5

Arba Minch University

College Of Natural and Computational Science

Department of Geology

Stream : Environmental Geology and Geohazard

Proposal Title:
Landslide susceptibility modeling using GIS and frequency ratio method in Kulfo River
Catchment, Southern Ethiopia

Prepared by Id

Isayas Habitamu PRPGNCS/022/17

Advisor: Muralitharan Jothimani (PhD)

Arba Minch, Ethiopia

 Chapter One

Introduction
These Chapter include the following

1.1 Background of the Study

Landslides occur when debris, rocks, or soil particles move downward. Examining the
susceptibility of landslides is essential for safeguarding human well-being and assessing the
consequences of landslides on the natural surroundings and ecosystems. This study utilized the
frequency ratio technique to evaluate the probability of landslides happening in the Kulfo River
watershed, situated in the Rift Valley area of Ethiopia. In order to ensure a comprehensive
analysis, many data sources were employed, including satellite images, geological data, and
historical records of landslides. This study developed a systematic approach to assess the
probability of landslides by considering ten (10) influential factors: land use/land cover, slope,
aspect, elevation, curvature, lithology, proximity to lineament, and normalized difference
vegetation index (NDVI). The ten influential elements were prioritised based on literature review,
expert knowledge, and preliminary study area analysis. The aforementioned causal elements are
integrated with a comprehensive landslide inventory map. The sensitivity of the study's area to
landslides was mapped using a frequency ratio (FR) model. Subsequently, it was classified as
representing various degrees of vulnerability, spanning from extremely minimal to quite
significant. The effectiveness of the suggested model was measured by evaluating the accuracy of
the generated map of landslide susceptibility by the area under the curve (AUC) technique. Based
on the most recent study results, the success rate curve has an area under the curve (AUC) of
79.6%, which indicates a highly satisfactory level of performance. Policymakers may utilize the
findings of this study to make educated decisions on how to mitigate the risks of landslides in
relation to land use and preparedness for disasters. The varied geological and geomorphological
features of the Rift Valley in Ethiopia raise substantial concerns about the possibility of
landslides.The landscapes encompass lofty plateaus, precipitous cliffs, extensively incised valleys,
and flat lowlands (Abay and Barbieri, 2012; Abebe et al., 2010Abebe et al., 2020; Van Westen et
al., 2006).
1.2. Study Area and Accessibility

The study area, known as the Kulfo River watershed, is located in the Abaya-Chamo subbasin of
the Southern Ethiopian rift valley and flows into Lake Chamo. The location is positioned within
the latitudes of 6°15′N and 6°16′N and the longitudes of 37°19′E and 37°35′E. The Kulfo River is
connected to a large number of tributaries. The upper section of the basin is drained by the streams
Baba, Gulando, and Yeremo, while the middle section of the watershed is drained by the
tributaries Wombale and Majale. Korzha, Ambule, and Titika rivers converge with Kulfo at the
bottom of the basin. Following substantial rainstorms in the elevated region upstream of the Gauge
Mountains, flooding is recurring from April to October. The primary drainage system the Arba
Minch city authority manages is the Kulfo River Basin. The research area encompasses a diverse
variety of elevations, encompassing both low-lying areas and elevated plateaus. The Kulfo River
catchment is located in the Southern Ethiopian Rift Valley, approximately 500 km southwest of
Addis Ababa, near Arba Minch, and drains into Lake Chamo. (Shano et al., 2021).

https://ars.els-cdn.com/content/image/1-s2.0-S2666033424000297-gr1_lrg.jpg

1.3. Climate Conditions and Physiography

The Kulfo River catchment area in Southern Ethiopia experiences a climate characterized by a dry
sub-humid climate zone, with average annual rainfall varying from 750 mm in the lowlands to
3342 mm in the mountainous regions, and a mean annual temperature ranging from 23.05 to
25.87°C and The average annual rainfall and potential evapotranspiration of the Lower Kulfo
Catchment were 904 mm and 1533 mm, respectively. The climate classification of the study was
found under dry sub-humid climate zone that includes arid and semi-arid regions.
1.4. Statement of the Problem
The Kulfo River Catchment, located in the Rift Valley area of Ethiopia, is Highly exposed to
landslides, leading to destruction of farmland, pastures, and buildings, impacting the livelihoods
and safety of the local population..Now aday the southern Ethiopia Region, landslide-generated
problems are becoming serious concerns to local people and administration. This area is one of
specific area in southern Ethiopia that has many active landslides in recent decades, causing
significant harm to infrastructure and farming and also loss of humans life.
1.5. Objectives

1.5.1. General Objective

The main objective of this research work is to produce a landslide susceptibility mapping using
GIS and frequency ratio method on Kulfo River Catchment, South Ethiopia

1.5.2. Specific Objectives

 Identify and analyze factors influencing landslides

 Create a landslide inventory map
 Classify landslide hazard zones
 Assess potential risks to human life, infrastructure, and land use
 To Recommend how to mitigate that hazardous that caused by landslide
1.6. Research Questions
The research questions that are concerned in this study are:

 How can frequency ratio method and GIS techniques are used to model landslide
susceptibility in study area?

 What factors are most important in determining landslide susceptibility in study area?

 How can frequency ratio and GIS techniques are used to predict landslides in study area?

 How can frequency ratio and GIS techniques are used to map landslide hazards in study
area?
 How to mitigate landslide hazards?

1.7. Significance of the Study

 To reduce damage to infrastructure, houses, and cultivated lands, as well as loss of life.

 To identify areas prone to landslides, this can be used to develop mitigation strategies and
safer land use practices.  

 To help decision-makers plan land use and implement landslide prevention and mitigation
measures.

 To develop warning system of hazards that caused by landslide.

