………….

Supervisor
Subodh Dhakal, PhD
Associate Professor

Department of Geology

Tri Chandra Multiple Campus

Ghantaghar, Kathmandu

iii
 TRIBHUVAN UNIVERSITY Tel: 01-4332147
01-4332711
Central Department of Environmental Science
Kirtipur,
Kathmandu, Nepal

Date: August, 2024

LETTER OF APPROVAL

On the recommendation of supervisor ―Associate Professor Dr. Subodh Dhakal‖ this

dissertation submitted my “Mr. Sunil Acharya‖ entitled ―Landslide Susceptibility
Assessment in Brahmayani Watershed, Sindhupalchowk‖ has been approved for the
examination and submitted to the Tribhuvan University in partial fulfillment of the
requirements M.Sc. in Environmental Science.

………………

Prof. Udhab Raj Khadka,,PhD

Head of Department

Central Department of Environmental Science

Tribhuwan University, Nepal

iv
 TRIBHUVAN UNIVERSITY Tel: 01-4332147
01-4332711
Central Department of Environmental Science

Kirtipur,
Kathmandu, Nepal

Date: August ,2024

CERIFICATE OF ACCEPTANCE

This dissertation entitled ―Landslide Susceptibility Assessment in Brahmayani Watershed,

Sindhupalchowk‖ Submitted by ―Mr. Sunil Acharya‖ has been examined and accepted as
a partial fulfillment of the requirements of M.Sc. in Environmental Science.

Evaluation Committee

…………….……………….. …………………………..

Supervisor External Examiner

Subodh Dhakal,PhD
Associate Professor
Department of Geology
Trichandra Multiple Campus
Ghantaghar, kathmandu

…………….………………..
…………………………..
Internal Examiner
Prof. Udhab Raj Khadka, PhD
Head of Department
Central Department of Environmental Science

Tribhuwan University, Nepal

v
 ACKNOWLEDGEMENT

I would like to express heartfelt gratitude to my supervisor Associate Prof. Subodh Dhakal,
PhD, Central Department of Environment Science for his guidance, suggestions, comments
and encouragement throughout the undertaking of this research. This dissertation work
would not have been completed without his suggestions and supervision.

I wish to express my thanks to Head of the Department Prof. Dr. Udhab Raj Khadka and
entire CDES staff for their support. I would like to extent my special thanks to my friends
and family for their support and guidance during the laboratory work.

I would like to thank my family members for their belief and encouragement throughout
my study period. I would also like to thank my dear friends who helped me during my field
visit. .

Sunil Acharya
August , 2024

vi
 ABSTRACT
Landslide is defined as the downward movement of a mass rock, debris or earth along a
slope and denotes a connected complex process influenced by various causative factors
like geology, morphometric and hydrological factors. The area of interest is Brahmayani
Watershed within Higher Himalaya, where landslide inventory map is prepared due to the
presence of small to large landslides affecting ecology and society. Landslides are a
significant natural hazard, causing loss of life, damage to infrastructure, and economic
disruption. Therefore, it is important to generate landslide susceptibility map which is a
fundamental tool for disaster management activities in mountainous terrains of Himalaya.
In this paper, the main objective is assessing landslide susceptibility using: Frequency
Ratio (FR) Model and Analytical Hierarchy Process (AHP) model considering thirteen
causative factors. Five distinct classes: very low, low, moderate, high and very high
susceptibility map was produced In FR low class has the highest area with 209.46 sq.km,
very low susceptibility has only 21.59 sq.km, similarly moderate has 116.57 sq.km high
has 143.98sq.km and very high has 188.40 sq. km of the total area where as in AHP low
class has the highest area with 181.96 sq.km, very low susceptibility has only 20.60 sq.km,
moderate has an area of 134.85. Similarly high has 161.67sq.km, and very high has 180.92
sq. km from the total area. Landslide susceptibility mapping and validation from
Frequency Ratio and Analytic Hierarchy Process models are validated from the success
rate and predictive rate of ROC-AUC curve. The success rate was 86.5 % for Frequency
Ratio and 80.2 % for Analytic Hierarchy Process while predictive rate curve 81.2 % for
Frequency Ratio and 77.9 % for Analytic Hierarchy Process respectively. This study
shows the potential of the models to predict landslide susceptible areas and can be used for
different purposes by different stakeholders including engineers, planners, and decision-
makers for infrastructure.

Keywords: AHP Model, Brahmayani Watershed, FR Model, Landslide Mapping,

Landslide Susceptibility

vii
 सोध-सार

पहिरोलाई चट्टान, माटोको र ढलानको साथ तलहतरको गहतमा गइरिे को मासको रूपमा पररभाहित
गररएको छ र यसले हिहभन्न कारणात्मक कारकिरू जस्तै आधारहिला भूहिज्ञान (हलथोलोजी,
संरचना, मौसमको हडग्री), आकारको गुणिरू जस्तै ढलान कोण, पक्ष िा उचाइ, जलहिज्ञानका

कारकिरू जस्तै नदीिरूमा दू री, भूहिज्ञान, माटो, SPI र भूहमको उपयोग भूहमको कभरले प्रभाि
पारे को जहटल प्रहियालाई जनाउँ छ॰ चासोको क्षेत्र उच्च हिमालयहभत्रको ब्रह्मायणी जलाधार िो, जिाँ

साना दे खि ठूला पहिरोिरू छन् जसले पाररखिहतकी र समाजलाई रसर गछ॰ ब्रह्मायनी जलाधार क्षेत्र
उच्च भू भागमा रिे को छ र यिा भू कम्पीय रूपमा सहिय क्षे त्रमा पदछ, जसले यसलाई पहिरोको लाहग

संिेदनिील बनाउँ छ॰ त्यसकारण, हिमालयको पिाडी भू बनोटमा हिपद् व्यििापन गहतहिहधिरूको

लाहग आधारभूत उपकरणको रूपमा पहिरो संिेदनिीलता नक्सा उत्पादन गनु मित्त्वपूण छ॰ यस

पत्रमा, मुख्य उद्दे श्य भनेको १३ कारणात्मक कारकिरूको हिचार गदै : हिक्वेन्सी रे हसयो (FR) मोडे ल
र एनाहलहटकल िाइराकी प्रोसेस (AHP) मोडे ल प्रयोग गरे र क्षेत्रीय स्तरमा पहिरो संिेदनिीलता

नक्सा पहिचान र हिश्लेिण गनु िो॰ पाँ च फरक िगिरू: धेरै कम, कम, मध्यम, उच्च र धेरै उच्च
संिेदनिीलता नक्सा उत्पादन गररएको हथयो॰ FR मा कम िगमा २०९.४६ िग हकमीको सबैभन्दा

ठूलो क्षेत्र छ, धेरै कम संिेदनिीलतामा केिल २१.५९ िग हकमी छ, यस्तै मध्यममा ११६.५७ िग
हकमी उच्चमा १४३.९८ िग हकमी र धेरै उच्चमा १८८.४० िग हकमी क्षेत्र छ, जबहक AHP मा कम

िगमा १८१.९६ िग हकमी क्षेत्र छ॰ , धेरै कम संिेदनिीलता मात्र २०.६० िग हकमी छ, मध्यममा
१३४.८५ िग हकमी क्षेत्र छ॰ यस्तै गरी उच्चमा १६१.६७ िग हकमी क्षेत्र र धेरै उच्चमा १८०.९२ िग

हकमी क्षे त्र रिे को छ॰ हिक्वेन्सी रे हसयो र एनाहलहटकल िाइराकी प्रोसेस मोडे लिरूबाट प्राप्त पहिरो
संिेदनिीलता नक्सा र मान्यकरणलाई ROC-AUC ििको सफलता दर र भहिष्यिाणी दरबाट

मान्य गररएको हथयो॰ सफलता दर हिक्वेन्सी रे हसयोको लाहग ८६.५% र एनाहलहटकल िाइराकी
प्रोसेसको लाहग ८०.२% हथयो भने भहिष्यिाणी दर िि हिक्वेन्सी रे हसयोको लाहग ८१.२% र

एनाहलहटकल िाइराकी प्रोसेसको लाहग ७७.९% हथयो॰ यस रध्ययनले पहिरो सं िेदनिील क्षेत्रिरू
भहिष्यिाणी गन मोडे लिरूको सम्भािनालाई दे िाउँ छ र इखजजहनयरिरू, योजनाकारिरू र

हनणयकतािरूले पूिाधारका लाहग हिहभन्न उद्दे श्यका लाहग प्रयोग गन सक्ने दे िाउँ छ॰

मुख्य िब्दिरू: पाहिरो, पाहिरो संिेदनिीलता, FR मोडे ल, AHP मोडे ल, ब्रह्मायणी जलाधार

viii
 TABLE OF CONTENTS

DECLARATION ................................................................................................................. ii
RECOMMENDATION ..................................................................................................... iii
LETTER OF APPROVAL ................................................................................................ iv
CERIFICATE OF ACCEPTANCE .................................................................................. v
ACKNOWLEDGEMENT ................................................................................................. vi
ABSTRACT ....................................................................................................................... vii
सोध-सार .............................................................................................................................. viii
TABLE OF CONTENTS ................................................................................................... ix
LIST OF TABLES ............................................................................................................ xii
LIST OF FIGURES ......................................................................................................... xiii
LIST OF PHOTOGRAPHS ............................................................................................. xv
ABBREVIATIONS AND ACRONYMS ........................................................................ xvi
CHAPTER 1: INTRODUCTION ...................................................................................... 1

1.1 Background .......................................................................................................... 1

1.2 Rationale ............................................................................................................... 3

1.3 Research question ................................................................................................. 4

1.4 Research objectives: ............................................................................................. 4

CHAPTER 2: LITERATURE REVIEW .......................................................................... 6

2.1 Landslide hazard................................................................................................... 6

2.2 Landslide inventory Maps .................................................................................... 7

2.3 Landslide risk assessment .................................................................................... 8

2.4 Landslide Susceptibility Mapping ........................................................................ 9

2.5 Landslides in Nepal ............................................................................................ 10

CHAPTER 3: MATERIALS AND METHOD ............................................................... 12

3.1 Study area ........................................................................................................... 12

ix
 3.2 Methodology ...................................................................................................... 13

3.3 Data collection .................................................................................................... 14

3.3.1 Data Collection ............................................................................................ 14

3.3.2 Preparation of Landslide Inventory Map ..................................................... 16

3.4 Landslide Susceptibility Mapping ...................................................................... 16

3.5 Models used in study .......................................................................................... 29

3.5.1 Frequency Ratio .......................................................................................... 29

3.5.2 AHP Method................................................................................................ 30

3.5.3 Validation of model ..................................................................................... 32

CHAPTER 4: RESULT .................................................................................................... 34

4.1 Landslide Inventory Map ................................................................................... 34

4.2 Landslide Causative Factors ............................................................................... 34

4.3 Landslide susceptibility ...................................................................................... 43

4.3.1 Frequency ratio ............................................................................................ 43

4.3.2 Analytical hierarchy Process ....................................................................... 48

4.3 Comparison of AHP and FR model ................................................................... 53

4.4 Validation of landslide susceptibility Models .................................................... 53

CHAPTER 5: DISCUSSIONS ......................................................................................... 56

5.1 Landslide Inventory ............................................................................................ 56

5.2 Landslide Causative factors................................................................................ 56

5.3 Landslide susceptibility mapping and Validation .............................................. 59

CHAPTER 6: CONCLUSION AND RECOMMENDATIONS ................................... 61

6.1 Conclusion .......................................................................................................... 61

6.2. Recommendations ............................................................................................. 62

Reference ............................................................................................................................ 63
Appendices ............................................................................................................................ I

x
Appendix-1: Photograph ............................................................................................. I

Appendix-2: Pairwise comparisons of each causative factor ..................................... II

Appendix-3: Table showing landslide number and area of landslide ....................... V

xi
 LIST OF TABLES
Table 1: Classification of landslide ....................................................................................... 6
Table 2: Data source of causative factors ............................................................................ 14
Table 3: Random Consistency index (Saaty, 2000) ............................................................ 31
Table 4 : Scale of preferences between two parameters in AHP Saaty,(1977). .................. 32
Table 5: Frequency ratio table of causative factor .............................................................. 45
Table 6: Pairwise comparisons of causative factor ............................................................. 50
Table 7: Table showing Number, Eigen value, Consistency index, Random index and
consistency ratio .................................................................................................................. 51

xii
 LIST OF FIGURES

Figure 1: Location map showing the Brahmayani Watershed in Sindhupalchowk District of

Nepal ................................................................................................................................... 12
Figure 2: Research methodology flow chart ....................................................................... 13
Figure 3: Slope Map ............................................................................................................ 17
Figure 4: Aspect Map .......................................................................................................... 18
Figure 5: Curvature Map ..................................................................................................... 19
Figure 6: NDVI Map ........................................................................................................... 20
Figure 7: Distance from Road ............................................................................................. 21
Figure 8: Distance from Drainage ....................................................................................... 22
Figure 9: Geological Map ................................................................................................... 23
Figure 10: TWI Map ........................................................................................................... 24
Figure 11: Soil Map............................................................................................................. 25
Figure 12: Elevation Map .................................................................................................... 26
Figure 13: Rainfall Map ...................................................................................................... 27
Figure 14: Land cover Map ................................................................................................. 28
Figure 15: SPI Map ............................................................................................................. 29
Figure 16: Landslide Inventory Map ................................................................................... 34
Figure 17: Landslide distribution with relation to slope ..................................................... 35
Figure 18: Landslide distribution with relation to aspect .................................................... 36
Figure 19: Landslide distribution with relation to curvature ............................................... 36
Figure 20: Landslide distribution with relation to distance from road:............................... 37
Figure 21: Landslide distribution with relation to drainage distance ................................. 38
Figure 22: Landslide distribution with relation to elevation ............................................... 38
Figure 23: Landslide distribution with relation to land cover ............................................. 39
Figure 24: Landslide distribution with relation to NDVI .................................................... 40
Figure 25: Landslide distribution with relation to TWI ...................................................... 40
Figure 26: Landslide distribution with relation to geology................................................. 41
Figure 27: Landslide distribution with relation to rainfall .................................................. 42
Figure 28: Landslide distribution with relation to Soil ....................................................... 42

xiii
sFigure 29: Landslide distribution with respect to SPI ....................................................... 43
Figure 30: Landslide susceptibility map obtained from FR model ..................................... 47
Figure 31: Landslide susceptibility area designated based on FR model ........................... 47
Figure 32: Landslide susceptibility map obtained from AHP model .................................. 52
Figure 33: Landslide susceptibility area designated based on AHP model ....................... 52
Figure 34: Assessment of FR and AHP model performance showing true positive rate and
false positive rate (a) Success rate curve (b) Predictive curve rate ................................... 54
Figure 35: Training and testing points to carryout ROC-AUC curve in a) FR susceptibility
map b) AHP susceptibility map .......................................................................................... 55

xiv
 LIST OF PHOTOGRAPHS
Photographs 1: Landslide at Golche ....................................................................................... I
Photographs 2: Landslide at Selang ....................................................................................... I
Photographs 3: Landslide at Hagam ...................................................................................... I
Photographs 4: Landslide at Gumba ..................................................................................... I
Photographs 5 : Landslide at Baramche ................................................................................ II
Photographs 6: Landslide at Yanglakot .............................................................................. II

xv
 ABBREVIATIONS AND ACRONYMS

AHP Analytical Hierarchy Process

ALOSPALSAR: Advance Land Observing Satellite phased Array L-band Synthetic

Aperture Network

AUC: Area Under Curve

DEM: Digital Elevation Model

FP: False positive

FR: Frequency ratio

GIS: Geographic Information System

GoN: Government of Nepal

LSM: Landslide Susceptibility Mapping

MBT: Main Boundary Thrust

NDVI: Normalized Difference Vegetation Index

OSM: Open street Map

SOTER: Soil And Terrain Analysis

TP: True Positive

TWI: Topographic Wetness Index

USGS: United states Geological Survey

WHO: World Health Organization

xvi
 CHAPTER 1: INTRODUCTION
1.1 Background
Nepal, a landlocked country lying in the Himalayan range, is characterized by a complex
and dynamic geological setting. The continuous tectonic movement of the Indian plate
underneath the Eurasian plate has given rise to the youngest and most tectonically active
mountain range in the world, the Himalayas (Webb et al., 2017; Bhandary et al., 2013).
This rugged and fragile topography, combined with high relief, variable climatic
conditions, and intricate geographic features, renders Nepal highly susceptible to various
geo-hazards, including landslides, erosion, floods, earthquakes, avalanches, and glacial
lake outburst floods (Pokharel & Thapa, 2019).This geological setting, coupled with the
country's monsoon climate and frequent seismic activity, makes Nepal highly susceptible
to landslides (Dahal et al., 2012). Landslides are a significant natural hazard in Nepal,
causing loss of life, damage to infrastructure, and economic disruption (Dhital, 2015).
These gravitational mass movements of rock, debris, or earth down a slope (Cruden, 1991)
pose a significant threat to the natural and built environments, causing injuries, damage to
infrastructure, and economic disruption (Nowicki et al., 2014). Along with this
anthropogenic factors such as rapid infrastructure development, improper land-use
practices, unmanaged settlements, and deforestation (Upreti & Dhital, 1996).

Landslide is defined as the downward movement of a mass rock, debris or earth along a
slope (Cruden, 1991) and denotes a connected complex process influenced by various
causative factors (van Westen et al., 2006). The most influencing factors are bedrock
geology (lithology, structure, degree of weathering), morphometric properties like slope
angle, aspect or elevation, hydrography factors like distance to rivers, geology, soils, SPI
and land use land cover. Due to gravitational pull, the section of hillside or a sloped area of
the bed loses the ability which causes mass to fall down (Wubalem, 2021).

The occurrence of landslides in Nepal is influenced by various causative factors, including

bedrock geology (lithology, structure, degree of weathering), morphometric properties
(slope angle, aspect, elevation), hydrography (distance to rivers), geology, soils, SPI, and
land use/land cover (van Westen et al., 2006). The steep slopes, a key feature of the
Himalayan geomorphology, are particularly susceptible to landslides. Improper land use,
1
haphazard migration, urbanization, and infrastructure development in landslide-prone areas
have further promoted slope instability in the mountainous regions (Dahal et al., 2011).
The scenes of heavy settlements on hilltops, landslide-prone areas, or old mass movement
landscapes are not uncommon in Nepal's context. These landslides degrade the mountain
environment and add sediment loads to streams and rivers, affecting people across Nepal,
especially during the monsoon periods (Shroder & Bishop, 1998;Shang et al., 2003).

Landslide susceptibility refers to the likelihood of a landslide occurring in a specific area

based on various geological, geomorphological, and environmental factors (van Westen et
al., 2008). Understanding landslide susceptibility is crucial for effective risk management,
land-use planning, and disaster preparedness in Nepal (Dahal et al., 2008).Landslide
susceptibility mapping is a valuable tool for reducing and managing landslide-related geo-
hazards. It identifies areas that are prone to future landslides based on the terrain
conditions that determine the natural behavior of landslide activity (Fell et al., 2008;
Pradhan & Lee, 2010). Landslide susceptibility zonation involves mapping regions with
equal chances of landslide occurrence within a given period (D. J. Varnes, 1984).
Susceptibility maps can be used for hazard mitigation planning, disaster management, and
informed decision-making.

Various methods have been developed for generating landslide susceptibility maps,
generally categorized into qualitative (subjective) and quantitative (objective) approaches
(Guzzetti et al., 1999).Quantitative methods for landslide susceptibility mapping involve
the use of numerical data and statistical or mathematical techniques to analyze the
relationship between landslide occurrences and causative factors. Examples of quantitative
methods include: Statistical methods: Frequency Ratio, Weights of Evidence, Logistic
Regression, Discriminant Analysis, etc. (Pradhan, 2010;Van Westen et al., 2003).Machine
Learning techniques: Artificial Neural Networks, Support Vector Machines, Decision
Trees, etc. (Pham et al., 2017;Merghadi et al., 2020).Qualitative methods rely on expert
knowledge, field observations, and subjective evaluation of causative factors. These
methods are often used when quantitative data is limited or when expert judgment is
deemed essential. Examples of qualitative methods include: Heuristic or knowledge-driven
methods: Based on expert opinions and field observations (Darcy, 1963). Analytic

2
Hierarchy Process (AHP) which involves pairwise comparisons of factors based on expert
judgment (Yalcin, 2008).

The Frequency Ratio method is a bivariate statistical approach used to analyze the
relationship between landslide occurrences and various causative factors (e.g., slope,
aspect, geology, land use/cover, etc.). It involves calculating the ratio of the probabilities of
landslide occurrence to non-occurrence for each factor class (Pradhan, 2010).The higher
the frequency ratio value, the stronger the correlation between the factor class and
landslide occurrence. This method is relatively simple and straightforward to implement,
but it does not consider the interdependence between causative factors (Pradhan & Lee,
2010). The Analytical Hierarchy Process is a multi-criteria decision-making technique that
involves pairwise comparisons of factors to derive their relative weights or importance
(Saaty, 1980). In landslide susceptibility mapping, AHP is used to integrate various
causative factors based on their relative importance in influencing landslide occurrence.
Experts' opinions are often used to assign weights to the factors through pairwise
comparisons (Yalcin, 2008). AHP allows for the consideration of interdependence between
factors, but it can be subjective due to the involvement of expert judgment (Yalcin et al.,
2011).

1.2 Rationale
Sindhupalchowk, a district in central Nepal, has witnessed numerous landslides due to its
rugged topography, high seismic activity, and intense precipitation. It has been severely
affected by 2015 Gorkha earthquake, which triggered numerous landslides and caused
significant damage to infrastructure and loss of life (Kargel et al.,2015). Multiple disasters
in Sindhupalchowk including the Jure landslide in 2014, the earthquake of April 2015,
Bhotekoshi flood 2016, Lidi Landslide in 2020 and Melamchi flood and landslide in 2021
have indicated a highly fragile landscape and high multi hazard risk of the district. From
2011-2021, 66 major landslide has taken place, where 128 death, 161 missing people and
684 family has been affected (Thapa et al., 2022).

Brahmayani Watershed encompasses steep slope, valleys, glaciers and river basins and
possesses risk to life and property. This region lies with in tectonically active areas that are
Main Boundary Thrust (MBT) and Main Central Thrust (MCT) this region is very fragile
3
and seismically sensitive making it even more susceptible to landslide making the
communities more vulnerable ( Thapa et al., 2022).

Landslide susceptibility mapping is a crucial tool for identifying areas prone to landslides
and mitigating the associated risks. Landslides can disrupt transportation networks,
damage agricultural lands, and impact local economies (Froude & Petley, 2018).
Sindhupalchowk is home to numerous settlements and infrastructure projects, such as
roads, bridges, and hydropower plants. By identifying susceptible areas, appropriate
mitigation measures can be implemented, reducing potential economic losses and ensuring
the sustainable development of the region (Kayastha et al., 2013). Landslide susceptibility
maps also provide a basis for developing risk management strategies, such as land-use
zoning, early warning systems, and evacuation plans (van Westen et al., 2008). These
strategies can help protect lives and property in high-risk areas.

In light of the above, this study aims to identify and analyze landslide susceptibility in the
Brahmayani Watershed region using the Frequency Ratio (FR) Model and Analytical
Hierarchical Process (AHP) model. The specific objectives are:

1.3 Research question

The research will answer the following questions.

i. How the landslide is spatially distributed in the study area?

ii. What are the factors responsible for causing the landslide?
iii. Which areas are most susceptible to landslides?
iv. Which model is more effective for the landslide susceptible mapping in
the study area?
1.4 Research objectives:
1.4.1 General objective

The general objective of this study is to assess landslide susceptibility in Brahmayani

Watershed.

4
1.4.2 Specific objectives

i. To determine the spatial distribution of landslides in Brahmayani Watershed

ii. To investigate the relation between landslides distribution and causative factors.
iii. To use Frequency Ratio (FR) model, Analytical Hierarchical Process Method
(AHP) to derive the landslide susceptibility map of the area.
iv. To assess and validate the accuracy of the obtained susceptibility map.

5
 CHAPTER 2: LITERATURE REVIEW
2.1 Landslide hazard
Landslides are one of the most devastating natural hazards, causing significant loss of life
and property damage worldwide (Fell et al., 2008).Landslides is the movement of soil,
rock or other materials down a slope and is contributed by both internal and external
variables (Getachew & Meten, 2021). The most influencing factors are bedrock geology
(lithology, structure, degree of weathering), morphometric properties like slope angle,
aspect or elevation, hydrography factors like distance to rivers, geology, soils, SPI and land
use land cover (van Westen et al., 2008).

Varnes classified a slope movement based on the type of movements and the nature of the
material. Rock, debris and earth are three different sorts of materials that influence slope
movements. The hard mass that is in the proper place and remains unchanging before
movements begin is referred to as "rock." "Earth" denotes a minimum of 80% of the
particles have a size of less than 2 mm. Coarse materials with 20%–80% of its particles
larger than 2 mm are referred to as debris. (Ahmed et al., 2020).Varnes (1978), classified
the landslide based on type of materials that flows:

Table 1: Classification of landslide

Type of movement Type of Material

Bed Rock Engineering Soils

Predominantly coarse Predominantly fine

Falls Rock fall Debris fall Earth fall
Topples Rock topple Debris topple Earth topple
Slides Rotational Rockslide Debris slide Earth slide
Translational
Lateral Spreads Rock spread Debris spread Earth spread
Flows Rock flow Debris flow Earth flow
Complex Combination of two or more principal types of movement

6
According to the World Health Organization (WHO), around 4.8 million people were
affected by landslides between 1998 and 2017, with over 18,000 deaths. Climate change
and intense rainfall patterns are expected to increase the likelihood of more landslides,
potentially leading to and increase fatalities (Van Den Eeckhaut et al., 2012). Communities
residing in hilly areas are more susceptible to landslides worldwide as a result of geology,
terrain, climate, and human activity. Increased precipitation, harsh weather, and climatic
oddities make this worse (Chen et al., 2018). Risk assessment, early warning systems,
land-use planning, sustainable land management practices, and stakeholder participation
for readiness, response, and recovery measures can all greatly lessen the effect (Wester et
al., 2019).

2.2 Landslide inventory Maps

Landslide inventories are the basis for assessing landslide susceptibility, hazard and risk.
Landslide inventory mapping is a crucial process in the field of geo-hazard assessment and
risk mitigation. It involves the systematic identification, delineation, and characterization
of landslides within a given area. Therefore we should know about the location of past
landslide events. This process is essential for understanding the spatial distribution,
magnitude, and potential impacts of landslide events, enabling informed decision-making
for land-use planning, infrastructure development, and disaster risk reduction strategies
(van Westen et al., 2008). Here the data are analyzed along with the condition when
landslide occurred and appropriate condition is used to predict the future landslide.
Therefore landslide inventory mapping is the fundamental phase to evaluated hazard and
risk (Aleotti & Chowdhury, 1999).

The first phase in preparing the landslide hazard mapping is mapping of landslide hazard
zonation (Irigaray & Chacón, 1996). It is the simplest form of landslide mapping (Guzzetti
et al., 2000; Hansen, 1984). Inventory maps shows location, date of the events and types of
landslide (Malamud et al., 2004). The preparation of landslide map is done by gathering
the historical information on each landslide events or analysis of aerial photographs along
with field visit (Kayastha et al., 2013). Historical information gives idea about location of
landslide occurring site and aerial photographs portray the distribution of landslide
deposits (Guzzetti et al., 2000).

7
The inventories can be done in different scale such as small, medium and large scale. The
small scale (1:200,000) can be compiled through inquiries to public organization, private
consultants by searching journal and scientific papers (Radbruch-Hall et al., 1982). They
can also be obtained through aerial photographs analysis (Cardinali et al., 1990). Medium
scale landslide inventories (1:25000 to 1:200000) are prepared by the systematic
interpretation of aerial photographs at print scales 1: 60000 to 1:20000 and by integrating
local field check with historical information. Large scale inventories are (1:25000) are
prepared for limited areas using both interpretation of aerial photographs at print scale
>1:200000 and extensive field investigation that uses varieties of tools and techniques
pertaining to engineering geology, geomorphology and geotechnical engineering
(Wieczorek, 1983).

Many factors affect the reliability, completeness and resolution of inventory map such as
freshness and age of landslide, the quality and scale of aerial photographs, morphological
complexity of study area, land use type and degree of experience of geomorphologic used.

2.3 Landslide risk assessment

Landslide risk assessment is a critical process that involves evaluating the potential for
landslide occurrences and their associated consequences. It is a multidisciplinary field that
combines various disciplines, including geology, geomorphology, hydrology, and civil
engineering (Dai et al., 2002). The primary objective of landslide risk assessment is to
identify areas susceptible to landslides and implement appropriate mitigation measures to
reduce the risk to human life, property, and infrastructure.

The assessment process typically begins with a thorough investigation of the study area,
including its geological, geomorphological, and hydrological characteristics. This involves
collecting and analyzing data from various sources, such as field surveys, remote sensing
techniques (e.g., aerial photography, satellite imagery), and historical records of past
landslide events (Guzzetti et al., 2005). The collected data is then used to develop landslide
susceptibility maps, which delineate areas with different levels of landslide susceptibility
based on the combination of predisposing factors, such as slope angle, lithology, soil
properties, and vegetation cover (Whitt et al., 2007).

8
Once the susceptibility maps are created, the next step is to assess the potential
consequences of landslides, including the exposure and vulnerability of people,
infrastructure, and economic activities within the identified hazard areas. This process
involves quantifying the potential losses and damages that could result from landslide
events, considering factors such as population density, land use patterns, and the value of
assets at risk (Whitt et al., 2007). By combining the susceptibility maps with the
consequence analysis, a comprehensive landslide risk assessment can be conducted, which
provides decision-makers with valuable information for developing and implementing risk
mitigation strategies (Corominas et al., 2013).

Effective landslide risk assessment requires a multidisciplinary approach, involving

collaboration among geologists, engineers, planners, and policymakers. It is an iterative
process that should be regularly updated as new data and information become available, or
as changes occur in the study area (Smoll et al., 2020). Furthermore, the integration of
advanced technologies, such as Geographic Information Systems (GIS), remote sensing,
and numerical modeling techniques, has significantly improved the accuracy and reliability
of landslide risk assessments (Guzzetti et al., 2012).

2.4 Landslide Susceptibility Mapping

Landslides are a significant geological hazard that can cause substantial damage to
infrastructure and loss of life. Landslide susceptibility mapping is a crucial tool for
identifying areas with an increased likelihood of landslide occurrence, allowing for
proactive mitigation efforts and risk management strategies. This mapping process
involves the analysis of various environmental factors that contribute to slope instability,
such as topography, geology, soil conditions, hydrology, and land cover (Reichenbach et
al., 2018). By integrating these factors, landslide susceptibility maps provide a spatial
representation of the relative degree of landslide hazard across a given area.

The development of landslide susceptibility maps typically follows a systematic approach

that combines field observations, remote sensing data, and advanced geospatial analysis
techniques. One widely employed method is the use of bivariate or multivariate statistical
models, which establish mathematical relationships between known landslide occurrences
9
and the associated environmental factors (Pourghasemi & Rossi, 2017). These models can
be further enhanced by incorporating machine learning algorithms, such as artificial neural
networks, decision trees, or support vector machines, which can capture complex non-
linear relationships and improve predictive accuracy (Chen et al., 2018). Additionally,
Geographic Information System (GIS) tools play a crucial role in data integration, spatial
analysis, and map visualization.

Landslide susceptibility maps have numerous practical applications in disaster risk

reduction and land-use planning. They serve as valuable resources for decision-makers,
urban planners, and emergency response teams, enabling them to identify high-risk areas
and implement appropriate mitigation strategies (Li et al., 2020). These maps can guide the
selection of suitable locations for infrastructure development, inform land-use zoning
regulations, and support the development of early warning systems. Furthermore, they
contribute to raising public awareness about landslide hazards, fostering preparedness and
resilience within communities. As technology advances and more robust data sources
become available, the accuracy and reliability of landslide susceptibility mapping are
expected to continue improving, ultimately enhancing our capacity to mitigate the impacts
of these formidable natural hazards

2.5 Landslides in Nepal

Landslides are a recurring natural hazard in Nepal, a country with a diverse and complex
topography. The rugged terrain, coupled with intense rainfall during the monsoon season,
makes the country highly susceptible to these destructive events. One of the most
devastating landslides in recent history occurred in August 2014, when a massive landslide
struck the Sindhupalchok district, resulting in the loss of over 150 lives and widespread
destruction (Thapa et al., 2022).