Chapter Two

2. Literature Review

2.1. Landslide Susceptibility and Environmental Factors

2.1.1. Landslide Susceptibility

Landslide susceptibility is a quantitative or qualitative assessment of the spatial distribution of

landslides, which exist or potentially may occur in an area (Landslide Studies Division 2018).
Landslide susceptibility is usually based on two assumptions: First, that areas which have
experienced landslides in the past are likely to experience them in the future, and Second, that
areas with a similar set of geo-environmental conditions, such as geology, slope, land use etc. as
that of the failed areas, are also likely to fail in future. This means that the quantitative estimates of
the spatial location of future landslide sources depend on the detailed information on the
distribution of past landslides and a set of thematic variables such as slope angle, lithology etc.
that has initiated these landslides. The second assumption facilitates in predicting the geographical
location of future landslides in passive areas (i.e., areas presently devoid of landslides) provided
the geo-environmental conditions remain the same. The above assumptions have been successfully
used for statistically quantifying landslide susceptibility using landslide as dependent variable. A
landslide susceptibility map identifies areas which are subject to landslides and is measured from
low to high (Geological Survey Ireland is a Division of Department of the Environment, Climate
and Communications © 2024).The landslide susceptibility map takes into account where the
landslides occur and what causes them (slope, soil type and the impact of the flow of water in an
area.

2.1.2. Landslide influencing factors

Landslide formation mechanisms are complex and diverse, and are generally determined by a
combination of internal and external factors (Sun et al. 2022b; Liu Y. et al., 2023). Generally,
there is no clear standard for the selection of factors, and it varies according to the objectives of
the study (Yong et al. 2022). Generally there are three main categories of factors: geological
environmental factors (including topography, meteorology and hydrology, geological conditions,
LULC, and vegetation cover), geological hazard factors (typhoons, rainfall,
temperature,earthquake, snow-melt, drought, desertification, etc.), and human activity factors
(including road construction and fortification).

2.1.2.1.Geological environmental factors

Geological environmental factors include the following: slope, aspect, curvature, lithology, soil,
NDVI, NDWI, Terrain moisture index (TWI), distance to fault, and distance to river. These
evaluation factors are calculated from DEM data. Aspect directly affects the stability of the slope
thus leading to landslides ((Hong et al., 2019; Liu Y. et al., 2023). Curvature is an important
parameter to express the geometric characteristics of slopes, and some literature shows that plane
curvature has a strong relationship with surface erosion and surface runoff as a way to accelerate
the probability of landslide occurrence ((Hong et al. 2019; Chen et al. 2021; Liu Y. et al., 2023).
Terrain moisture index (TWI) is a physical factor of the influence of regional topography on the
direction and accumulation of runoff, and it can influence the regional hydrological processes; the
higher the TWI value, the higher the soil moisture content in the area, which makes the soil less
capable of consolidation, and the soil is prone to loosening and landslides ((Zhou et al. 2021; Liu
Y. et al., 2023). The physical and chemical properties of the lithology can lead to changes in the
permeability of the rock, and lower permeability can disintegrate the stability of the slope (Liu Y.
et al., 2023). Changes in pore pressure due to changes in soil porosity during landslides are related
to the initial porosity, the relative time scales of soil deformation and pore-pressure diffusion (Liu
Y. et al., 2023). The distance to the faults and the distance to the river are considered to be related
to landslide occurrence.

2.1.2.2.Geological hazard factors

Geological disaster factors include the land surface temperature (LST), rainfall, drought degree
index, and desertification degree index (DDI). Land surface temperature is more common in LSM
and the relationship between land surface temperature anomalies and rock slope instability has not
been effectively demonstrated yet. Rainfall is more frequently used in landslide susceptibility
evaluation; the increase of rainfall can put the void water under great pressure (Liu Y. et al.,
2023). Drought degree index is closely related to soil water content, similar to the effect of soil
water content on landslides.

2.1.2.3.Human activity factors

The road construction along the steep slope, mining, and fortification leads to the occurrence of
landslides (Ayalew and Yamagishi 2005; Liu Y., et al., 2023).
2.2. Remote Sensing for Landslide Susceptibility Mapping

many remote sensing data have been used to study landslide processes, including space-borne
synthetic aperture radar (SAR) and optical remote sensing, airborne light detection and ranging
(LiDAR), ground-based SAR and terrestrial LiDAR, incorporating in situ measurements and
observations of environmental factors (Bozzano et al., 2017). In particular: SAR data have been
widely used in landslide research because of their broad coverage and high spatial resolution and
the ability to operate under all weather conditions. Satellite SAR data used include archived ERS
and Envisat ASAR (Bozzano et al., 2017). Optical remote sensing images were mainly applied to
generate landslide inventory, considering long time-series of Landsat TM/ETM, SPOT 1–5,
ASTER, and Rapid Eye (Golovko D et al., 2016). Multi-temporal LiDAR images and ortho photos
can be compared to quantify landscape changes caused by an active landslide (Kamps M 2017).
The ground-based terrestrial laser scanner (TLS) LiDAR can produce highly detailed three
dimensional (3D) images within minutes, allowing the study of 3D surface changes of landslides
(Luo et al 2017). Among the most useful applications derived from the analysis of remote sensing
images is the development of digital terrain models (DEMs), such as those generated from Indian
remote sensing satellite (IRS) P5 images ( Ren Z et al., 2017) and TerraSAR-X/TanDEM-X
images by InSAR [108]. DEM can then be used to assess erosion, landslide, and topographic multi
temporal differences (Ren Z 2017).

2.3. GIS Applications in Landslide Susceptibility Analysis

GIS based spatial data analysis for landslide phenomena and susceptibility mapping which is
carried out in Debresina area of the Afar rift margin (in Ethiopia) and in Rio San Girolamo basin
at the margin of Campidano graben (in Sardinia, Italy). Both of these distant study areas are prone
to various types of landslide and landslide-generated hazards with tremendous damages such as
loss of human lives, failure of infrastructures, and damage on agricultural fields and on the natural
environment. Landslides and related hazards have no geographic boundaries. They occur in both
developed countries (like Italy) and in developing countries (like Ethiopia) causing a continuous
threat to human beings all over the world. Geographic Information System (GIS) is a key tool for
analyzing landslide susceptibility. It can handle large amounts of data, both in terms of file size
and geographical scale. They can create thematic data layers that show factors that contribute to
landslides, such as slope, elevation, rainfall, and land use. GIS can perform statistical analysis to
calculate the cumulative effects of these factors on landslides. They can create landslide
susceptibility maps that show the spatial distribution of potential landslide hazard zones. These
maps can help with land use planning and engineering practices. GIS can perform dynamic and
ongoing landslide susceptibility zonation.