The impact of landslides in Nepal extends far beyond the immediate loss of life and
property damage. These events can disrupt vital infrastructure, such as roads and
communication lines, hampering relief efforts and exacerbating the suffering of affected
communities. Additionally, landslides can have long-lasting environmental consequences,
including soil erosion, water contamination, and the destruction of fragile ecosystems

10
(Dhital, 2015). Efforts to mitigate the risk of landslides in Nepal involve the
implementation of early warning systems, the development of hazard maps, and the
promotion of sustainable land-use practices in vulnerable areas (Shrestha et al., 2020).

11
 CHAPTER 3: MATERIALS AND METHOD
3.1 Study area
The study area, Brahmayani watershed is in Sindhupalchowk Nepal and has a total area of
683 km2 (Fig 1). It is approximately located between 27°42’41‖–28°31’05‖ N latitude and
85°42’59‖–86°19’37‖ E longitude. The elevation of this watershed varies between 644 and
6883 m above the mean sea level (msl). Most of the area in this watershed is covered by
snow and glacier (above 5000 to 6883 m). Administratively, this watershed lies within a
boundary of three Municipality that is Jugal Rural Municipality, Balefi Rural Municipality
and Chautara-Sagachowkgadi Municipality.

The geology of the area is young and fragile located at both lesser Himalaya (Precambrian
age) and higher Himalaya where the main process of landform development is tectonic
upliftment, weathering, erosion, and slope failure (Dhital, 2015).

Figure 1: Location map showing the Brahmayani Watershed in Sindhupalchowk District of

Nepal

12
3.2 Methodology
The research examines probable areas within the study area landslide inventory and
landslide conditioning factors. To obtain the landslide susceptibility map (LSM), the
methodology followed these major processing steps: data collection, reclassification of
landslide factors and model development, model verification and preparation of the LSM.
The causative factors of landslide, inventory of landslide as well as the model and process
used in the study are illustrated (Figure 2).

Google Earth, Satellite Landslide Inventory Destination

images, Past literature Map Map
and Feild vist

Training Sample(
Validation
80%)
Sample (20%)

Slope
Landslide Susceptibility

Frequency Ratio
Aspect
Curvature
Conditioning
Landslide

Elevation
Map
Factors

Analytic Hierarchy Process

Distance to road
Distance to river
SPI
LSM validation (ROC-
Geology
AUC approach)
Rainfall
Land cover/use
Topographic Wetness
index
Normalized Difference
Vegetation Index(NDVI)
Soil type

Objective 1 Objective 2 Objective 3

Figure 2: Research methodology flow chart

13
3.3 Data collection
3.3.1 Data Collection
The study comprises two major methods of data collection as primary data collection and
secondary collection.

Primary Data Collection

 Field work
Data collection was started with the pre-field visit from 20th Jan-22nd Jan 2024. It
helped to plan well for the field visit. Field work was carried out from 12nd Apr-15th
Apr 2024. During field work, 50 landslides were visited and used for validation.
From the field visit, GPS location, elevation, condition of landslide were observed
and recorded.

Secondary Data

 Remote Sensing Data

 Topographic and Geological data

For the preparation of landslide susceptibility map, remote sensing data and ancillary
dataset were collected from various sources. A brief description of the dataset and
respective sources are given below in the Table 2.

14
 Table 2: Data source of causative factors

Factors Data Used/Resolution Technique Classification

Slope ALOS PALSAR DEM By using the DEM the slope range (0-81°) which was 6 Classes
Gradient (°) (12.5*12.5m resolution) classified into 6 classes (Natural break)

ALOS PALSAR DEM By using the DEM slope aspect was determined using GIS
Slope Aspect (12.5*12.5m resolution) software. 9 classes

Plane ALOS PALSAR DEM By using the DEM plane curvature was classified into
Curvature (12.5*12.5m resolution) concave convex and plane using GIS software. 3 classes

Distance
from Drainage Map and Euclidean distance was used using GIS
Drainage GIS made drainage Map software 6 classes

Distance OSM map and Euclidean distance was used using GIS
from Road OSM (Open street Map) software 6 classes

SOTER Nepal Map compiled

Soil Map by FAO Clipping of Required area 5 classes

14
Elevation ALOS PALSAR DEM By using the DEM the elevation range (640-6881m) was 6 classes
(m) (12.5*12.5m resolution) classified into six classes using GIS software. (Natural break)

SPI = As tanβ (Moore et al., 1991) where As is the

ALOS PALSAR DEM specific catchment area and β is the local slope gradient 4 classes
SPI (12.5*12.5m resolution) measured in degrees. (Natural break)

TWI = ln (As / tanβ) (Moore et al., 1991) where As is the

ALOS PALSAR DEM cumulative upslope area draining through a point per unit
TWI (12.5*12.5m resolution) contour length and tanβ is the slope angle at the point. 4 classes

NDVI = (NIR-R) / (NIR+R), where NIR is near inferred

NDVI Landsat 8- image from USGS band R is the red band (NIR=B8, Red=B4) 5 classes

ERSI Map with resolution of Based on the information on ERSI websites, it was
Land cover 12.5*12.5 classified into 7 classes 7 classes
Geology DMG, 2024 Digitization process 11 classes
Annual
average Monthly rainfall data from 1991-2022 was used and
Rainfall From DHM data analyzed in Excel 4 classes

15
Software for Data analysis: GIS is the widely used system for analysis and processing
remote sensing data. Here Arc GIS 3.8 was used for the mapping. MS Excel (2010) was
used for data analysis and visualization.

3.3.2 Preparation of Landslide Inventory Map

Field observation with GPS device and a checklist is used to collect primary data.
Landslide Inventory map was prepared from the combined use of Google maps and field
investigation along with analysis with GPS. Landslide leaves sign on topographic surface
in the form of shape, size and appearance which can be recognized through visual
interpretation from satellite image (Guzzetti et al., 2005). The attributes helps us to
recognize which are digitized and analyzed in Geographic Information System (GIS)
environment. After the demarcation of landslide from both field visit and Google Earth,
inventory data was randomly categorize into two parts of 80% for building model and 20%
for validating the model performance (Nohani et al., 2019).

Landslide areas were measured from the polygonal representation produced through
Google earth and GIS.

3.4 Landslide Susceptibility Mapping

GIS requires the input of relevant data of two types: attribute data (descriptive information
about the geographic data) and spatial data (representing geographic data features like line
and polygon). Landslide susceptibility mapping was conducted from the above data.
Digital Elevation Model (DEM)

A digital elevation model (DEM) representing the terrain is a key to generate various
topographic parameters, which influences the landslide activity in an area. From DEM,
thematic data layers like slope, aspect, elevation, curvature drainage density was prepared.
DEM was downloaded from ALOSPALSAR (12.5*12.5).

Slope Angle
Slope has significant importance in terms of the formation, development, and susceptibility
to landslides. And it is the expression of the rate of the vertical distance to the horizontal
distance between two specified points with the tangent angle. Slope is the angle between
each surface section and horizontal reference point that measures the speed of change in

16
height and that supports the flow of water and other materials in the direction of slope in
terms of the steepest drop in slope for elevation (Thapa et al., 2022).

In the study area slope ranges from 0-81 degree and it was classified into six on the basis
of equal interval classes as shown in figure (3) similar to (Dahal et al., 2011) as slope range
is similar in both study (3).

Figure 3: Slope Map

Aspect:
An aspect map shows both the direction (Poudyal et al., 2010) and grade of a terrain at the
same time . Therefore, it is an important factor in the analysis and production of landslide
susceptibility maps. In mountainous terrain the role of aspect is important to understand
the effects of local climate since various aspect of terrain receives different amount of
climatic factors.
In the study area aspect was classified into nine categories as shown in figure (4).

17
 Figure 4: Aspect Map
Curvature:
Curvature influences the driving and resisting stresses in the direction of mass movement..
Normally, convex slopes are well built as they dispense the runoff equally down the slope
while concave slopes are regarded as potentially unstable as they concentrate water at the
lowest point and contribute to the build-up of adverse hydrostatic pressure. This parameter
is crucial in geomorphology and hydrology for understanding erosion and deposition
processes, as well as flow accumulation patterns (Florinsky, 2016).

The curvature in this study was extracted from the DEM and is classified into (i) convex
(<-0.5), (ii) flat (-0.5–0.5), and (iii) concave (>0.5) similar to (Thapa et al., 2022) as shown
in figure (5).

18
 Figure 5: Curvature Map

Normalized Difference Vegetation Index (NDVI):

The NDVI was used to examine the presence of vegetation cover by measuring surface
reflectivity. Vegetation provides both hydrological and mechanical effects that increase the
stability of slopes by anchoring roots with soil and hence contribute to reducing speed of
rainfall/run-off movement by creating a barrier. This index is extensively used in
agriculture, forestry, and ecological studies to monitor crop health, assess drought impacts,
and track changes in vegetation cover over time (Pettorelli et al., 2005). The NDVI value is
inversely correlated to landslide susceptibility. NDVI is calculated using the following
equation:

NDVI = (NIR - RED) / (NIR + RED) (1)

Where: NIR = Reflectance in the near-infrared band RED = Reflectance in the red band

The NDVI was calculated from Landsat 8 image and was classified into five categories as
(i) -1 –( -0.5)(ii) -0.5 - 0.1,(iii) 0.1–0.2, (iv) 0.2–0.3, and (v) 0.3–0.98 which were used in
landslide hazard assessment similar to (Thapa et al., 2022) because of similar data range.

19
 Figure 6: NDVI Map

Distance from Road

Distance to roads is one of the major anthropogenic factors influencing landslide
occurrences (Nourani et al., 2014) . As roads are extensively built in the terrain this affects
the stability of slopes. Road alters the gully drainage and can affect the local hydrology and
exposes cuts for seepages that can trigger slope failures. Road cuts can destabilize slopes
by removing lateral support and altering natural drainage patterns. Areas close to roads
often experience increased human activity, which can lead to changes in land use and
vegetation cover. landslide frequency often decreases with increasing distance from roads,
although the exact relationship can vary depending on local conditions and road
construction practices (Brenning et al., 2015).

Road network map was obtained from the Euclidean distance and the distance ranged from
0-16 km .It was classified into six classes as shown in figure (6) based on the similar study
of (Thapa et al., 2022) because of similar data range.

20
 Figure 7: Distance from Road

Distance from Drainage:

Rivers play a major role in landslide development (Park et al., 2013).The stability of a
slope depends on the degree of saturation of the material on the slope, and proximity to
streams is considered to be an aggravation factor due to its contribution in saturation. They
can lead the failure of banks because of the sub-quotation of slopes, and the modification
of the ground caused by gully erosion may also influence landslide initiation (Dai & Lee,
2002). The relationship between drainage distance and landslide susceptibility can be
complex. In some cases, very close proximity to drainage might coincide with flatter
terrain, potentially decreasing landslide risk. The influence of this factor often depends on
the local topography and hydrological conditions (Yalcin, 2008).
Distance from drainage map was obtained from the Euclidean distance and the distance
ranged from 0-3.2 km and was classified into six class as shown in figure (7) based on the
similar study of (Thapa et al., 2022) because of similar data range.

21
 Figure 8: Distance from Drainage

Geology:
Lithological and structural variation often leads to difference in strength and permeability
of rock and soil. Amongst all rock types, phyletic dominant formation (Himal formation)
of Higher Himalaya zone has occupied most of the study area and that exhibits most of the
district area is landslide hazards prone area . The other formation that has rock limestone
has also potential of instability of the slopes because of high weathering and the thick soil
depth on the moderate to high gradient slopes. Therefore, the geological map shows the
instability potential for the landslide. Geological factors often interact with other
environmental variables. For example, certain rock types may be more susceptible to
landslides under specific climatic conditions or when subjected to particular land use
practices (N. R. Regmi et al., 2010).

22
 Figure 9: Geological Map

Topographic Wetness Index (TWI):

The topographic wetness index (TWI), also known as the compound topographic index
(CTI), is a steady state wetness index. It is commonly used to quantify topographic control
on hydrological processes. The index is a function of both the slope and the upstream
contributing area per unit width orthogonal to the flow direction. Accumulation numbers in
flat areas will be very large, so TWI will not be a relevant variable. TWI is calculated
using the following equation:

TWI = ln(a / tan β) (2)

Where: a = Upslope contributing area per unit contour length β = Local slope gradient

This index is widely used in hydrology, ecology, and soil science to predict soil moisture
patterns, groundwater flow, and vegetation distribution (Sörensen et al., 2006).In the study

23
area aspect was classified into four classes as i) 0.801-3.3, ii) 3.3-6.69, iii) 6.64-12.89 iv)
12.86-24.72 similar to (Thapa et al., 2022).

Figure 10: TWI Map

Soil:
Lithology is a frequently used factor in landslide susceptibility analyses (Althuwaynee &
Pradhan, 2017).Soil type has a crucial role in determining the terrain characteristics, slope
failure, mass movement and effect of rainfall, irrigation, flood, earthquake and human
factors like construction. Soil categories for watershed were prepared as shown in the
figure (11).
It was downloaded from SOTER and was classified into five classes with description and
dominant soil present in it as:
i) LPi- Steeply to very steeply sloping mountainous terrain rock headwalls- Glelic
Gleysols
ii) CMu- Past glaciated mountainous terrain above upper altitudinal limit of arable
agriculture very steep slopes- Humic Cambisols

24
 iii) CMx- Steeply sloping mountainous terrain- Chromic Cambisols
iv) CMe- Moderately to steeply sloping mountainous terrain- Eutric Cambisols
v) RGe- Past glaciated mountainous terrain below upper altitudinal limit of arable
agriculture moderate to steep slopes- Eutric Regosols

Figure 11: Soil Map

Elevation:
Elevation was considered for landslide hazard because it is affected by geological
processes. It commands the spatial disparity of hydro-meteorological condition and slope
stabilities. It is also an influencing factor for flood as it affects runoff direction, moisture,
temperature, wind direction, and the extent and the depth of the flood. Elevation is a
crucial factor in landslide susceptibility assessment. It influences various processes that
contribute to slope instability: Higher elevations often experience more intense weathering
due to greater exposure to wind and temperature fluctuations. Elevation affects vegetation
types and density, which in turn impacts soil cohesion and hydrology. Steeper slopes, often
associated with higher elevations, are more prone to landslides due to increased

25
gravitational forces. The study area has the elevation ranging from 644m -6883m which
was classified into 6 classes as shown figure(12) similar to (Gautam et al., 2021) because
of the similar data range .

Figure 12: Elevation Map

Annual Precipitation:
Heavy rainfalls trigger floods and landslides. it's not just the total amount of precipitation
that matters, but also its intensity and duration. Sudden, intense rainfall events can trigger
rapid landslides, while prolonged periods of rainfall can lead to deep-seated landslides
(Gariano & Guzzetti, 2016).

The inverse distance weighted (IDW) interpolation approach was applied to create the five
buffer zones of rainfall as (i) <2000 mm, (ii) 2000-2500 mm, (iii)2500- 3000 mm, iv)
>3000 as shown in map below. As per the interpolated map, it is obvious that northern part
of the district receives more rainfall each year compared to the southern part which
intensifies towards North West.

26
 Figure 13: Rainfall Map

Land cover:
The land use influences all four hazards considered in this study. It directly or indirectly
affects some hydrological processes such as surface runoff, evapotranspiration, and
infiltration and physical infrastructures such as open space, building stock, road and
transport infrastructure, and critical facilities. For landslide hazard assessment, the
agricultural land with shallow rooted nature of most of the agricultural crops and the lack
of proper drainage system compared to natural forest areas have different actions. Changes
in land use, such as deforestation or urban expansion, can significantly alter landslide
susceptibility patterns in a region (Reichenbach et al., 2014). In the study area, it was
classified into seven classes based on the data provided by Ersi as shown in figure (3):

27
 Figure 14: Land cover Map

Stream Power Index(SPI)

Stream Power Index (SPI) is a geomorphological parameter used to quantify the erosive
power of flowing water in a landscape. It is derived from the concept that the potential for
erosion and sediment transport in a stream is proportional to the product of discharge and
slope. The SPI is particularly useful in identifying areas of potential erosion and deposition
within a watershed (Moore et al., 1991).

The Stream Power Index is calculated using the following equation:

SPI = A * tan(β) (3)

Where: A = Specific catchment area (upslope contributing area per unit width) β = Local
slope gradient

This index is widely used in geomorphology, hydrology, and soil science to assess
landscape vulnerability to erosion and to predict the spatial distribution of erosion and

28
deposition processes (Florinsky, 2016). Here it was classified into 4 classes with ranges i)-
10-(-5) ii)-5-(-1) iii) (-1)-1 and iv) 1-12

Figure 15: SPI Map

3.5 Models used in study

3.5.1 Frequency Ratio
The relationship between the landslide occurrence area and the landslide causative factors
can be inferred from the relationship between the non-slippery area and the landslide
causative factors. To determine how close their relationship is, a simple statistical
technique was applied to infer it with the frequency ratio approach. In addition, the FR
model has become valuable for ranking preferred causal factors based on their ability to
control landslides (Kannan et al., 2013) as FR can describe clearly the difference of each
score between the landslide causative factors in the class and the landslide occurrence.
Therefore, the number of pixels where landslide occurred on the area must be combined
between causal factors. Then, the rate for each factor was calculated by dividing the
landslide occurrence rate by the proportion of each class in the causative factors (Lee &

29
Talib, 2005). The scale value in each class shows how strongly the attribute of certain
factors is related between landslide occurrences, and where a ratio greater than one
indicates a stronger correlation, a ratio less than one showed a weaker correlation (Lee &
Sambath, 2006).
The calculation steps for an FR for a class of the landslide-influencing factors are below

(4)
Where,
Npix (1) = the number of pixels containing landslides in a particular class
Npix (2) = the total number of landslide pixels of each class of the causative factor
∑Npix (3) = Total number of pixel of each class of the causative factor
∑Npix (4) = Total number of pixel in the study area

𝑃𝑅 = (𝑀𝑖𝑛 𝑅𝐹 − 𝑀𝑎𝑥 𝑅𝐹)/ (𝑀𝑖𝑛 𝑅𝐹 − 𝑀𝑎𝑥 𝑅𝐹) 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 (5)

Scale values obtained using FR are assigned as weighted values to the layers of each factor
map to generate weighted, overlaid and detailed factor thematic maps using a raster
calculator to generate a Landslide Sensitivity Index (LSI) map.
LS=Σ𝐹𝑅 (6)

The calculated values of FR for each pixel of the LSI indicate the relative susceptibility to
the occurrence of landslides. The high pixel values of the LSI are most sensitive to
landslides and the lower pixel values are the least sensitive.

3.5.2 AHP Method

Analytic hierarchy process (AHP) is one of the most common and widely used multi-
criteria methods. The AHP helps integrate various landslide conditioning factors (e.g.,
slope, aspect, geology, land use, etc.) and assign weights to them based on their relative
importance in influencing landslide occurrence (Ayalew et al., 2005). The process involves
the following steps: Ranking / selection is based on the overall purpose, which is
categorized into several criteria. Decisions are usually based on the perception of the
person who is supposed to make the final decision, and evaluate priorities, emphasizing the

30
importance of the consistency and correlation of alternatives compared throughout the
decision-making process (Saaty, 1980).
The AHP procedure is very flexible because it provides an easy way to find the
relationship between criteria and alternatives. Using this method, you can break down
complex problems into specific hierarchies and include both quantitative and qualitative
aspects of the problem in your analysis. The AHP connects all levels of the hierarchy. This
allows you to see how changing one standard affects other standards and alternatives..

The ability to identify rating inconsistencies by consistency index (CI) is a crucial element
of the AHP. In this study, the consistency ratio (CR) is determined by calculating the ratios
between the values of the random index (RI), which is the average consistency index, and
the consistency index (CI) matrix, whose expression is displayed in the equation. A CR of
less than 0.1 would be considered acceptable (Ayalew et al., 2005).

CI= λmax-N/N-1 (7)

λmax is the largest Eigen value and N is the order of the comparison matrix. Saaty,(1980)
developed an average random consistency index (RI) for different matrix orders and
defined the consistency ratio (CR) as the ratio of the consistency index (CI) and the
random consistency index(RI). CR can be calculated by using formula:

CR=CI/RI (8)

If CR is greater than 0.1, the comparison matrix is inconsistent and should be revised. The
main disadvantage of this method is that subjective preference in the ranking of factors
may differ from one expert to another.

Table 3: Random Consistency index (Saaty, 2000)

N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

RI 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.15 1.49 1.51 1.53 1.56 1.27 1.59

31
 Table 4 : Scale of preferences between two parameters in AHP Saaty,(1977).

Preference Factor Degree of Preference Explanation

1 Equally Two factor contribute equally to the

objective

3 Moderately Experience and judgment slightly to

moderately favor one factor over another

5 Strongly Experience and judgment strongly or

essentially favor one factor over another

7 Very Strongly A factor is strongly favor over another and

its dominance is showed in practice

9 Extremely The evidence of favoring one factor over

another is of the highest degree possible of
an affirmation

2,4,6,8 Intermediate Used to represent compromises between

the preference in the weights 1,3,5,7 and 9

Reciprocals Opposites Used for inverse comparison

Finally, the acquisitive weights were integrated the various causative classes in a single
landslide susceptibility index using equation 9:

LSI = Σ(n=1 to m) Ri * Wi (9)

Where Ri is the rating classes, each layer and Wi is the weights for the each of the
landslide conditioning factors. The LSI represents the landslide susceptibility index.

3.5.3 Validation of model

Validation is very important for landslide susceptibility mapping. For the model validation,
20% of the total dataset were used as test dataset. The Receiver Operating Characteristic
(ROC) curve is a widely used graphical technique for evaluating the performance of binary

32
classification models, including landslide susceptibility models. The ROC curve plots the
true positive rate (sensitivity) against the false positive rate (1 - specificity) at various
threshold settings (Yesilnacar & Topal, 2005). The area under the ROC curve (AUC)
provides a quantitative measure of the model's ability to discriminate between landslide
and non-landslide areas (Remondo et al., 2003).

The receiver Operating Characteristics (ROC) and Area under the Curve (AUC) methods
are used to validate the model using the twenty percent (20%) of the total landslide. Each
threshold that is taken into account for calculation result in four different types of pixel or
binary confusion matrix: true positive (TP), false positive (FP), true negative (TN), and
false negative (FN). Within the classes above and below the threshold, the TP and FN
pixels, respectively, are landslides pixels (Vakhshoori & Zare, 2018).

Based on the number of pixels for each threshold, two statistics were calculated

TP rate=TP/ (TP+FN (10)

FP rate= FP/ (TN+FP) (11)

The ROC curve plots TP rates on the y-axis and FP rates on the x-axis. Employing the
ArcSDM ROC tool, the AUC value represents the model performance on the training
dataset and its ability to predict the test dataset. The AUC value indicates the models
success rate in capturing past landslide occurrence and its predictive capability. According
to Nohani et al.,(2019) AUC value ranges from 0.5-1. Generally, an AUC value of 0.5
indicates a random prediction, while values between 0.7 and 0.8 are considered acceptable,
and values above 0.8 are considered excellent (Yesilnacar & Topal, 2005).

33
 CHAPTER 4: RESULT
4.1 Landslide Distribution Map
Landslide inventory map was prepared through field visit, from Google Earth . Altogether
190 landslides were identified that were dividing into training and testing landslide in the
ration of 80:20 for landslide susceptibility mapping and validation of the model. Among
them 140 of them were identified form remote sensing from 2016-2022 and remaining 50
of them were identified form field survey (fig16).Large number of landslides were
observed to the upper section of the study area that is in Jugal Rural Muncipality especially
in Ward no 2 (Golche) and Ward no 3(Gumba).

Figure 16: Landslide Inventory Map

4.2 Landslide Causative Factors

The following graphs were used to depict how landslides were distributed geographically.
A total of 190 landslides were identified. 13 different landslide cumulative factors were
analyzed.

34
 a) Slope
Slope is one of the important landslide conditioning factors. Slope in a study area ranges
from 0-81⸰ where landslides are distributed. 55 of the landslides were reported in the range
between 42⸰-54⸰ with total area of 1.125 sq.km. Similarly, 54⸰-68⸰ has 45 landslide number
with area of 0.6389 sq. km. Minimum number of landslide was found in 0-14⸰ with 8
landslide with an area of 0.0175 sq.km.

1.2 Lanslide area (sq.km) Landslide number 60

1 50

0.8 40

Number of landslide
Landslide area (sq.km)

0.6 30

0.4 20

0.2 10

0 0
68⸰-81⸰ 54⸰-68⸰ 42⸰-54⸰ 28⸰-42⸰ 14⸰-28⸰ 0-14⸰

Slope

Figure 17: Landslide distribution with relation to slope

b) Aspect
Aspect is another important factor where 48 landslides were counted on the southern
aspect with total area of 0.8093 sq.km. Following that southwest has the highest number
with 38 with an area of 0.59189 sq. km. Minimum number of landslide was found in flat
aspect 0 landslide followed by north with 11 landslide number and an area of 0.0934
sq.km.

35
 Landslide area (sq.km) Landslide number
0.9 60

0.8
50
0.7
Landslide arae (sq.km)

0.6 40

Number of landslide
0.5
30
0.4

0.3 20

0.2
10
0.1

0 0

Aspect

Figure 18: Landslide distribution with relation to aspect

c) Curvature
In curvature 83 landslides were counted on the concave aspect with total area of 1.2543
sq.km. 77 landslide were of convex planar with and area of 1.0137 while 31 landslide is of
flat nature with 0.4106 sq. km.

Landslide area (sq. km) Landslide number

1.4 90
1.2 80
Landslide area (sq.km)

Number of landslide

70
1
60
0.8 50
0.6 40
30
0.4
20
0.2 10
0 0
Concave Flat Convex
Curvature

Figure 19: Landslide distribution with relation to curvature

36
 d) Distance form Road
Distance form Road is another important causative factor of landslide. It ranged from 0 to
16000 m .Maximum number of landslide was 68 found from 3000-7000m from road
section with and area of 1.003 sq. km. following that was 1500-3000 m with and area of
0.5895 sq. km. From 12000-16000 m from road section no landslide was found because of
mountain.

Landslide area(sq.km) Number of landslide

1.2 80

70
1
Landslide area (sq.km)

60
0.8

Number of landslide
50
0.6 40

30
0.4
20
0.2
10
0 0
<500 500-1500 1500-3000 3000-7000 7000-12000 >12000
Distance from Road

Figure 20: Landslide distribution with relation to distance from road

e) Distance from Drainage
In this causative factor maximum number of landslide was found in the range of 1200-
2400 m with an area of 1.039 with 74 landslide followed by 300-700 with 40 with and area
of 5.706 sq.km. <100m has the lowest number of landslide with an area of 0.1214 sq.km
also >2400 m has a total of 11 landslide number with an area of 0.1673 sq.km.

37
 Landslide area(sq.km) Number of landslide
1.2 80

70
1
Landslide area (sq.km) 60

Number of landslide
0.8
50

0.6 40

30
0.4
20
0.2
10

0 0
<100 100-300 300-700 700-1200 1200-2400 >2400
Distance from drainage

Figure 21: Landslide distribution with relation to drainage distance

f) Elevation
Elevation ranges from 644-6881 m. The maximum number of landslides was mapped in
the range of 3200-400m with an area of 1.055 sq. km followed by the range of 4000-5000
with sq.km with 45 landslides and area of 0.7268 sq.km. <1500 has no landslide mapped
and from 1500-2400 m 7 landslide were mapped with 0.118 sq. km.

Landslide area Number of landslide

1.2 90
80
1
70 Number of landslide
0.8 60
Landslide area (sq.km)

50
0.6
40
0.4 30
20
0.2
10
0 0
>5000 4000-5000 3200-4000 2400-3200 1500-2400 <1500
Elevation

Figure 22: Landslide distribution with relation to elevation

38
 g) Land cover
In this causative factor maximum number of landslide was found in the Rangeland with 68
landslide and with 1.067 sq.km followed by Forest with an area of 0.9156 sq. km. with 61
landslide. Minimum number was found in Built up and Water body area with 0 land slide
followed by an area covered with Snow with 3 landslides and an area of 0.093 sq.km.

1.2 Landslide area(sq.km) Number of landslide 80

70
1
60
Landslide area (sq.km)

0.8

Number of landslide
50
0.6 40
30
0.4
20
0.2
10
0 0
Water Forest Built up Agricultural Barren Snow/Ice Rangeland
area Land Land area

Land cover

Figure 23: Landslide distribution with relation to land cover

h) NDVI
In NDVI maximum number of landslide was 67 found in range of < -0.05 with an area of
1.08 sq.km followed by 53 landslide in the range of (-0.05-0.1) with an area of 0.7265
sq.km. 11 was lowest number of landslide was found in <0.3sq.km with an area of
0.041875 sq.km. Similarly class 0.1-0.2 has an area of 0.5320 with 40 landslides whereas
from 0.2-0.3 it has 19 landslides with an area of 0.2983 sq.km.

39
 Landslide area(sq.km) Number of landslide
1.2 80

70
Landslide area(sq.km) 1
60
0.8
50

Number of landslide
0.6 40

30
0.4
20
0.2
10

0 0
<-0.5 (-0.5-0.1) 0.1-0.2 0.2-0.3 >0.3
NDVI

Figure 24: Landslide distribution with relation to NDVI

i) TWI
In TWI, maximum number of landslide was found in range of 6.64-12.89 with 82 landslide
with an area of 1.12 sq.km followed with an area of 0.815 sq.km in class >12.86 with 50
landslides. Minimum number of landslide was found in <3.3 with 22 landslide and an area
of 0.2746 sq.km

1.2 Landslide area(sq.km) Number of Landslide 90

80
1
70
Landslide area (sq.km)

Number of landslide

0.8 60
50
0.6
40
0.4 30
20
0.2
10
0 0
>12.86 6.64-12.89 3.3-6.69 <3.3
TWI

Figure 25: Landslide distribution with relation to TWI

40
 j) Geology
In Geology maximum number of landslide was found in Himal formation with 86
landslides and with an total area of 1.82 sq.km. Beside that Ghanapokhara formation has
the highest landslide number with 37 landslides and an area of 0.422 sq.km. Ba, Syangja
and Naudada formation has no landslide.

Landslide area(sq.km) Number of landslide

2 100
1.8 90
1.6 80

Number of landslide
Landslide area (sq.km)

1.4 70
1.2 60
1 50
0.8 40
0.6 30
0.4 20
0.2 10
0 0

Geology

Figure 26: Landslide distribution with relation to geology

k) Rainfall
In rainfall 3000> has the highest number of landslide with 76 landslide with an area of 1.46
sq.km followed up by 2500-3000with 63 landslide with an area of 0.8335 sq.km. <2000
has the lowest landslide count with 7 landslides and an area of 0.074375 sq.km.

41
 landslide area Number of landslide
1.6 80

1.4 70
Landslide area (sq.km)

1.2 60

NNumber of landslide
1 50

0.8 40

0.6 30

0.4 20

0.2 10

0 0
>3000 2500-3000 2000-2500 <2000
Rainfall

Figure 27: Landslide distribution with relation to rainfall

l) Soil

In TWI, maximum number of landslide was found in Lpi with 62 with total area of 1.18
sq.km followed by RGe 51 landslide with area 0.71 sq.km. Similarly, CMu has 50
landslide with 0.70 sq.km area. Minimum number of landslide was found in CMx with
0.0219 sq.km whereas CMe has 20 landslide with area 0.0514 sq.km.

Landslide area(sq.km) Number of landslide

1.4 70

1.2 60
Landslide area(sq.km)

Number of landslide

1 50

0.8 40

0.6 30

0.4 20

0.2 10

0 0
LPi CMu CMe CMx RGe
Soil

Figure 28: Landslide distribution with relation to Soil

42
 m) SPI

In SPI, maximum number of landslide was found in >1 with 84 landslide total area of 1.63
sq.km followed by -1-1 class with 55 landslides with area 0.0.788 sq.km. Similarly, -5-(-1)
has 32 landslide with 0.15 sq.km area. Minimum number of landslide was found in <-5
with 0.0917 sq.km with 12 landslides.