2.4.Factors Affecting Landslide Occurrence and Triggering

Landslide occurrence is influenced by both natural and human factors, with triggers like intense
rainfall, earthquakes, and human activities like deforestation and road construction playing a
significant role.major faults that run parallel to the Main East African Rift (MER) have formed
release surfaces for structurally controlled deep-seated landslides.Most of the large-scale
landslides in Ethiopia have occurred along the MER scarps and also developed in plateau regions
(e.g., Abramson et al. 1996; Ayalew 1999; Temesgen et al. 2001; Ayalew and Yamagishi 2004;
Ayenew and Barbieri 2005; Nyssen et al. 2006; Fubelli et al. 2008, 2013; Moeyersons et al.

Climate

Long-term climatic changes can significantly impact soil stability. A general reduction in
precipitation leads to lowering of water table and reduction in overall weight of soil mass, reduced
solution of materials and less powerful freeze-thaw activity. A significant upsurge in precipitation
or ground saturation would dramatically increase the level of ground water. When sloped areas are
completely saturated with water, landslides can occur. If there is absence of mechanical root
support, the soils start to run off (Sonia Madam 2024).

Earthquakes

Seismic activities have, for a long time, contributed to landslides across the globe. Any moment
tectonic plates move, the soil covering them also moves along. When earthquakes strike areas with
steep slopes, on numerous occasion, the soil slips leading to landslides. In addition, ashen debris
flows instigated by earthquakes could also cause mass soil movement (Sonia Madam 2024).
 Weathering

Weathering is the natural procedure of rock deterioration that leads to weak, landslide-susceptible
materials. Weathering is brought about by the chemical action of water, air, plants and bacteria.
When the rocks are weak enough, they slip away causing landslides (Sonia Madam 2024).

Érosion

Erosion caused by sporadic running water such as streams, rivers, wind, currents, ice and waves
wipes out latent and lateral slope support enabling landslides to occur easily.

Volcanoes

Volcanic eruptions can trigger landslides. If an eruption occurs in a wet condition, the soil will
start to move downhill instigating a landslide. Stratovolcano is a typical example of volcano
responsible for most landslides across the globe (Sonia Madam 2024).

Slope Angle and Material

Steeper slopes and unstable soil or rock formations are more prone to landslides.

Forest fires
Forest fires instigate soil erosion and bring about floods, which might lead to landslides by making
soil exposed to erosion
Gravity
Steeper slopes coupled with gravitational force can trigger a massive landslid related to flat land
surface .
2.4.2. Human causes of landslides
.
Mining
Mining activities that utilize blasting techniques contribute mightily to landslides. Vibrations
emanating from the blasts can weaken soils in other areas susceptible to landslides. The
weakening of soil means a landslide can occur anytime (Sonia Madam 2024).

Clear cutting (Deforestation)

Clear cutting is a technique of timber harvesting that eliminates all old trees from the area. This
technique is dangerous since it decimates the existing mechanical root structure of the area.

2.5. Global and Regional Studies on Landslide Susceptibility Mapping

Global and regional landslide susceptibility mapping studies, using techniques like
remote sensing and GIS, aim to identify areas prone to landslides, aiding in
infrastructure planning and disaster mitigation. These studies analyze various factors
like slope, rainfall, and geology to predict landslide risks. These approaches are very useful
mainly in very large geographical areas where landslides are the most common yet highly
devastating disasters, such as in Nepal, U.S., Philippines, and in many other countries. In the U.S.,
landslides caused 25 to 50 deaths each year, whereas extreme rainfall is the most common cause
of landslides in the Philippines (Talubo et al 2015; Ram et al., 2020; Ray and Jacobs (Ray et al.,
2007) studied landslides in California, U.S., Leyte, Philippines and, Dhading, Nepal. They
established the relationship between landslides, satellite soil moisture (Advanced Microwave
Scanning Radiometer (AMSR-E)), and satellite precipitation (Tropical Rainfall Measuring
Mission (TRMM)). In Nepal, Amatya et al. 2019) used high resolution optical data for landslide
mapping and susceptibility analysis along the Karnali Highway in Nepal. And (O. Kebeba et al.,
2024) used DEM, satellite imagery data to detect landslides in the Maze catchment, omo Valley to
generate a susceptibility map using the frequency ratio and the AHP approaches in an integrated
strategy to map and analyze the area’s vulnerability to landslides. This study intends to
characterize the processes and failure types associated with landslides to provide a more detailed
assessment of the relationships between landslides and their causes. The study also attempts to
produce a detailed landslide susceptibility map by merging data from frequency ratio and AHP
investigations. Mersha and Meten 2020 used DEM data, LiDAR sensors and satellite imagery data
to detect landslide susceptibility mapping and assessment using bivariate statistical methods in
Simada area, northwestern Ethiopia. . In this study, Frequency Ratio (FR) and Weights of
Evidence (WoE) models were applied to evaluate the landslide causative factors and generate
landslide susceptibility maps (LSMs).

2.6. Methodologies for Landslide Risk Assessment Using Remote Sensing and GIS

According to (Shano et al., 2020) several approaches and techniques have been proposed for

landslide risk assessment.These include landslide inventories, heuristic terrain and susceptibility
zoning, statistical methods, deterministic methods, probabilistic methods etc. ((Reichenbach et

al. 2018; Corominas et al. 2014; Raghuvanshi et al. 2014a; Negassa and Kala 2015; Kanungo et

al., 2006; Aleotti and Chowdhury 1999; Guzzetti et al., 1999; Shano et al., 2020). Broadly, all

these techniques or approaches may be further classified into qualitative and quantitative
approaches. These all approaches/methods are used for landslide risk assessment by using GIS

and remote sensing.

There are different methodologies for landslide risk assessment using remote sensing and GIS.

Those are

 Use a variety of techniques to evaluate the landslide inventory, including heuristic

techniques, geomorphological mapping, and physical and statistical modeling.

 Analyze the landslide susceptibility assessment with the temporal distribution, landslide
frequency magnitude relationship, and major triggers.

 Create a landslide risk assessment map that highlights landslide risk hotspots. This map can be
classified into five classes, such as very low, low, moderate, high, and very high.

 Analyze the geospatial database using a spatial multi-criteria evaluation technique.

 Use LiDAR to collect high-resolution terrain data. This can be used to create highresolution
digital elevation models (hrDEMs).