Landslide area(sq.km) Number of landslide

1.8 100
1.6 90

1.4 80
Landslide area(sq.km)

Number of landslide
70
1.2
60
1
50
0.8
40
0.6
30
0.4 20
0.2 10
0 0
<-5 (-5-(-1) (-1-1) >1
SPI Index

Figure 29: Landslide distribution with respect to SPI

4.3 Landslide susceptibility

In the study of landslide susceptibility mapping of Brahmayani Watershed, Frequency
Ratio and Analytic Hierarchy Process were applied. It analyzed and assessed different
landslide causative factor coupled with past landslide to delineate risk zone accordingly. It
was classified into five distinct categories using natural breaks method: very low, low,
moderated, high and very high.

4.3.1 Frequency ratio

Landslides were observed to be more common on slopes greater than42⸰. The slope class
42⸰-54⸰ had the highest FR value of 1.56 followed by 54⸰-68⸰. The slope class between 0-
13⸰ also had an FR value of 0.29. When assessing from aspect perspective south has the
highest FR value, followed by Southwest and southeast with value 1.99, 1.56 and 1.04
respectively.FR of concave curvature was 1.143 while that of convex was 0.903. FR in

43
distance from drainage was highest in the distance from 300-700m with value 1.01,
followed by 1200-2400m with value 1.08 and >2400 m with value from 1.41. In distance
from road highest FR value is from 3000-7000m which is 2.054 while distance >12000 has
0 FR. 500-1500 has second highest value with 1.254 followed by 1500-3000 with 1.1054.

In NDVI causative factor, < (-0.5) has highest FR value with 1.38 followed by (-0.5-0.1)
which has the value of 1.37 while >0.3 has the lowest FR value with 0.27.In Land cover
and land use parameter, Rangeland has the highest FR value with 1.4517 while classes
like Built up area and Water has the lowest FR value with 0. In Topographic Wetness
index class with range 6.64-12.89 has the has the highest FR value 1.41 followed by 3.3-
3.64 with 0.955 <3.3 has the lowest FR value with 0.66. When taking about the soil, LPi
with 1.166 has the highest FR value, followed by Cmu and Rge with value 1.37l and 1.08
while that of CMx has the lowest value with 0.0592.

In the causative factor of elevation 3200-4000msl has the highest FR value with 1.9148
followed by 4000-5000 and >5000 with values 1.52 and 1.41 respectively. The lowest FR
value is 0 in class of <1500.In Geology also Himal formation has the highest FR value
with 1.32 followed by Ghanapokhara formation with values 1.05 respectively. The lowest
FR value is 0 in class of Ba, Syangja and Naudada formation. In a parameter SPI class with
>1 has the highest FR with 1.27 whereas class with (-1-(1)) has the FR with 0.8091
whereas least FR is in class <-5 with value 0.1978. Similarly in rainfall parameter average
annual rainfall with class >3000 has the highest FR with 1.462 whereas class <2000 has
the lowest FR with 0.1631.

44
 Table 5: Frequency ratio table of causative factor
Causative Class Landslide
factors Classes Pixel % Pixels Pixels % Pixel FR PR
>68° 11692 0.269 62 0.362 1.347
54°-68° 195417 4.488 1069 6.235 1.389
Slope

42°-54° 707148 16.241 4349 25.367 1.562

10.76
28°-42° 1568808 36.031 7184 41.904 1.163
13°-28° 1537347 35.309 4089 23.851 0.675
<13° 344673 7.916 391 2.281 0.288
Curvature

Concave 1787112 40.941 8028 46.827 1.144

Flat 803376 18.405 2628 15.329 0.833 3.13
Convex 1774597 40.654 6488 37.844 0.931
(Flat)1 47815 1.098 0 0.000 0.000
(North)2 437353 10.045 630 3.675 0.366
Northeast(3) 462019 10.611 1571 9.164 0.864
East(4) 591520 13.586 1656 9.659 0.711
Aspect

Southeast(5) 575210 13.211 2376 13.859 1.049 10.32

South(6) 659500 15.147 5180 30.215 1.995
Southwest(7) 625722 14.371 3861 22.521 1.567
West(8) 542881 12.469 856 4.993 0.400
Northwest(9) 423065 9.717 1014 5.915 0.609
<100 298109 6.829 777 4.532 0.664
Distance from

100-300 439989 10.079 1616 9.426 0.935

Drainage

300-700 918260 21.035 3652 21.302 1.013

6.29
700-1200 957895 21.943 3374 19.680 0.897
1200-2400 1557859 35.686 6654 38.812 1.088
2400-3200 192973 4.420 1071 6.247 1.413
0-500 1430874 32.777 3773 22.008 0.671
Distance from

500-1500 591555 13.551 2914 16.997 1.254

Road

1500-3000 478428 10.959 2077 12.115 1.105

7.58
3000-7000 807367 18.494 6422 37.459 2.025
7000-12000 732787 16.786 1958 11.421 0.680
>12000 324074 7.424 0 0.000 0.000
-1-(-0.5) 1509675 34.585 8192 47.783 1.382
(-0.05-0.1) 216663 4.964 1169 6.819 1.374
NVDI

0.1-0.2 1263466 28.945 5325 31.060 1.073 10.76

0.2-0.3 1129864 25.884 2190 12.774 0.494
0.3-0.98 245417 5.622 268 1.563 0.278
Water 14094 0.323 0 0.000 0.000
Forest 1700905 38.967 6724 39.221 1.007
LULC

Built up area 42594 0.976 0 0.000 0.000 17.80

Agricultural Land 96816 2.218 111 0.647 0.292
Barren Land 159215 3.648 786 4.585 1.257

45
 Continued table of FR

Causative Class Landslide

factors Classes Pixel % Pixels Pixels % Pixel FR PR
LULC Snow/Ice area 684406 15.680 18 0.105 0.007
Rangeland 1667055 38.192 9505 55.442 1.452
>12.86 34823 0.800 101 0.589 0.737
6.64-12.89 670561 15.401 3038 17.720 1.151
TWI

4.66
3.3-6.69 3256587 74.795 12247 71.436 0.955
<3.3 403114 9.258 1758 10.254 1.108
Lpi(Rock) 1654882 38.081 7609 44.383 1.165
Cmu(Cryothens 985342 22.674 5328 31.078 1.371
Soil

Cme 235340 5.416 329 1.919 0.354 14.57

CMx 619351 14.252 140 0.817 0.057
RGE 870170 20.024 3738 21.804 1.089
5000-6883 513939 11.774 2857 16.665 1.415
4000-5000 860509 19.713 5152 30.051 1.524
Elevation

3200-4000 640496 14.673 4817 28.097 1.915

14.25
2400-3200 706453 16.184 3111 18.146 1.121
1500-2400 950273 21.770 1207 7.040 0.323
644-1500 693415 15.885 0 0.000 0.000
Himal Formation 3022313 69.818 15907 92.785 1.329
Ulleri formation 84342 1.932 304 1.773 0.918
Lakhapatra formation 134580 3.083 70 0.408 0.132
Galyang formation 182629 4.184 91 0.531 0.127
Ba 44194 1.012 0 0.000 0.000
Geology

Syangja formation 64858 1.486 0 0.000 0.000

15.70
Ranimata formation 551485 12.634 353 2.059 0.163
Ghanapokhara
formation 50629 1.160 210 1.225 1.056
Kushma formation 23630 0.541 46 0.268 0.496
Naudada Formation 70635 1.618 0 0.000 0.000
Gn 135790 3.111 163 0.951 0.306
>3000 371625 8.514 2135 12.453 1.463
Rainfall

2500-3000 2664711 61.046 12566 73.297 1.201

11.41
2000-2500 585494 13.413 1967 11.473 0.855
<2000 743255 17.027 476 2.776 0.163
-10 –(-5) 127454 2.920 99 0.577 0.198
((-5-(-1)) 391961 8.979 1016 5.926 0.660
SPI

9.82
(-1-1) 1742163 39.911 5536 32.291 0.809
1-10 2103507 48.189 10493 61.205 1.2701

46
The FR model produced a landslide susceptibility map with five distinct classes: very low,
low, moderate, high and very high susceptibility. And from the figure (30) it can be
depicted that the class with low has the highest area with 209.46 sq.km. Very low
susceptibility has only 21.59 sq.km, similarly moderate has 116.57 sq.km high has 143.98
and very high has 188.40 sq. km from the total area.

Figure 30: Landslide susceptibility map obtained from FR model

250
209.46
200 188.40

143.98
150
116.57
Area (sq.km)

100

50
21.59

0
Very low Low .
Moderate High Very high

Figure 31: Landslide susceptibility area designated based on FR model

47
4.3.2 Analytical hierarchy Process
The application of the Analytical Hierarchy Process (AHP) in this study aims to assess the
relative importance of various factors influencing the decision-making process. Table [6]
displays the pairwise comparison matrix and the resulting criteria weights, which indicate
the significance of each factor. This section provides a detailed analysis of each factor, its
comparison with other factors, and its derived weight.

Slope is an essential factor, particularly in geographical and environmental studies, as it

influences water runoff, soil erosion, and land stability. In the pairwise comparisons, Slope
is equally important to distance from drainage and aspect (both with a value of 1), but less
significant than elevation and land Cover, reflected by values of 1/3 and 1/5, respectively.
Its criteria weight of 0.0591 indicates a moderate level of importance. Aspect, the direction
a slope faces, affects microclimates and vegetation growth. It holds equal importance to
Slope, as shown by a comparison value of 1. Compared to other factors, Aspect is less
crucial than Elevation and Land Cover, with comparison values of 1/3 and 1/5,
respectively. The criteria weight for Aspect is 0.0544, signifying its moderate role in the
analysis.

Curvature refers to the shape of the land surface, influencing water flow and erosion
patterns. It is comparatively less important, as indicated by its lower criteria weight of
0.0216. It shows a relatively less importance to SPI and similar weight distance from
drainage with values of 1/2 and 1/3, respectively, but it is significantly less important than
elevation and land Cover. Distance from drainage plays a role in determining the
availability of water resources and the potential for flooding. It is of moderate importance,
with a criteria weight of 0.0328. In the pairwise comparison matrix, it is less significant
than elevation and land Cover, reflected by values of 1/3 and 1/5, respectively. SPI used to
quantify the erosive power of flowing water in a landscape This factor similar criteria
weight as of slope with weight of 0.0576 , relatively minor importance in this analysis. It
is less important than parameters like land cover, rainfall and soil with comparison values
of 1/4,1/2 and 1/2 respectively.

48
Elevation is one of the significant factor in this analysis, with the criteria weight of 0.1035.
It influences climate, vegetation, and hydrology. Elevation is compared as more important
than most other factors, including slope (value of 3) and land Cover (value of 7). TWI
indicates potential soil moisture and water accumulation areas. It is moderately important,
with a criteria weight of 0.0315. In pairwise comparisons, TWI is less significant than
elevation (1/2) and land cover (2), but equally important as aspect and distance from SPI.
Land Cover, representing the physical material at the earth's surface, is highly significant
with a criteria weight of 0.2128.. It impacts biodiversity, climate, and human activities. In
the matrix, land cover is more important than many factors, including slope and aspect.

Soil characteristics affect plant growth, water retention, and land use. It holds substantial
importance, with a criteria weight of 0.125. Soil is less important than Elevation (value of
3) but equally important as land cover rain influences water availability, soil moisture, and
ecosystem health. It shares the same criteria weight as soil, indicating a significant impact.
Rain is highly important compared to factors like distance from SPI and curvature.

49
 Table 6: Pairwise comparisons of causative factor

All causative 1 2 3 4 5 6 7 8 9 10 11 12 13 Criteria

factor weights

1 Slope 1 0.0522

2 Aspect 1 1 0.0480

3 Curvature 1/5 1/3 1 0.0184

4 Distance from 1 1 5 1
Road 0.0520

5 SPI 2 1 2 1 1 0.0576

6 Distance from 1/3 ½ 3 ½ 5 1

Drainage 0.0292

7 Elevation 3 3 7 2 8 5 1 0.1095

8 TWI ½ ½ 2 ½ 3 1 ¼ 1 0.0293

9 Land cover 5 5 9 5 9 7 3 7 1 0.2128

10 Soil 3 3 7 3 8 5 1 5 ½ 1 0.1200

11 Rain 3 3 7 3 8 5 1 5 ½ 1 1 0.1200

12 NDVI 1 1 5 1 7 3 ½ 2 ¼ ½ ½ 1 0.0584

13 Geology 1 1 5 1 7 3 ½ 2 ¼ ½ ½ 1 1 0.0639

50
 Table 7: Table showing Number, Eigen value, Consistency index, Random index and
consistency ratio

Causative factors N λmax CI RI CR

All 13 13.53 0.04417 1.56 0.02831
Slope 6 6.35 0.07 1.25 0.056
Aspect 9 9.432 0.054 1.45 0.03724
Curvature 3 3.044 0.022 0.52 0.04231
Distance from road 6 6.55 0.11 1.25 0.088
Distance from
6 6.6 0.12 1.25 0.096
drainage
SPI 4 4.136 0.045 0.6 0.0506
Elevation 6 6.61 0.122 1.25 0.0976
Geology 11 11.35 0.035 1.51 0.02318
NDVI 5 5.314 0.0785 1.11 0.07072
Land cover 7 7.648 0.108 1.35 0.08
TWI 4 4.12 0.04 0.89 0.04494
Rainfall 4 4.19 0.06333 0.89 0.07116

The AHP model produced a landslide susceptibility map with five distinct classes: very
low, low, moderate, high and very high susceptibility. And from the figure (32) it can be
depicted that the class with low has the highest area with 181.96 sq.km. Very low
susceptibility has only 20.60 sq.km, moderate has an area of 115.75. Similarly high has
161.67 sq.km and very high has 180.92 sq. km from the total area.

51
 Figure 32: Landslide susceptibility map obtained from AHP model

200
181.96 180.92
180
161.67
160

140 134.85

120
Area (sq.km)

100

80

60

40
20.60
20

0
Very low Low Moderate
. High Very high

Figure 33: Landslide susceptibility area designated based on AHP model

52
4.3 Comparison of AHP and FR model
The Frequency Ratio (FR) and Analytic Hierarchy Process (AHP) models exhibit similar
yet distinct patterns in landslide susceptibility mapping. Both models show a similar extent
of very low susceptibility areas, with the FR model showing a slightly higher area. The FR
model indicates a significantly larger area (209.46 sq. km) compared to the AHP model
(181.96 sq. km). The AHP model shows a larger area (134.85 sq. km) compared to the FR
model (116.57 sq. km). The AHP model has a higher area (161.67 sq. km) compared to the
FR model (143.98 sq. km).

These variations emphasize the impact of different methodologies on susceptibility

mapping outcomes. The FR model indicates a larger extent of low susceptibility areas,
while the AHP model points to more significant moderate and high susceptibility zones.
High and Very High susceptibility areas are prominently located in the central and
northern parts of the map in FR model while in AHP high and very high susceptibility
areas are mainly in the central and northern regions, but there are noticeable differences in
the extent and distribution compared to the FR model.

4.4 Validation of landslide susceptibility Models

The ROC-AUC curve represents a powerful tool for evaluating and visualizing the
performance of binary classification models. The effectiveness of FR and AHP models was
evaluated in the study using True Positive Rate (TPR) and False Positive Rate (FPR)
metrics, with the support of ARC SDM tools. The validation process was carried out in
ARC GIS 10.8, where new landslide data used for testing purposes to generate the ROC-
AUC curve. TPR was employed to assess the precision of landslide classification, while
FPR indicated the number of landslides that were incorrectly identified within non-
landslide areas. For validation, 80% of the training sample was used to find out the success
rate while 20% of the testing sample was used to find out predictive rate.

The finding demonstrates that the ROC-AUC curve values of success rate for FR and AHP
was 0.865 and 0.802 while the predictive rate for FR and AHP was 0.812 and 0.779
respectively. These results indicate a significant level of accuracy in the prediction.

53
Figure 34: Assessment of FR and AHP model performance showing true positive rate
and false positive rate a) Success rate b) Predictive rate

54
Figure 35: Training and testing points to carryout ROC-AUC curve in a) FR susceptibility
map b) AHP susceptibility map

55
 CHAPTER 5: DISCUSSIONS
Landslides are one of the most devastating natural hazards, causing significant loss of life
and property damage worldwide (Fell et al., 2008). The occurrence of landslides in Nepal
is influenced by various causative factors, including bedrock geology (lithology, structure,
degree of weathering), morphometric properties (slope angle, aspect, elevation),
hydrography (distance to rivers), geology, soils and land use/land cover (van Westen et al.,
2006). Therefore there is prominent rise in the frequency of landslides and their impacts
(Froude & Petley, 2018).