 Use ground-based sensors, such as ground-based interferometric radar, Doppler radar,

 and lidar, to monitor smaller areas. These sensors can be used in real-time monitoring and

 early warning of landslides.

Chapter Three

Methodology and Materials

3.1. Methodology

The methodology followed for Landslide Susceptibility Modeling in the Kulfo River Catchment
Using frequency ratio and GIS Techniques is divided into secondary and primary data collection.

3.1.1. Secondary Data Collection

Before field work collected from base map preparation, image downloading, geo referencing and
projection, topographic map preparation and secondary data collection will be carried out. The
secondary data for this study will be collected from various sources. These include collecting
relevant literatures from published and unpublished papers, DEM data from USGS, a regional
geological map from Geological Survey of Ethiopia at a scale of 1:250,000, rainfall data from
National Metrological Agency of Ethiopia, a topographic map from Ethiopian Geospatial
Information Agency at a scale of 1:50000 and Google Earth image from Google Earth. Elevation,
slope, Normalized Difference Vegetation Index, geology, land use/land cover, distance from the
stream, distance from the road and rainfall map layers will be generated from DEM of resolution
30m that is downloaded from Advanced Space-borne Thermal Emission and Reflection
Radiometer (ASTER). h on-site observations and remote sensing data from Google Earth Pro. The
research area’s geomorphometric properties, including elevation, slope, aspect, and curvature,
were derived from a Digital Elevation Model (DEM) with a spatial resolution of 30 by 30 m. A
cartographic representation of the utilization and coverage of land was generated with sentinel
photos with diverse spectral bands and resolutions.
3.1.2. Primary Data Collection

During the fieldwork, landslide locations will be identified and marked with GPS, land use (land

cover) types around the landslide areas, drainage networks and spring locations, lithological units

and human activities will be investigated to prepare the landslide susceptibility maps. Landslide

inventory mapping on active landslide areas will be measured by their length, width,accumulation
zone and depth (if possible), and Selection of landslide causative factors should be identified
based up on the nature of the study area and the availability of data. After compilation of the
actual field investigation, the data will be systematically processed and analyzed first in ArcGIS
and Frequency ratio followed by Microsoft Excel and then finally in ArcGIS.

3.1.3. Data Analysis methods

3.1.3.1. Frequency ratio method

The frequency ratio technique is a probabilistic model that utilizes the observed correlation
between the pattern of landslide distribution and each element linked to landslides. The frequency
ratio measures the occurrence of landslides in the analyzed area and evaluates the probability of a
landslide occurring compared to the probability of it not occurring, depending on certain criteria
(Bonham-Carter and Bonham-Carter, 1994Bonham-Carter, 1994). A value of 1 represents the
average number of landslides that occur over the whole region. A value over one indicates an
increased likelihood of landslide occurrence and amore pronounced correlation. The frequency
ratio is determined by utilizing (Eq. (1))

FRd = %Ls/ %Am

The frequency ratio for the causative factor class is represented by the term “FRd", the proportion
of landslides in a causative factor class is shown as "%Ls", and the area of the causative factor
class as a fraction of the whole map is described as "%Am". Lower correlation implies a reduced
probability of landslide occurring. The frequency ratios of each component type or class were
summed to create the landslide susceptibility index (LSI), as specified in (Eq. (2))
 n

∑ LR
0

LSI is an acronym for landslide susceptibility index, whereas LR represents the frequency ratio of
every variable type or class. The index will be normalized and partitioned into homogenous zones,
resulting in a map that represents the probability of landslides. A higher LSI value indicates a
higher susceptibility to landslides, while a lower value indicates a reduced sensitivity to landslides.

3.1.4. Method of landslide susceptibility: Remote Sensing and GIS Techniques

In the Kulfo River catchment, landslide susceptibility mapping can be achieved using remote
sensing data and GIS techniques, specifically by mapping thematic layers and utilizing frequency
ratio analysis to evaluate factors influencing landslide susceptibility. it is also essential to have
advanced tools and detailed spatial information to develop an effective landslide susceptibility
map. Digital tools such as GIS and global positioning system (GPS) are mostly used to analyze
spatial data and developing landslide susceptibility and hazard maps. Moreover, remotely sensed
data and technologies are widely used for effective landslide susceptibility mapping, hazard
assessment, and risk assessment, which further helps in awareness, mitigation, and management of
potential landslide threats (Ray et al., 2020).

GIS (Geographic Information Systems) in Landslide Susceptibility

GIS is a powerful tool that is commonly used in the analysis of landslide susceptibility (Lin et al.

2018) . It allows for the integration, visualization, and analysis of spatial data, and enables the
integration and analysis of various data sources, including topographic, geologic, and
meteorological data. It can be used to create maps and models that identify areas prone to
landslides based on various factors such as slope, land use, and geology.

Remote Sensing (RS) in Landslide Susceptibility

RS is powerful tool that is commonly used in the analysis of landslide susceptibility as it provides
high-resolution images and other data from satellite and aerial platforms, and allows for the
mapping and monitoring of changes in the landscape over time (Kim et al. 2002). Remote
sensing provides valuable information on various landslide-related factors, including slope
gradient, aspect, land cover, and geological characteristics. High-resolution satellite imagery and
LiDAR data allow for the extraction of accurate topographic information, which is crucial for

landslide susceptibility analysis. Remote sensing technology is an effective and widely established
analytical method in geological investigations, especially at the inaccessible sites (Kruse and
Dietz, 1991; Al Rawashdeh et al., 2016; Schetselar, 2001).

3.1.5. Landslide susceptibility and Modeling

Geographic Information Systems(GIS) can be used to create susceptibility models by combining

and analyzing data layers to identify patterns and relationships between contributing factors and
landslide occurrences. GIS can also be used to visualize the results, which can help
decisionmakers and emergency management agencies plan for and respond to landslides.