5.1 Landslide Inventory

Landslide inventory map gives the past records of the landslide events, their location and
size which is essential (Bhandari & Dhakal, 2020) for the prediction of future (Guzzetti et
al., 1999). Past data were used assuming that landslide will occur in the similar
circumstances (D. J. Varnes, 1984). In the present study, a total of 190 landslides were
mapped using Google Earth and Arc GIS 3.8. Among 50 landslides observed on the field,
majority of landslides were debris flows and followed by rock slide. The distribution of
landslide in the study area is not homogenous and found to be varying with geology and
soil type showing different types of landslides. Most of the landslides were found to be in
the rangeland and forest of the study area. Various types of new road construction and
expansion activities are also going in the study area.

5.2 Landslide Causative factors

Thirteen landslide causative factors were analyzed and calculated that shows relation with
the observed landslides Combination of factors such as slope angle, elevation and rainfall
makes more susceptible to landslide. Thus, both anthropogenic and natural factor are
causes of landslide. With respect to slope as it increases, shear strength on the surface
generally increases and it is expected to have high possibility of landslides (Pradhan,
2010). For slope in frequency ratio 42⸰-54⸰, has the highest FR value with 1.56 while in
AHP slope with 54⸰-68° has the highest criteria weight with 0.316.Also, the number of
landslides found in 42⸰-54⸰ was highest with 55 landslides. According to (Meten et al.,
2015) landslide tend to occur more frequently on steeper slopes because as the slope
increases it results in unstable terrain (Gyawali et al., 2021). In a study (Gautam et al.,

56
2021) slope classes 30° to 40° and 40° to 50° show a positive relation with landslide
occurrence similar to this study. However > 68° has lower frequency ratio because their
higher angles can lead to better drainage and less accumulation of excess water which
reduces the probability of landslide occurrence (A. D. Regmi et al., 2014).

An aspect map shows both the direction and grade of a terrain at the same time. . Slopes oriented
in specific directions may be more vulnerable to triggering factors such as precipitation,
freeze-thaw cycles, or seismic activity, thus heightening the risk of slope instability
(Guzzetti et al., 2005).In this study south aspect has the highest FR where as in AHP
southwest has higher weightage with value 1.99 and 0.399 respectively. The weightage of
the aspect was lowest on northeast facing slope and in flat surface. (Khan et al., 2019) also
had a similar finding. Also, a study in China found that hill slopes facing southeast and
south were especially prone to landslides mostly due to summer rainfall (Ma et al., 2022).
Elevation of the study area ranges from 681-6881m msl. Most landslides were observed in
the elevation class 3200-4000. The landslide number increases as elevation increases
(Pachauri & Pant, 1992). In both FR and AHP model, most of the landslides are found in
the 3200-4000m, with frequency ratio of 2.3864 and 0.447 respectively where as in <1500
m there was no landslide found and has least weightage. This pattern can be attributed to
the steep slopes and rugged terrain typical of higher elevations, which are more susceptible
to landslides due to gravitational forces (Smith et al.,2021).

Curvature analysis allows division of the area into concave, convex, and planar surfaces
and consequently help to identify zones that exhibit proneness to landslide (Mancini et al.,
2010). Here in both FR and AHP concave surface has higher possibility of landslide
followed by the convex surface. It is possibly due to concave slopes generally holding
more water during rainfall, increasing the weight and contributing to slope instability (A.
D. Regmi et al., 2014). (Gautam et al., 2021) also had the similar finding but it contrasted
with the study of which contrast the study of Gyawali et al., (2021).

Environmental variables frequently interact with geological factors which is important

factor in landslide susceptibility. Certain rock types might be more prone to landslides in
certain climates or when exposed to particular land use techniques (N. R. Regmi et al.,
2010). Himal formation, Gn and Naudada formation have higher FR and criteria (AHP) in

57
comparison to other formation. Comparable results has been seen in the paper by (Gautam
et al., 2021;Thapa et al., 2022) which states of landslide were observed in Himal
formation and has positive weights in correlating with landslide occurrences. This finding
corroborates the work of Highland, (2008) who emphasized the importance of geological
structures in landslide susceptibility. Type of rocks found during field survey in the study
area is mostly phylite, slate and schist. There were other deposit like quartzite, and
quaternary deposit. . This aligns with the findings of Smith et al.,(2021), who also
identified metamorphic formations as highly susceptible to landslides. However, our study
found a lower susceptibility for the Galyang formation (FR = 0.1258) compared to their
results, which may be due to regional geological variations.

Land use types with frequency ratios greater than one, such as bare rock, grassland, and
other forested areas, are more likely to experience landslides. On the other hand, certain
land use types, such as agricultural land, snow covered exhibit frequency ratio below 1
indicating lesser landslide. There was no landslide seen at water body and in built up area.
But in proximity to build- up area landslides were found during field visit. When analyzing
meticulously, rangeland and bare land which have sparse vegetation to retain soil have
higher FR values (1.25 and 1.452 respectively) and AHP weightage (0.113 and 0.36
respectively). Akgun et al., (2008)) used FR and AHP to assess landslide susceptibility and
similarly found that areas with minimal vegetation, including rangeland and bare land,
were more prone to landslides, indicating higher susceptibility. Landslide was highly
influenced by the distance from the source of water as river plays an important role in
undercutting the hill slope (Meten et al., 2015). The area which is near to the river have
more landslides (Acharya et al., 2006). This study revealed that areas with greater than
300-700m distance from the river have higher susceptibility of landslide. Areas more than
2400 meters away from drainage channels have a lower FR and less weightage in AHP
suggesting a reduced landslide risk. This is likely due to the increased water infiltration
and erosion near drainage channels, which can destabilize slopes (Wu & Sidle, 1995). The
analysis of distance from roads shows that areas within 500-1500 meters from roads have
the highest FR of 1.2543. This finding suggests that road construction and associated
activities, such as excavation and vibration, may contribute to slope instability (Sun et al.,

58
2018). Areas more than 12000 meters away from roads have the lowest FR of 0.6804,
indicating a minimal impact of roads on landslide occurrence in these regions.

Normalized Difference Vegetation Index (NDVI) values indicate that areas with dense
vegetation (NDVI > 0.3) have a lower FR of 0.2780, suggesting that vegetation plays a
protective role in slope stability. This is consistent with the understanding that vegetation
roots help to bind soil particles and reduce surface erosion (Sidle et al., 1985). Areas with
sparse vegetation (NDVI < -0.5) have a higher FR of 1.3816, indicating a greater
susceptibility to landslides. Regarding SPI in FR the highest susceptibility class (>1)
encompassed 48.19% of the study area with frequency ratio (FR) of 1.27. This aligns with
findings from other studies that consistently identify steep SPI with higher values as highly
susceptible to landslides (Reichenbach et al., 2018).The Topographic Wetness Index
(TWI) is an essential factor in determining landslide susceptibility. Higher range TWI
greater than exhibit the highest incidence of landslides suggesting that regions with higher
moisture accumulation are more prone to landslides. This finding aligns with previous
research indicating that soil moisture content is a critical determinant of slope stability
(Iverson, 2000). From FR and AHP, rainfall within the range of >3000mm and 2500-3000
mm have higher landslide occurring probabilities. This finding is consistent with the
studies by (Guzzetti et al., 2008), which indicate that intense and prolonged rainfall events
are primary triggers for landslides.

5.3 Landslide susceptibility mapping and Validation

Landslide susceptibility mapping and Validation were prepared from the frequency and
Analytic Hierarchy Process points. Five different susceptibility classes were prepared from
the FR and AHP. The FR model highly susceptible classes found to be very high
susceptibility classes found to be very high landslide susceptible classes which is similarity
from the AHP. For the validation of the model AUC curve was prepared.

Result shows the northern regions of have been more susceptible than southern region.
Geology, rainfall, elevation, land use has been the major triggering factor for the
occurrence of landslides. The reliability of the obtained result is validated by one of the
widely used validation technique(Devkota et al., 2013;Kayastha, et al., 2013).

59
Testing data was used for finding predictive rate where Frequency Ratio model has AUC
value of 0.810 and Analytic Hierarchy Process has 0.779 while training data was used for
finding the success rate which has AUC value of 0.865 for Frequency Ratio model and
0.802 for Analytical Hierarchy Process respectively. A similar study was carried out by
Bacha et al., (2018) in the Hunza-Nagar valley ,demonstrated that the frequency ratio
model had a success rate of 79% and prediction accuracy of 80%. Similarly in a study
carried out in Indonesia using similar factor by Rasyid et al., (2016) where success rate and
predictive rate was 85.8% and 85.1% respectively.

In my study both success and predictive rate in FR was more than AHP, indicating that FR
model is more reliable and accurate than AHP. This is because, FR data driven approach
that calculates the relationship between observed occurrences of a landslide and causative
factors (Youssef, 2015) which reduces the biasness in decision making since AHP relies
heavily on expert judgment to assign relative weights to various criteria through pairwise
comparisons (Saaty, 1980). The simplicity of the FR model makes it a preferred choice
when handling large datasets (Pradhan, 2013) than AHP. While AHP can be highly
effective in scenarios where expert knowledge is paramount, its accuracy can be
compromised by inconsistent judgments or biases inherent in the decision-making process
(Forman & Gass, 2001).

60
 CHAPTER 6: CONCLUSION AND RECOMMENDATIONS
6.1 Conclusion
Landslide susceptibility mapping is one of the crucial methods for hazard and risk
management in the landslide-prone areas. In this study, 50 landslides were obtained from
field survey and while other landslide were mapped in Google earth. The landslide data
were from the Department of Hydrology and Meteorology, UN website, Ministry of Home
Affairs, and others. Landslide causative factors used for analysis were elevation, slope,
curvature, geology, rainfall, land use/land cover, distance from the river, and distance from
the road. The causative factors and processes have been mapped and included in the
landslide susceptibility map.

The distribution and movement of landslides in the study area were found to be influenced
by the geological composition, slope, aspect, curvature, distance from road, distance from
river, average annual rainfall, and soil types. Among the formation, Himal formation
exhibiting a higher concentration of landslides based on Geology. The highest number of
landslides seemed to occur at the southern slope angle of 42°-54°, at southern slope,
convex curvature, at a distance of <3000-7000 m from the road, at a distance of 1200-
2400m m from drainage and with average annual rainfall of >4000 mm. In terms of soil
types of the study area, LPi and CMu which are found above upper altitudinal limit are
found to play a more prominent role for the higher frequency of debris flow landslide.
Among thirteen LCFs, LULC, NDVI, rainfall, soil emerged as the most significant
landslide-conditioning factor. In order to predict the susceptibility of areas prone to
landslides, a landslide susceptibility map was created utilizing both the Frequency Ratio
(FR) and Analytic Hierarchy Process (AHP) models. The susceptibility of the area
categorized into five distinct groups. From FR, the largest area in the study region falls
under the low susceptibility zone, covering 30.83% of the total area. Additionally, the very
high, high, moderate, and very low susceptibility zones cover 27.70%, 21.17%, 17.30%,
and 3.47% of the area, respectively. From AHP, the largest area in the study region falls
under the low susceptibility zone, covering 26.75% of the total area. Additionally, the very
high, moderate, high and very low susceptibility zones cover 26.60%, 19.83%, 23.77%,
and 3.02% of the area, respectively.

61
Landslide susceptibility mapping and validation from Frequency Ratio and Analytic
Hierarchy Process models were validated from the success rate and predictive rate of
ROC-AUC curve. The success rate was 86.5 % for Frequency Ratio and 80.2 % for
Analytic Hierarchy Process while predictive rate curve 81.2 % for Frequency Ratio and
77.9 % for Analytic Hierarchy Process respectively. This study shows the potential of the
models to predict landslide susceptible areas and can be used for different purposes by
different stakeholders including engineers, planners, and decision-makers for
infrastructure.

6.2. Recommendations
The recommendations drawn from the study are as follows:

• Landslide data at a finer resolution, further research can be carried out.

• Different multivariate models can also be used to examine the performance in landslide
susceptibility mapping.

62
 Reference
Acharya, G., De Smedt, F., & Long, N. T. (2006). Assessing landslide hazard in GIS: A
case study from Rasuwa, Nepal. Bulletin of Engineering Geology and the
Environment, 65(1), 99–107. doi. 10.1007/s10064-005-0025-y

Ahmed, B., Rahman, M. S., Sammonds, P., Islam, R., & Uddin, K. (2020). Application of
geospatial technologies in developing a dynamic landslide early warning system in a
humanitarian context: the Rohingya refugee crisis in Cox’s Bazar, Bangladesh.
Geomatics, Natural Hazards and Risk, 11(1), 446–468. doi.
10.1080/19475705.2020.1730988

Akgun, A., Dag, S., & Bulut, F. (2008). Landslide susceptibility mapping for a landslide-
prone area (Findikli, NE of Turkey) by likelihood-frequency ratio and weighted linear
combination models. Environmental Geology, 54, 1127–1143.

Aleotti, P., & Chowdhury, R. (1999). Landslide hazard assessment: summary review and
new perspectives. Bulletin of Engineering Geology and the Environment, 58, 21–44.

Althuwaynee, O. F., & Pradhan, B. (2017). Semi-quantitative landslide risk assessment

using GIS-based exposure analysis in Kuala Lumpur City. Geomatics, Natural
Hazards and Risk, 8(2), 706–732. doi 10.1080/19475705.2016.1255670

Ayalew, L., Yamagishi, H., Marui, H., & Kanno, T. (2005). Landslides in Sado Island of
Japan: Part II. GIS-based susceptibility mapping with comparisons of results from
two methods and verifications. Engineering Geology, 81(4), 432–445. doi.
10.1016/j.enggeo.2005.08.004

Bacha, A. S., Shafique, M., & van der Werff, H. (2018). Landslide inventory and
susceptibility modelling using geospatial tools, in Hunza-Nagar valley, northern
Pakistan. Journal of Mountain Science, 15(6), 1354–1370.

Bhandari, B. P., & Dhakal, S. (2020). Spatio-temporal dynamics of landslides in the

sedimentary terrain: a case of Siwalik zone of Babai watershed, Nepal. SN Applied
Sciences, 2(5), 1–17. doi. 10.1007/s42452-020-2628-0

Bhandary, N. P., Yatabe, R., Dahal, R. K., Hasegawa, S., & Inagaki, H. (2013). Areal

63
 distribution of large-scale landslides along highway corridors in central Nepal.
Georisk, 7(1), 1–20. doi 10.1080/17499518.2012.743377

Brenning, A., Schwinn, M., Ruiz-Páez, A. P., & Muenchow, J. (2015). Landslide
susceptibility near highways is increased by 1 order of magnitude in the Andes of
southern Ecuador, Loja province. Natural Hazards and Earth System Sciences, 15(1),
45–57.

Cardinali, M., Guzzetti, F., & Brabb, E. E. (1990). Preliminary maps showing landslide
deposits and related features in New Mexico. US Geological Survey.

Chen, W., Zhang, S., Li, R., & Shahabi, H. (2018). Performance evaluation of the GIS-
based data mining techniques of best-first decision tree, random forest, and naïve
Bayes tree for landslide susceptibility modeling. Science of the Total Environment,
644, 1006–1018. doi 10.1016/j.scitotenv.2018.06.389

Corominas, J., Smith, J. T., Van Den Eeckhaut, M., Winter, M. G., Van Westen, C.,
Cascini, L., Pitilakis, K., Mavrouli, O., Agliardi, F., & Malet, J.-P. (2013).
Recommendations for the quantitative assessment of landslide risk. Aristotle
University of Thessaloniki.