FR model

The FR model is one of the bivariate probability methods, which is applicable to determine the

correlation between landslide occurrence and landslide causative factor classes (Mersha and
Meten 2020). . In this model, processing the input data, computations and output-processes are
very simple and can be easily understood. It is simple and relatively flexible to use and implement
a landslide susceptibility map with accurate results (Lee and Pradhan 2007; Yilmaz 2009; Choi et
al. 2012; Mohammady et al. 2012; Park et al. 2012; Mersha and Meten). The frequency ratio
model is one of the probabilistic models which are based on the observed relationship between the
distribution of landslides and each landslide related factor (Lee and Talib 2005). When the ratio
value is greater than one, it indicates the strong correlation between

factor class and landslide occurrence in a given terrain; however, the ratio value less than one
indicated that weak coloration between landslide occurrence and landslide factors, which means

a low probability of landslide occurrence (Mersha and Meten 2020). It can be calculated using
equation

F=a/b=Nslpix⁄Ntslpix (1)

Ncpix ⁄Ntcpix
Where FR is the frequency ratio, Nslpix is the landslide pixel area in a landslide factor class,

Ntslpix is the total area of a landslide in the entire study area (a), Ncpix is the area of the class in
the study area, and Ntcpix is the total pixel area in the entire study area (b). In this research work,

the FR for each causative factor class is calculated using equation above.

After calculated the FR for each landslide factor class using Microsoft Excel and GIS, the FR

value for each factor class assigned through the join in the ArcGIS tool. The LSI indicated the

degree of susceptibility of the area for landslide occurrence. The LSI of the study area will be

calculated by carefully summing up the weighted factor raster maps using equation below by the

raster calculator in Map Algebra of the spatial analysis tool. To get the LSI, the FR of each factor

type or class is summed as in equation below.

i=1
LSI= ∑ FRiXi (2)
n

Where LSI is the landslide susceptibility index, n is the number of landslide factors, Xi is the

landslide factor, and FRi is the FR of each landslide factor type or classes. After the LSI was

calculated, the index values will be classified into a different level of landslide susceptibility

zones using natural breaks in the ArcGIS tool. The higher the value of the LSI, the higher the

probability of landslide occurrence and but the lower the LSI, the lower the probability of

landslide occurrence. Based on the natural break classification, the landslide susceptibility map

of the study area will have five classes such as very low, low, moderate, high, and very high

landslide susceptibility class.

3.1.6. Validation of Results

Without model validation, landslide susceptibility maps will not be meaningful. As a result,
validation of the predictive model is an important step for landslide susceptibility mapping
(Mersha and Meten 2020). There are various types of validation techniques for landslide
susceptibility maps. In the current study, the performance of the LSMs produced by FR models

that will be evaluated using Area Under the Curve (AUC) and Landslide Density Index (LDI).

Area under the curve (AUC)

The area-under-curve (AUC) method works by creating success rate and prediction rate curves

(Mersha and Meten 2020). Landslide susceptibility maps can be validated by comparing the
susceptibility maps with both the training landslide (70%) and validation landslide (30%). The

success and predictive rate curves can be created for FR model. The success rate curve is based

on the comparison between the predictive model and the training landslide. The predictive rate

curve is based on the comparison between the predicted map and the validation landslide. The
Area Under the Curve (AUC) of the success rate represents the quality of the model to reliably

classify the occurrence of existing landslides whereas the AUC of the predictive rate explains the

capacity of the proposed landslide model for predicting landslide susceptibility (Mersha and

Meten 2020). AUC was calculated by reclassifying LSI into a number of classes with descending

order of the values of pixels in the study area and combined with a landslide inventory. Then the

rate curves will be drawn through the cumulative percentage of both the training and validation

landslide (y-axis) and cumulative area percentage (x-axis).

Landslide density index (LDI)

For validation of the model, landslide pixels which have not been used for constructing the

models are generally considered as the future landslide area (Mersha and Meten 2020). The
landslide density index, which is the ratio between the percentage of landslide pixels and the
percentage of class pixels in each class on landslide susceptibility map, will be used to validate the
model (Mersha and Meten2020)). If the value of the landslide density index is increased from

low to a very high susceptibility classes, then the landslide susceptibility map is considered to be

valid. LDI can be calculated using the formula in equation below and its output will be presented

in Table format. The suitability of any susceptibility map can be validated if more percentages of

landslides occur in the high and very high susceptibility zones as compared to other zones (Mersha
and Meten 2020)).

LD=Percentage of validation landslide pixel ⁄percentage of area pixel (3)

Where LD =Landslide density

Landslide data should lie in either the high or very high susceptibility classes for successful

validation of a LSM (Mersha and Meten 2020). Finally all validated landslide susceptibility is

represented by different diagrams and graph.

Chapter Four

4. Expected Results/output

4.1. Landslide inventory mapping

Due to the fact that it encompasses valleys that contain debris flowsas well as regions of
vulnerability such as banks of streams, cut slopes,and cliffs, the polygon-based inventory was
selected for this study. This was due to the fact that it provides wide coverage of both places
wherelandslides begin and areas that are vulnerable to landslides. Landslides are commonly found
in regions with lineaments or faults, river banks, and steep terrain. In this inquiry, a total of 348
landslides were gathered.The landslide inventory map was subsequently partitioned into two
categories—the training dataset, which included around 70% of the whole dataset, consisted of
244 landslides. The remaining 30% of the dataset, or 104 landslides, formed the testing dataset
sshows the landslide inventory map. A landslide inventory is a thorough compilation of the
geographical distribution and observable indications of landslides, created by on-site observations
or information acquired using remote sensing techniques (Shano et al., 2021).

4.2. Landslide causative factors

Landslides can occur for a variety of reasons, both environmental and human-caused. Field
surveys and in-depth remote sensing image analysis should be conducted to find such elements.
Ten causative factors were chosen after a thorough analysis. These variables include “rainfall,
NDVI, Slope, aspect, curvature, elevation, lithology, distance from streams, distance from
lineaments, and land use land cover” were considerd in this present study.

4.2.1. Land use/land cover (LULC)

LULC component in this study is categorized into five distinct classifications based on the
interpretation of Google Earth imagery. The area comprises forests, settlements, agricultural lands,
barren lands, and bushes . In this present study, the LULC map was created using the ArcGIS
program through the implementation of the supervised classification method. This study entailed
computing the frequency ratio for every variable class by combining the training landslide with 10
influencing factors associated with landslides. As a result, agricultural land has a higher FR value
of 1.93, and settlement areas have a higher FR value of 1.10 when considering land use and land
cover. This outcome suggests a significant likelihood of a landslide happening. The remaining
classes score below one, indicating a low probability of landslide occurrence. The anthropogenic
factor is a contributing element that pushes the slope stability towards the point of sliding. Both
mankind and their surroundings face perpetual peril of destruction. For example, they obliterate
vegetation and disrupt the course of water bodies, resulting in erosion and landslides.Higher
frequency ratio (FR) values for agricultural and settlement regions indicate a higher landslide risk.
This is likely due to farming land disturbance, construction and disturbance in slope, and
inadequate water management, causing soil saturation and plant loss. Human activities and land
use changes degrade slope stability, increasing landslide risk.According to astudy conducted by
Shano et al. (2021), the main element determining rainfall-triggeredlandslides is how land is used
and covered.
Fig.4 a. Landslide inventory, b. LULC, c. Slope, d. Aspect, e. Curvature, f. Lithology, g. Distance to
lineament, h. Distance to stream, i. NDVI, and j. Rainfall.(T. Mulugeta et al.)
4.2.2. Slope

The slope gradients is a pivotal determinant that directly impacts the soil’s moisture level, soil
growth, erosion probability,and slope stability.Slopes were classified into five categories such as

0–5◦,5–12,◦12–30◦, 30–45◦ , and >45 (Fig. c). For the slope degree, FR rises as the slope gradient
increases, as shown in this categorization. This shows that as the gradient of the slope increases in
certain classes, there is a significant likelihood of landslides happening. The frequency ratio values
for the other classes are less than one. In the research region, it was shown that the likelihood of a
landslide occurring rises as the slope gradient increases. According to Sun (2009), this result
supports the idea that the likelihood of a landslide increases with slope gradient. When the slope’s
material is loose and anthropogenic and natural activities have altered the slope, a landslide may
have happened on a moderate slope (Shano et al., 2021).

4.2.3. Aspect

The aspect parameter indicates the slope face direction (0–360◦). It does this by decreasing
sunlight so that the earth may be heated more efficiently, preserving the soil’s moisture content
(Shano et al., 2021). According to (Shano et al., 2021Shano et al., 2020), the aspect was divided
into 10 categories in this study: “flat (− 1◦), north (N), northeast (NE), east (E), southeast (SE),
south (S), southwest (SW), west (W), and northwest (NW)" (Fig. 2d). Curved shadows over an
area result in the accumulation of more rainfall in stored waters, significantly raising the
probability of a landslide (Shano et al., 2021Shano et al., 2020). Aspect analyses show that the
maximum FR value (1.68) is for the south, followed by west (1.52), southwest (1.33), and
southeast (1.27), indicating the maximal probability of landslide occurrences. This means these
classes have a high correlation with landslide occurrences. The other aspect classes have less than
one FR value, indicating less correlation with landslides (Shano et al., 2021).

4.2.4. Elevation

The study area’s height varies between 1169 and 3544 m, as determined using a Digital Height
Model (DEM) generated from data acquired by the Advanced Spaceborne Thermal Emission and
Reflection Radiometer (ASTER). The study categorized elevation into five classes: 1169–1643 m,
1643–2068 m, 2068–2500 m, 2500–2915 m, and 2915–3544 m, using the natural breaks approach
(Firomsa and Abay,2019). These categories were then used to compile landslidetraining data for
landslide analysis. Weathering and erosion will intensify as height increases. Consequently, it is
logical to consider elevation as a fundamental factor that can be altered to manage the landslide
process (Firomsa and Abay, 2019).

4.2.5. Curvature

The curvature of the research area (Fig.e ) was calculated using ArcGIS software and classified
into three categories based on the methodology proposed by Genene and Meten (2021). Fig. 2e
displays the map of the research area’s curvature. When it comes to curvature, the curvature radius
(FR) is higher for a concave slope, measuring 1.26. On the other hand, for a convex slope, the FR
number is 1.08. The presence of a flat curvature suggests a lower value of 0.08. This finding is
consistent with the results reported by (Shano et al., 2021). Concave topography is indicated by
considerable negative values on the curvature map, while a flat surface is characterized by values
close to zero on both sides.

4.2.6. Lithology

Landslides are triggered by a multitude of reasons, one of which is lithology. A landslide will
likely occur if the rock unit is brittle and loose (Shano et al., 2021).in this study, lithology has two
classes according to lithology type. These are basalt and rhyolite (Fig. f). Thus, basalt and
rhyolites are two lithological units that predominantly cover the study area, making up 80% of the
basalt and 20% of the rhyolites, respectively. The research area’s flat terrain and a few low
elevations are covered in rhyolites. For the lithology factor, frequency analysis shows that the

basalt has a high susceptibility to landslides with an FR value of (4.27)

4.2.7. Distance from lineament

Lineaments significantly influence the composition of surface materials and profoundly affect the
ability of the terrain to allow the flow of fluids, ultimately impacting the stability of slopes. The
proximity of specific characteristics to a slope impacts its stability, increasing the probability of
landslides (Shano et al., 2021). The features of the regions were determined based on the ASTER
image, which had a resolution of 30 m. The Euclidean distance technique was utilized within the
ArcGIS system to classify lineaments into five distinct categories based on how close they are to
the lineament: “0–100 m, 100–200 m, 200–300 m, 300–400 m", and greater than 400 m (Fig. g).
The buffering intervals were determined using on-site observation and analysis of Google pictures.
The frequency of landslides often diminishes as the distance away from the lineament rises. The
FR analysis for the distance from the lineament indicates that the ranges of “0–100 m, 100–200 m,
200–300 m, 300–400 m, and >400 m” have varying effects on the occurrence of landslides, with
corresponding values of 1.42, 1.50, 0.56, 0.58, and 0, respectively. This outcome demonstrates
that the FR falls as the distance from the lineament rises.