Cruden, D. M. (1991). A simple definition of a landslide. Bulletin of the International

Association of Engineering Geology - Bulletin de l’Association Internationale de
Géologie de l’Ingénieur, 43(1), 27–29. doi 10.1007/BF02590167

Dahal, R. K., Hasegawa, S., Bhandary, N. P., Poudel, P. P., Nonomura, A., & Yatabe, R.
(2012). A replication of landslide hazard mapping at catchment scale. Geomatics,
Natural Hazards and Risk, 3(2), 161–192. doi 10.1080/19475705.2011.629007

Dahal, R. K., Hasegawa, S., Nonomura, A., Yamanaka, M., Masuda, T., & Nishino, K.
(2008). GIS-based weights-of-evidence modelling of rainfall-induced landslides in
small catchments for landslide susceptibility mapping. Environmental Geology, 54(2),
311–324. doi 10.1007/s00254-007-0818-3

Dahal, R. K., Hasegawa, S., Yamanaka, M., & Bhandary, N. P. (2011). Rainfall-induced
landslides in the residual soil of andesitic terrain, western Japan. Journal of Nepal

64
 Geological Society, 42, 137–152.

Dai, F. C., & Lee, C. F. (2002). Landslide characteristics and slope instability modeling
using GIS, Lantau Island, Hong Kong. Geomorphology, 42(3–4), 213–228. doi
10.1016/S0169-555X(01)00087-3

Dai, F. C., Lee, C. F., & Ngai, Y. Y. (2002). Landslide risk assessment and management:
An overview. Engineering Geology, 64(1), 65–87. doi 10.1016/S0013-
7952(01)00093-X

Darcy, L. (1963). Research library. In Educational Forum (Vol. 27, Issue 2). doi
10.1080/00131726309340664

Devkota, K. C., Regmi, A. D., Pourghasemi, H. R., Yoshida, K., Pradhan, B., Ryu, I. C.,
Dhital, M. R., & Althuwaynee, O. F. (2013). Landslide susceptibility mapping using
certainty factor, index of entropy and logistic regression models in GIS and their
comparison at Mugling-Narayanghat road section in Nepal Himalaya. Natural
Hazards, 65(1), 135–165. doi 10.1007/s11069-012-0347-6

Dhital, M. R. (2015). Geology of the Nepal Himalaya: Regional perspective of the classic
collided orogen. In Geology of the Nepal Himalaya: Regional Perspective of the
Classic Collided Orogen. doi 10.1007/978-3-319-02496-7

Nohani,E., Moharrami.,M & Sharaf.,S. (2019). Landslide Susceptibility Mapping Using

Diff erent. Water, 11(7), 1402.

Fell, R., Corominas, J., Bonnard, C., Cascini, L., Leroi, E., & Savage, W. Z. (2008).
Guidelines for landslide susceptibility, hazard and risk zoning for land use planning.
Engineering Geology, 102(3–4), 85–98. doi 10.1016/j.enggeo.2008.03.022

Florinsky, I. (2016). Digital terrain analysis in soil science and geology. Academic Press.

Forman, E. H., & Gass, S. I. (2001). The analytic hierarchy process—an exposition.
Operations Research, 49(4), 469–486.

Froude, M. J., & Petley, D. N. (2018). Global fatal landslide occurrence from 2004 to
2016. Natural Hazards and Earth System Sciences, 18(8), 2161–2181.

65
Gariano, S. L., & Guzzetti, F. (2016). Landslides in a changing climate. Earth-Science
Reviews, 162, 227–252.

Gautam, P., Kubota, T., Sapkota, L. M., & Shinohara, Y. (2021). Landslide susceptibility
mapping with GIS in high mountain area of Nepal: a comparison of four methods.
Environmental Earth Sciences, 80(9), 1–18. doi 10.1007/s12665-021-09650-2

Getachew, N., & Meten, M. (2021). Weights of evidence modeling for landslide
susceptibility mapping of Kabi-Gebro locality, Gundomeskel area, Central Ethiopia.
Geoenvironmental Disasters, 8(1). doi 10.1186/s40677-021-00177-z

Guzzetti, F., Cardinali, M., Reichenbach, P., & Carrara, A. (2000). Comparing landslide
maps: A case study in the upper Tiber River basin, central Italy. Environmental
Management, 25(3), 247–263. doi 10.1007/s002679910020

Guzzetti, F., Carrara, A., Cardinali, M., & Reichenbach, P. (1999). Landslide hazard
evaluation: A review of current techniques and their application in a multi-scale
study, Central Italy. Geomorphology, 31(1–4), 181–216. doi 10.1016/S0169-
555X(99)00078-1

Guzzetti, F., Mondini, A. C., Cardinali, M., Fiorucci, F., Santangelo, M., & Chang, K.-T.
(2012). Landslide inventory maps: New tools for an old problem. Earth-Science
Reviews, 112(1–2), 42–66.

Guzzetti, F., Peruccacci, S., Rossi, M., & Stark, C. P. (2008). The rainfall intensity–
duration control of shallow landslides and debris flows: an update. Landslides, 5, 3–
17.

Guzzetti, F., Reichenbach, P., Cardinali, M., Galli, M., & Ardizzone, F. (2005).
Probabilistic landslide hazard assessment at the basin scale. Geomorphology, 72(1–4),
272–299. doi 10.1016/j.geomorph.2005.06.002

Gyawali, Pradeep; Aryal, Yagya Murti; KC,Prajwol; Ansari, K. (2021). Landslide

Susceptibility Assessment Using Bivariate Statistical Methods: A Case Study of
Gulmi District, western Nepal. VW Engineering International, Volume: 3(ssue: 2),
29–40. doi 10.36297/vw.jei.v3i2.60

66
Hansen, A. (1984). Landslide hazard analysis. Slope Instability.

Highland, L. M. (2008). Introduction The Landslide Handbook-A Guide to Understanding

Landslides. The Landslide Handbook - A Guide to Understanding Landslides, 4–42.
http://landslides.usgs.gov/

Irigaray, C., & Chacón, J. (1996). Methodology for the analysis of landslide determinant
factors by means of a {GIS}: {A}pplication to the {C}olmenar area ({M}alaga,
{S}pain). Proceedings of the \begin{math} 8^{th} \end{math} ICFL’96, May 2016,
163–171.

Iverson, R. M. (2000). Landslide triggering by rain infiltration. Water Resources Research,

36(7), 1897–1910.

J. S. Kargel, G. J. Leonard, Dan H. Shugar, U. K. Haritashya, A. Bevington, E. J. Fielding,

K. Fujita, M. Geertsema, and E. S. M., & This. (2015). Geomorphic and Tectonic
Controls of Geohazards Induced by the 2015 Gorkha Earthquake , Nepal. June, 1–13.
https://www.sciencedirect.com/science/article/pii/S0169555X04001114.

Kannan, M., Saranathan, E., & Anabalagan, R. (2013). Landslide vulnerability mapping
using frequency ratio model: A geospatial approach in Bodi-Bodimettu Ghat section,
Theni district, Tamil Nadu, India. Arabian Journal of Geosciences, 6(8), 2901–2913.
doi 10.1007/s12517-012-0587-5

Kayastha, P., Bijukchhen, S. M., Dhital, M. R., & De Smedt, F. (2013). GIS based
landslide susceptibility mapping using a fuzzy logic approach: A case study from
Ghurmi-Dhad Khola area, Eastern Nepal. Journal of the Geological Society of India,
82(3), 249–261. doi 10.1007/s12594-013-0147-y

Kayastha, P., Dhital, M. R., & De Smedt, F. (2013a). Application of the analytical
hierarchy process (AHP) for landslide susceptibility mapping: A case study from the
Tinau watershed, west Nepal. Computers and Geosciences, 52, 398–408. doi
10.1016/j.cageo.2012.11.003

Kayastha, P., Dhital, M. R., & De Smedt, F. (2013b). Application of the analytical
hierarchy process (AHP) for landslide susceptibility mapping: A case study from the

67
 Tinau watershed, west Nepal. Computers and Geosciences, 52, 398–408. doi
10.1016/j.cageo.2012.11.003

Khan, H., Shafique, M., Khan, M. A., Bacha, M. A., Shah, S. U., & Calligaris, C. (2019).
Landslide susceptibility assessment using Frequency Ratio, a case study of northern
Pakistan. Egyptian Journal of Remote Sensing and Space Science, 22(1), 11–24. doi
10.1016/j.ejrs.2018.03.004

Lee, S., & Sambath, T. (2006). Landslide susceptibility mapping in the Damrei Romel
area, Cambodia using frequency ratio and logistic regression models. Environmental
Geology, 50(6), 847–855. doi 10.1007/s00254-006-0256-7

Lee, S., & Talib, J. A. (2005). Probabilistic landslide susceptibility and factor effect
analysis. Environmental Geology, 47(7), 982–990. doi 10.1007/s00254-005-1228-z

Li, S., Xiong, L., Tang, G., & Strobl, J. (2020). Deep learning-based approach for
landform classification from integrated data sources of digital elevation model and
imagery. Geomorphology, 354, 107045. doi 10.1016/j.geomorph.2020.107045

Ma, S., Shao, X., & Xu, C. (2022). Characterizing the distribution pattern and a physically
based susceptibility assessment of shallow landslides triggered by the 2019 heavy
rainfall event in Longchuan County, Guangdong Province, China. Remote Sensing,
14(17), 4257.

Malamud, B. D., Turcotte, D. L., Guzzetti, F., & Reichenbach, P. (2004). Landslide
inventories and their statistical properties. Earth Surface Processes and Landforms,
29(6), 687–711. doi 10.1002/esp.1064

Mancini, F., Ceppi, C., & Ritrovato, G. (2010). GIS and statistical analysis for landslide
susceptibility mapping in the Daunia area, Italy. Natural Hazards and Earth System
Science, 10(9), 1851–1864. doi 10.5194/nhess-10-1851-2010

Merghadi, A., Yunus, A. P., Dou, J., Whiteley, J., ThaiPham, B., Bui, D. T., Avtar, R., &
Abderrahmane, B. (2020). Machine learning methods for landslide susceptibility
studies: A comparative overview of algorithm performance. Earth-Science Reviews,
207, 103225.

68
Meten, M., Bhandary, N. P., & Yatabe, R. (2015). Application of GIS-based fuzzy logic
and rock engineering system (RES) approaches for landslide susceptibility mapping
in Selelkula area of the Lower Jema River Gorge, Central Ethiopia. Environmental
Earth Sciences, 74(4), 3395–3416. doi 10.1007/s12665-015-4377-8

Moore, I. D., Grayson, R. B., & Ladson, A. R. (1991). Digital terrain modelling: a review
of hydrological, geomorphological, and biological applications. Hydrological
Processes, 5(1), 3–30.

Nourani, V., Hosseini Baghanam, A., Adamowski, J., & Kisi, O. (2014). Applications of
hybrid wavelet-Artificial Intelligence models in hydrology: A review. Journal of
Hydrology, 514, 358–377. doi 10.1016/j.jhydrol.2014.03.057

Nowicki, M. A., Wald, D. J., Hamburger, M. W., Hearne, M., & Thompson, E. M. (2014).
Development of a globally applicable model for near real-time prediction of
seismically induced landslides. Engineering Geology, 173, 54–65. doi
10.1016/j.enggeo.2014.02.002

Pachauri, A. K., & Pant, M. (1992). Landslide hazard mapping based on geological
attributes. Engineering Geology, 32(1–2), 81–100. doi 10.1016/0013-7952(92)90020-
Y

Park, S., Choi, C., Kim, B., & Kim, J. (2013). Landslide susceptibility mapping using
frequency ratio, analytic hierarchy process, logistic regression, and artificial neural
network methods at the Inje area, Korea. Environmental Earth Sciences, 68(5), 1443–
1464. doi 10.1007/s12665-012-1842-5

Pettorelli, N., Vik, J. O., Mysterud, A., Gaillard, J.-M., Tucker, C. J., & Stenseth, N. C.
(2005). Using the satellite-derived NDVI to assess ecological responses to
environmental change. Trends in Ecology & Evolution, 20(9), 503–510.

Pham, B. T., Tien Bui, D., Prakash, I., Nguyen, L. H., & Dholakia, M. B. (2017). A
comparative study of sequential minimal optimization-based support vector machines,
vote feature intervals, and logistic regression in landslide susceptibility assessment
using GIS. Environmental Earth Sciences, 76(10). doi 10.1007/s12665-017-6689-3

69
Pokharel, B., & Thapa, P. B. (2019). Landslide susceptibility in Rasuwa District of central
Nepal after the 2015 Gorkha Earthquake. Journal of Nepal Geological Society,
59(August), 79–88. doi 10.3126/jngs.v59i0.24992

Poudyal, C. P., Chang, C., Oh, H. J., & Lee, S. (2010). Landslide susceptibility maps
comparing frequency ratio and artificial neural networks: A case study from the Nepal
Himalaya. Environmental Earth Sciences, 61(5), 1049–1064. doi 10.1007/s12665-
009-0426-5

Pourghasemi, H. R., & Rossi, M. (2017). Landslide susceptibility modeling in a landslide

prone area in Mazandarn Province, north of Iran: a comparison between GLM, GAM,
MARS, and M-AHP methods. Theoretical and Applied Climatology, 130(1), 609–
633.

Pradhan, B. (2010). Landslide susceptibility mapping of a catchment area using frequency

ratio, fuzzy logic and multivariate logistic regression approaches. Journal of the
Indian Society of Remote Sensing, 38(2), 301–320. doi 10.1007/s12524-010-0020-z

Pradhan, B. (2013). A comparative study on the predictive ability of the decision tree,
support vector machine and neuro-fuzzy models in landslide susceptibility mapping
using GIS. Computers & Geosciences, 51, 350–365.

Pradhan, B., & Lee, S. (2010). Landslide susceptibility assessment and factor effect
analysis: backpropagation artificial neural networks and their comparison with
frequency ratio and bivariate logistic regression modelling. Environmental Modelling
and Software, 25(6), 747–759. doi 10.1016/j.envsoft.2009.10.016

Radbruch-Hall, D. H., Colton, R. B., Davies, W. E., Lucchitta, I., Skipp, B. A., & Varnes,
D. J. (1982). Geological Survey Professional Paper 1183. US Geological Survey,
Washington.

Rasyid, A. R., Bhandary, N. P., & Yatabe, R. (2016). Performance of frequency ratio and
logistic regression model in creating GIS based landslides susceptibility map at
Lompobattang Mountain, Indonesia. Geoenvironmental Disasters, 3, 1–16.

Regmi, A. D., Yoshida, K., Pourghasemi, H. R., DhitaL, M. R., & Pradhan, B. (2014).

70
 Landslide susceptibility mapping along Bhalubang — Shiwapur area of mid-Western
Nepal using frequency ratio and conditional probability models. Journal of Mountain
Science, 11(5), 1266–1285. doi 10.1007/s11629-013-2847-6

Regmi, N. R., Giardino, J. R., & Vitek, J. D. (2010). Modeling susceptibility to landslides
using the weight of evidence approach: Western Colorado, USA. Geomorphology,
115(1–2), 172–187.

Reichenbach, P., Busca, C., Mondini, A. C., & Rossi, M. (2014). The influence of land use
change on landslide susceptibility zonation: the Briga catchment test site (Messina,
Italy). Environmental Management, 54, 1372–1384.

Reichenbach, P., Rossi, M., Malamud, B. D., Mihir, M., & Guzzetti, F. (2018). A review
of statistically-based landslide susceptibility models. Earth-Science Reviews, 180, 60–
91.

Remondo, J., González, A., Díaz de Terán, J. R., Cendrero, A., Fabbri, A., & Chung, C. J.
F. (2003). Validation of landslide susceptibility maps; examples and applications
from a case study in northern Spain. Natural Hazards, 30(3), 437–449. doi
10.1023/B:NHAZ.0000007201.80743.fc

Saaty, T. L. (1977). A scaling method for priorities in hierarchical structures. Journal of

Mathematical Psychology, 15(3), 234–281. doi 10.1016/0022-2496(77)90033-5

Saaty, T. L. (1980). The analytic hierarchy process mcgraw hill, new york. Agricultural
Economics Review, 70(804), 10–21236.

Saaty, T. L. (2001). Fundamentals of decision making and priority theory with the analytic
hierarchy process (Vol. 6). RWS publications.

Shang, Y., Yang, Z., Li, L., Liu, D., Liao, Q., & Wang, Y. (2003). A super-large landslide
in Tibet in 2000: Background, occurrence, disaster, and origin. Geomorphology,
54(3–4), 225–243. doi 10.1016/S0169-555X(02)00358-6

Shrestha, B. R., Rai, R. K., & Marasini, S. (2020). Review of flood hazards studies in
Nepal. The Geographic Base, 7, 24–32.