4.2.8. Distance from stream

The study classified the distance from the stream into five distinct groups: 0–100, 100–200, 200–
300, 300–400, and above 400 (Fig. 2h), as specified by Mersha and Meten (2020). Streams and
rivers can trigger slope instabilities and landslides by eroding the stabilizing geological
components, increasing the moisture content, and elevating the groundwater level (Desalegn et al.,
2022). A drainage system was created by utilizing a shape file that contains stream networks.
Afterwards, the Euclidean distance was calculated using ArcGIS software to determine the
distance to the nearest drainage point. The Euclidean distance method was utilized in the ArcGIS
software to create a themed layer that illustrates the closeness to streams. The layer is categorized
into five groups based on elevation: 0–100 m, 100–200 m, 200–300 m, 300–400 m, and above 400
m. The frequency ratio for a distance to stream category spanning from 0 to 100 m was more than
1, indicating the this category had the highest likelihood of experiencing a landslide. The FR
values for distances of 0–100 m, 100–200 m, 200–300 m, 300–400 m, and over 400 m from the
stream are 1.89, 0.68, 0.50, 0.09, and 0.01 respectively.

4.2.9. Normalized difference vegetation index (NDVI)

NDVI helps in evaluating the density and distribution of vegetation in an area. Vegetation
stabilizes slopes and prevents landslides by binding the soil with its roots. Areas with dense
vegetation cover are generally more stable and less prone to landslides (Desalegn et al., 2022;
Shano et al., 2021). NDVI provides a quantitative evaluation of vegetation health, allowing
researchers to analyse the quantity of vegetation and its impact on slope stability (Desalegn et al.,
2022; Shano et al., 2021). NDVI can help identify areas with low or sparse vegetation cover
(Desalegn et al., 2022; Shano et al., 2021). Such areas are more susceptible to landslides as they
lack the stabilizing effect of vegetation. Mapping areas with low NDVI values can pinpoint
regions where landslide hazards are more likely to occur (Desalegn et al., 2022; Shano et al.,
2021). NDVI can be used to monitor changes in vegetation cover over time. Sudden decreases in
NDVI values may indicate changes in vegetation health, such as deforestation, land degradation,
or vegetation loss due to natural or human-induced factors (Desalegn et al., 2022; Shano et al.,
2021).

4.2.10. Rainfall

PERSIANN, also known as Precipitation Estimation from Remotely Sensed Information using
Artificial Neural Networks, is a rainfall estimation method that relies on satellite data and artificial
neural networks (ANNs) to estimate precipitation accurately. This system aims to offer

detailed data on rainfall patterns worldwide, focusing on high-quality resolution (Ayalew, 1999;
Desalegn et al., 2022; Shano et al., 2021; Oyda et al., 2024). The current study involved retrieving
PERSIANN precipitation data in a GeoTIFF format from 2021 to 2023. Subse quently, average
values were computed for the study region. The PERSIANN rainfall data was utilized in the
following recent research to map landslide susceptibility (Desalegn et al., 2022; Shano et al.,
2021). Fig. j shows the rainfall map of the study area. The study region’s rainfall and categorized
it into five distinct ranges, as presented in Table 1. Each range was assigned a specific FR value
based on their sequence of FR values. The rainfall classes of 699–767 mm/yr, 768–828 mm/yr,
and 829–896 mm/yr had FR values of 1.32, 4.44, and 1.86, correspondingly. This result indicates
that these classes have substantial correlations with the occurrence of landslides. The remaining
categories of rainfall have FR levels below one, showing a poor correlation with the occurrence of
landslides.

4.3. Results of frequency ratio analysis

4.3.1. Relationship between landslide and causative factors

This study utilized the statistical frequency ratio to investigate the correlations between 9 potential
factors contributing to landslides (O. Kebede et al., 2024). The causal factors were categorized
into fifty-eight categories (Table below), determining the weights given to each class of the FR
model. The weight values derived from the FR model demonstrate the spatial correlation between
the causative factor types and the occurrence of landslides. This study investigated the correlation
between landslides and various parameters such as “slope, aspect, curvature, distance to stream,
distance to road, distance to fault, lithology, land use/ land cover, and NDVI.
Landslide susceptibility mapping

The study area was classified into five susceptibility groups, namely very low, low, moderate,
high, and extremely high, using the frequency ratio approach. This study investigated and linked
several regions that are susceptible to landslides, referred to as hotspots, and the elements that
contribute to their occurrence. The susceptibility study takes into account several elements, such
as geology (specifically lithology), topographic features (including “slope gradient, aspect,
curvature, and elevation), proximity to lineaments and streams, land use/land cover, NDVI
(Normalized Difference Vegetation Index”), and rainfall parameters. The study area’s
susceptibility thresholds were based on the 10 selected conditioning parameters. High-frequency
ratio (FR) values indicated landslide susceptibility, and these values were used to assign pixels to
each class. The classification criteria were validated and refined by expert judgment and field
observations.The class with very low susceptibility includes 54,119 pixels, which accounts for
roughly 12.47% of the entire area analyzed. High Resistance: This class covers an area of 79,596
pixels, which is approximately 18.35% of the total area. Moderate Susceptibility: It occupies the
greatest percentage, accounting for roughly 32.43% of the area, with a total of 140,689 pixels. The
High Susceptibility class consists of 94,338 pixels, which accounts for approximately 21.74% of
the entire area. The susceptibility is quite high, with a coverage of 65,113 pixels, accounting for
15.00% of the whole area displays the map indicating the likelihood of landslides occurring.
Analyzing the spatial relationship between each factor class and landslide using the frequency ratiomodel( (Desalegn
et al., 2022; Shano et al., 2021))
 Chapter five

5. Work plan

No. Tasks 2025 2025 2025 2025 2025 2025

1. Proposal preparation and presentation March April May June July Aug
2. Literature review
3. Progress meeting
4. Secondary data collection
5. Secondary data analysis
6. Progress meeting
7. Primary data collection
Data analysis & modeling
Progress meeting
8. Interpretation of results
Draft report preparation
9. Progress meeting
10. Final report preparation
11. Presentation of final report
 Chapter Six

6. Budget of the Project

6.1 Personnel

S. No Items Amount required((pieces / Unit cost Subtotal

number/packs / litter) (Birr/unit) cost (Birr)

1 Daily laborer during 1 person x 13days 200/day 2600

field work

2 Per diem for advisor 1 person * 2days 2000/days 4000

3 Draft Color print (20 pages)x3copies 10.00/page 600.00

4 Draft Normal print (50 pages)x3copies 5.00/page 750.00

5 Final Color print (23pages)x3copies 10.00/page 690.00

6 Final Normal print (50 pages)x3copies 5.00/page 750.00

7 Photocopies 150 pages 5.00/page 750.00

Total 10,140

6.2 Supplies
S. No Items Amount required((pieces / Unit cost Subtotal cost
number/packs / litter) (Birr/unit) (Birr)

1 Note book 1 piece 200/piece 200

2 Paper 2pack 200/pack 400

3 Pen 1 pack 200/pack 200

4 Pencil 1 pack 50/pack 50

5 Marker 1 pack 150/pack 150

6 Topographic maps 5 sheets 200/sheet 1000

7 Landsat image 2 sheets 200/sheet 400

8 Satellites Image 4 images 500/ image 2000

Total 4,400

6.3 Transportation and related cost

 S. Items Amount Unit cost Subtotal cost
No required(numbers) (Birr/unit) (Birr)

1 Travel for field For 2 days, two round 1000/round 1000

visits/overview

2 Travel for field For 7 days, two round 1000/round 2000

mapping

Total 3000

6.4 Cost summary

S. No Items Subtotal cost(birr)

1 Personnel cost 10,140

2 Supplies 4,400

4 Transportation cost 3000

5 Unexpected expenses 5000

Grand total 22,540

Reference
1. Asmare, D., Tesfa, C., Zewdie, M.M.A., 2023. GIS-
based landslide susceptibility assessment and mapping
around the Aba Libanos area, Northwestern Ethiopia.
Appl Geomat 15, 265–280.
https://doi.org/10.1007/s12518-023-00499-7.
2. Ayenew, T., Barbieri, G., 2005. Inventory of landslides and
susceptibility mapping in the Dessie area, northern
Ethiopia. J. Eng. Geol. 77, 1–15. https://doi.org/10.1016/J.
ENGGEO.2004.07.002.
3. Getnet M. and Filagot M. Assessment of Landslide
Risk in Ethiopia: Distributions, Causes, and Impacts
DOI: http://dx.doi.org/10.5772/intechopen.101023.
4. Hamza, T., and Raghuvanshi T.K., 2016. GIS based
landslide hazard evaluation and zonation. A case from Jeldu
District, Central Ethiopia. Journal of King Saud University
– Science 29, 151–165.
5. Irjesh S., Jayant N., Swarnim. 2022. Remote sensing and
GIS-based landslide susceptibility mapping using
frequency ratio method in Sikkim Himalaya. Quaternary
Science Advances, 8, 2-12.
6. Jiao W., Ya Zh., Liu Y., Yongxian Zh., Jun L., Meixia Zh.,
Guorui Ma. 2023. Identifying the essential infuencing
factors of landslide susceptibility models based on
hybrid-optimized machine learning with diferent grid
resolutions: a case of Sino-Pakistani Karakorum Highway.
Environmental Science and Pollution Research.100676, 5-7.
7. Kumar, V., Gupta, V., Jamir, I., 2018. Hazard evaluation of
progressive Pawari landslide zone, Satluj valley, Himachal
Pradesh, India. Nat. Hazards 93 (2), 1029–1047.
https://doi.org/10.1007/s11069-018-3339-3.
8. Landslide Studies Division GHRM Centre, Geological
Survey of India Practical Tutorial #2 Updated May 2018.
9. Lemma T., Abera U·, Thomas T. 2024. Landslide vulnerability mapping
using
multi-criteria decision-making approaches: in Gacho
Babba District, Gamo Highlands Southern Ethiopia.
Discover Applied Sciences. |
https://doi.org/10.1007/s42452-024- 05693-9 3-24.

10. Leulalem Sh. . T. R . Matebie M. 2021. Landslide Hazard

Zonation using Logistic Regression Model: The Case of
Shafe and Baso Catchments, Gamo Highland, Southern
Ethiopia. Geotechnical and Geological Engineering. 4-20.
11. Leulalem Sh., Bisirat G., Wondwosen J., Desta E. ,
Tigabu B., Tariku D., Geberemedin Ch., Zeleke D.,
Muralitharan J. 2024. Fatal landslides in Kencho, Shacha
& Gozdi villages, Gofa zone, Ethiopia: A detailed
investigation (Geological, Geotechnical, geophysical &
geospatial) of the July 22, 2024 catastrophe and its
socioeconomic repercussions. Quaternary Science
Advances.16, 2-13.
12. Leulalem Sh., Tarun K. and Matebie M. 2020. Landslide
susceptibility evaluation and hazard zonation techniques
– a review Geoenvironmental Disasters. 7, 2-19.
13. Likinaw M., Assayew N., Muralitharan J., Biniyam T.
2024. Landslide susceptibility assessment in Addi
Arkay, Ethiopia using GIS, remote sensing, and AHP.
Quaternary Science Advances. 15, 1-10.
 14. Lulseged Ayalew, 1999. The effect of seasonal rainfall on
landslides in the highlands of Ethiopia. Bull. Engineering
geology Environ 58: 9 -19.
15. Lynn M. 2008. The Landslide Handbook A Guide to
Understanding Landslides. United States Geological
Survey, and Peter Bobrowsky, Geological Survey of
Canada. 30-34.
16. Mandler, B. 2017. Investigating Volcanic Landslide
Hazards Modern tools improve risk assessment for
volcanic landslides and debris flows.
Americangeosciences.org/critical- issues.
17. Mesfin M. Abebe S. 2020. Remote Sensing and GIS Based
Potential Landslide Hazard Zonation in Ambo Woreda:
Central Ethiopia. Journal of Environment and Earth
Science. 10, 19-25.
18. Michael V., Ryan L., John G. 2023. Landslide susceptibility
modeling derived from remote sensing, multi-criteria
decision analysis, and GIS techniques: A case study in the
Southeast Bohol Province, Philippines.Research square. 4-
36.
Azemeraw W. 2020. Modeling of Landslide susceptibility in
a part of Abay Basin, northwestern Ethiopia. Open
Geoscience.12, 4-26.

19. Obse K., Leulalem Sh., Yadeta Ch., Muralitharan J.

2024. Integration of geospatial analysis, frequency ratio,
and analytical hierarchy process for landslide
susceptibility
20.Abay, A., Barbieri, G., 2012. Landslide susceptibility and causative factors
evaluation of

the landslide area of Debresina, in the southwestern Afar Escarpment, Ethiopia.

Journal of Earth Science and Engineering 2, 133–144.