71
Shroder, J. F., & Bishop, M. P. (1998). Mass movement in the Himalaya: New insights and
research directions. Geomorphology, 26(1–3), 13–35. doi 10.1016/S0169-
555X(98)00049-X

Sidle, R. C., Pearce, A. J., & O’Loughlin, C. L. (1985). Hillslope stability and land use.
American geophysical union.

Smith, Michael J., Goodchild, Michael F., Longley, P. A. (2021). Geospatial Analysis 6th
Edition, 2021 update - de Smith, Goodchild, Longley and Colleagues (Issue June).

Smoll, L. F., Martinez, A. G., Loo, J. Z., Mata, M. V., & Jackson, L. E. (2020).
Investigation of the origin and magnitude of debris flows from the Payhua Creek
basin, Matucana area, Huarochirí Province, Perú. Landslide Risk Management, 477–
484. doi 10.1201/9781439833711-30

Sörensen, R., Zinko, U., & Seibert, J. (2006). On the calculation of the topographic
wetness index: evaluation of different methods based on field observations.
Hydrology and Earth System Sciences, 10(1), 101–112.

Sun, X., Chen, J., Bao, Y., Han, X., Zhan, J., & Peng, W. (2018). Landslide susceptibility
mapping using logistic regression analysis along the Jinsha river and its tributaries
close to Derong and Deqin County, southwestern China. ISPRS International Journal
of Geo-Information, 7(11), 1–29. doi 10.3390/ijgi7110438

Thapa, N. B., Kafle, K. R., Niraula, R. R., Banskota, T. R., Kunwar, S., Pokharel, N.,
Shrestha, A., Neupane, R., & Shrestha, A. (2022). Multi Hazard Assessment and
Vulnerability Mapping of Sindhupalchowk District.

Upreti, B. N., & Dhital, M. R. (1996). Landslide studies and management in Nepal. (No
Title).

Vakhshoori, V., & Zare, M. (2018). Is the ROC curve a reliable tool to compare the
validity of landslide susceptibility maps? Geomatics, Natural Hazards and Risk, 9(1),
249–266. doi 10.1080/19475705.2018.1424043

Van Den Eeckhaut, M., Hervás, J., Jaedicke, C., Malet, J. P., Montanarella, L., & Nadim,
F. (2012). Statistical modelling of Europe-wide landslide susceptibility using limited

72
 landslide inventory data. Landslides, 9(3), 357–369. doi. 10.1007/s10346-011-0299-z

van Westen, C. J., Castellanos, E., & Kuriakose, S. L. (2008). Spatial data for landslide
susceptibility, hazard, and vulnerability assessment: An overview. Engineering
Geology, 102(3–4), 112–131. doi 10.1016/j.enggeo.2008.03.010

Van Westen, C. J., Rengers, N., & Soeters, R. (2003). Use of geomorphological
information in indirect landslide susceptibility assessment. Natural Hazards, 30(3),
399–419. doi 10.1023/B:NHAZ.0000007097.42735.9e

van Westen, C. J., van Asch, T. W. J., & Soeters, R. (2006). Landslide hazard and risk
zonation - Why is it still so difficult? Bulletin of Engineering Geology and the
Environment, 65(2), 167–184. doi 10.1007/s10064-005-0023-0

Varnes, D. (1978). Slope Movement Types and Processes. Special Report, 176, 11–33.

Varnes, D. J. (1984). Landslide hazard zonation: a review of principles and practice (Issue
3).

Webb, A. A. G., Guo, H., Clift, P. D., Husson, L., Müller, T., Costantino, D., Yin, A., Xu,
Z., Cao, H., & Wang, Q. (2017). The Himalaya in 3D: Slab dynamics controlled
mountain building and monsoon intensification. Lithosphere, 9(4), 637–651. doi
10.1130/L636.1

Wester, P., Mishra, A., Mukherji, A., & Shrestha, A. B. (2019). The Hindu Kush Himalaya
Assessment: Mountains, Climate Change, Sustainability and People. In The Hindu
Kush Himalaya Assessment: Mountains, Climate Change, Sustainability and People.
doi 10.1007/978-3-319-92288-1

Whitt, G., Fell, R., Corominas, J., Bonnard, C., Cascini, L., Leroi, E., & Savage, W. Z.
(2007). planning Australian Geomechanics. Engineering Geology, 102(3–4), 13–36.

Wieczorek, G. F. (1983). Preparing a Detailed Landslide-Inventory Map for Hazard

Evaluation and Reduction. In Bulletin of the Association of Engineering Geologists
(Vol. 21, Issue 3, pp. 337–342). doi 10.2113/gseegeosci.xxi.3.337

Wu, W., & Sidle, R. C. (1995). A distributed slope stability model for steep forested

73
 basins. Water Resources Research, 31(8), 2097–2110.

Wubalem, A. (2021). DEM Resolutions for Landslide Susceptibility Modeling. 1–24.

Yalcin, A. (2008). GIS-based landslide susceptibility mapping using analytical hierarchy

process and bivariate statistics in Ardesen (Turkey): Comparisons of results and
confirmations. Catena, 72(1), 1–12. doi 10.1016/j.catena.2007.01.003

Yalcin, A., Reis, S., Aydinoglu, A. C., & Yomralioglu, T. (2011). A GIS-based
comparative study of frequency ratio, analytical hierarchy process, bivariate statistics
and logistics regression methods for landslide susceptibility mapping in Trabzon, NE
Turkey. Catena, 85(3), 274–287. doi 10.1016/j.catena.2011.01.014

Yesilnacar, E., & Topal, T. (2005). Landslide susceptibility mapping: A comparison of

logistic regression and neural networks methods in a medium scale study, Hendek
region (Turkey). Engineering Geology, 79(3–4), 251–266. doi
10.1016/j.enggeo.2005.02.002

Youssef, A. M. (2015). Landslide susceptibility delineation in the Ar-Rayth area, Jizan,

Kingdom of Saudi Arabia, using analytical hierarchy process, frequency ratio, and
logistic regression models. Environmental Earth Sciences, 73, 8499–8518.

74
 Appendices
Appendix-1: Photographs

Photographs 1: Landslide at Golche Photographs 2: Landslide at Selang

Photographs 3: Landslide at Hagam

Photographs 4: Landslide at Gumba

I
 Photographs 6 : Landslide at Baramche Photographs 5: Landslide at Yanglakot

Appendix-2: Pairwise comparisons of each causative factor

Slope:

criteria
slope 0-13° 13°-28° 28°-41° 41°-55° 55°-68° 68°-81° weight
0°-13° 1 1 0.2 0.111111 0.142857 0.333 0.045866973
13°-28° 1 1 0.333 0.143 0.2 0.333 0.050291182
28°-42° 5 3 1 0.333 0.2 0.33 0.105455441
42°-54° 9 7 3 1 0.33 0.33 0.210136363
54°-68° 7 5 5 3 1 0.3 0.316157231
68°-81° 3 3 3 3 3 1 0.272226916
Aspect:

Aspect Flat North Northeast East Southeast South Southwest West Northwest criteria weight
Flat 1 0.2 0.33 0.333 0.333 0.3333 0.1111 0.3333 1 0.027734396
North 5 1 2 2 1 1 0.2 2 7 0.11844542
Northeast 3 0.5 1 1 0.333 0.333 0.1473 0.5 5 0.059770301
East 3 0.5 1 1 0.5 0.5 0.1473 0.5 5 0.063552017
Southeast 3 1 3 2 1 1 0.2 2 7 0.116966215
South 3 1 3 2 1 1 0.2 2 7 0.116966215
Southwest 9 5 7 7 5 5 1 7 5 0.399817997
West 3 0.5 2 2 0.5 0.5 0.14287 1 5 0.078946572
Northwest 1 0.1472 0.2 0.2 0.1473 0.1473 0.2 0.2 1 0.02456367

II
Curvature:

Curvature Concave Flat Convex Criteria weight

Concave 1 5 2 0.602823959
Flat 0.2 1 0.333 0.113858085
Convex 0.5 3 1 0.319930498

Distance from road:

Distance from Road <500 500-1500 1500-3000 3000-70007000-12000>12000 criteria weight
<500 1 2 3 5 7 9 0.390159163
500-1500 0.5 1 4 5 7 7 0.31358841
1500-3000 0.333 0.25 1 3 5 3 0.146022523
3000-7000 0.2 0.2 0.333 1 1 2 0.062529923
7000-12000 0.1473 0.1473 0.2 1 1 1 0.046604585
>12000 0.1111 0.1473 0.333 0.5 1 1 0.041095395

Distance from river:

Distance from River

<100 100-300 300-700 700-1200 1200-2400 >2400 Criteria weight(%)
<100 1 5 3 4 5 7 0.42985919
100-300 0.2 1 2 3 3 5 0.204590493
300-700 0.333 0.5 1 2 4 3 0.158300808
700-1200 0.25 0.333 0.5 1 1 2 0.080250549
1200-2400 0.2 0.333 0.25 1 1 1 0.062036397
>2400 0.1473 0.2 0.333 0.5 1 1 0.049979618

Elevation:

Elevation <1500 1500-2400 2400-32003200-40004000-5000>5000 criteria weight

<1500 1 1 0.5 0.1473 0.333 0.5 0.054780055
1500-2400 1 1 0.5 0.1473 0.333 0.5 0.054780055
2400-3200 2 2 1 0.2 0.5 7 0.175689876
3200-4000 7 7 5 1 5 2 0.447225011
4000-5000 3 3 2 0.2 1 2 0.161511469
>5000 2 2 0.1473 0.5 0.5 1 0.106013534

III
NDVI:

NDVI <-0.5 (-0.5-0.1) 0.1-0.2 0.2-0.3 >0.3 Criteria weight

<-0.5 1 2 3 4 5 0.413590949
(-0.5-0.1) 0.5 1 2 3 3 0.246934965
0.1-0.2 0.333 0.5 1 2 4 0.175288738
0.2-0.3 0.333 0.333 0.5 1 1 0.091417679
>0.3 0.2 0.333 0.25 1 1 0.072767668

TWI:

TWI <3.3 3.3-6.649 6.64-12.89 >12.89 Criteria weight

<3.3 1 2 2 3 0.411846715
3.3-
6.64.69 0.5 1 2 3 0.29304457
6.64-
12.89 0.5 0.5 1 2 0.187200715
>12.89 0.333 0.333 0.5 1 0.107908

Landcover:

Landcover Water Forest Built up areaAgriculturalbarren

land Area Snow/Ice Rangeland Criteria weight
Water 1 0.1473 1 2 0.1473 1 0.1111 0.0543161
Forest 7 1 7 1 7 7 1 0.290996815
Builtup area 1 0.1473 1 1 0.5 1 0.142857143 0.04745697
Agricultural land 0.5 1 1 1 0.5 2 0.2 0.087721798
Barren 7 0.1473 2 2 1 3 0.125 0.113416773
Snow/Ice 1 0.1437 1 0.5 0.33 1 0.1111 0.038523177
Rangeland 9 1 7 5 8 9 1 0.367571402

Rainfall:

Rainfall <2000 2000-2500 2500-3000 >3000 criteria weight

<2000 1 0.3333 0.1473 0.111 0.045598423
2000-2500 3 1 0.333 0.1473 0.101399202
2500-3000 7 3 1 0.33 0.261584612
>3000 9 7 3 1 0.592213872

IV
Soil:

criteria
Soil Lpi Cmu Rge CMx Cmu weight
Lpi 1 2 3 4 5 0.516988687
Cmu 0.5 1 2 3 3 0.308668706
RGe 0.333 0.5 1 2 4 0.219110923
CMx 0.333 0.333 0.5 1 1 0.114272099
CMu 0.2 0.333 0.25 1 1 0.090959586

Geology:
Geology Himal Ulleri Lakhapatra Galyang Ba Syanga Ranimatta Ghanapok Kushma Naudada GN criteria weight
Himal 1 2 5 5 8 8 5 0.333 8 8 0.2 0.186812313
Ulleri 0.2 1 1 1 1 3 3 0.3333 3 8 0.2 0.08155375
Lakhapatra 0.2 1 1 1 1 1 1 0.1473 1 3 0.142857 0.043306262
Galyang 0.125 1 1 1 3 3 1 0.3333 0.2 0.33 0.33 0.04420141
Ba 0.125 1 1 1 1 3 1 0.1473 0.111111 0.125 0.125 0.03320485
Syanga 0.2 0.33 1 1 5 1 0.2 1 0.142857 0.2 0.2 0.051024405
Ranimatta 0.2 0.33 1 1 8 5 1 0.1473 0.333 1 0.333 0.053833016
Ghanapok 0.333 3 7 7 3 3 7 1 3 7 3 0.203584078
Kushma 0.125 1 1 1 9 7 1 0.33 1 2 0.1473 0.067616028
Naudada 0.125 1 3 1 8 5 1 0.1473 0.5 1 1 0.071057036
GN 0.2 5 7 7 8 5 3 0.333 7 1 1 0.163847487

Appendix-3: Table showing landslide number and area of landslide

Area (Sq.km)

Landslide
Frequency ratio Classes Area number
Slope 0.12359375 25
54°-68° 0.63890625 45
42°-54° 1.1225 55
28°-42° 0.60921875 35
14°-28° 0.16703125 22
<13° 0.0175 8
Total 2.67875 190
Parameters Classes
Curvature Concave 1.25437435 82
Flat 0.4106255 31
Convex 1.01375015 77
Total 2.67875 190
Parameters Classes
Aspect Flat 0 0
North 0.0934375 11

V
 Northeast 0.24546875 20
East 0.25875 16
Southeast 0.37125 28
South 0.809375 48
Southwest 0.591875 38
West 0.13375 15
Northwest 0.1584375 14
Total 2.67874975 190
Parameters Classes
Distance from
Drainage <100 0.12140625 8
100-300 0.2525 18
300-700 0.570625 40
700-1200 0.5271875 39
1200-2400 1.0396875 74
>2400 0.16734375 11
Total 2.67875 190
Parameters Classes
Distance from Road <500 0.32453125 21
500-1500 0.4553125 37
1500-3000 0.58953125 43
3000-7000 1.0034375 68
7000-12000 0.3059375 21
>12000 0 0
Total 2.67875 190
Parameters Classes
NVDI <-0.5 1.08 67
(-0.05-0.1) 0.7265 53
0.1-0.2 0.53203125 40
0.2-0.3 0.2983 19
<0.3 0.041875 11
Total 2.67870625 190
Landslide
Parameters Classes Area(sq.km) number
Land cover Water 0 0
Forest 0.915625 61
Built up area 0 0
Agricultural Land 0.02315675 9
Barren Land 0.5790625 47
Snow/Ice area 0.0933868 5
Rangeland 1.06746875 68
Total 2.6786998 190
Landslide
Parameters Classes Area(sq.km) number

VI
TWI >12.86 0.8157 50
6.64-12.89 1.113 82
3.3-6.69 0.4746 36
<3.3 0.27468 22
2.6787 190
Landslide
Parameters Classes Area(sq.km) number
soil Lpi 1.18890625 62
Cmu 0.7025 50
Cme 0.05140625 20
CMx 0.021875 7
RGE 0.7140625 51
Total 2.67875 190
Landslide
Frequency ratio Classes Area(sq.km) number
Elevation >5000 0.35265625 25
4000-5000 0.726875 45
3200-4000 1.05578125 80
2400-3200 0.42484375 33
1500-2400 0.118859375 7
<1500 0 0
Total 2.679015625 190
Landslide
Parameters Classes Area(sq.km) number
Geology Himal formation 1.82296875 86
Ulleri formation 0.1986 18
Llakhapatra formation 0.0419375 3
Galyang formation 0.01421875 3
Ba 0 0
Syangja formation 0 0
Ranimata formation 0.05515625 3
Ghanapokhara
formation 0.4222125 37
Kushma formation 0.0071875 17
Naudada Formation 0 0
Gn 0.11671875 23
Total 2.679 190
Landslide
Casautive factor Classes Area(sq.km) number
Rainfall >4000 1.4634375 76
3000-4000 0.83359375 63
2000-3000 0.30734375 44
<2000 0.074375 7
2.67875 190

VII
 Landslide
Casautive factor Classes Area(sq.km) number
SPI <-5 0.09171875 12
(-5-(-1)) 0.15875 32
(-1-1) 0.78875 55
>1 1.63953125 91
Total 2.67875 190

VIII