Luận văn thạc sĩ: Landslides susceptibility assessment using GIS and remote sensing method: A case study in mountainous area of Nghe An province, Vietnam

THUY LOI UNIVERSITYMASTER THESIS LANDSLIDES SUSCEPTIBILITY ASSESSMENT USING GIS AND REMOTE SENSING METHOD: A CASE STUDY IN MOUNTAINOUS AREA OF NGHE AN PROVINCE, VIETNAM Participant: Nguy

THUY LOI UNIVERSITY

MASTER THESIS

LANDSLIDES SUSCEPTIBILITY ASSESSMENT USING GIS AND REMOTE

SENSING METHOD: A CASE STUDY IN MOUNTAINOUS AREA OF

NGHE AN PROVINCE, VIETNAM

Participant: Nguyen Hoang Ninh Class: Niche V

Student code: 1481440225004

Major codes:

SUPERVISORS:

1.Supervisor 1: Dr Ngo Le An

2 Supervisor 2: Dr Pham Thi Thanh Nga

Department : Faculty of Hydrology and Water Resources

Ha Noi, 2017

First of all, I would like to give my deepest appreciation to my supervisors Dr.Ngo Le An and Dr Pham Thi hanh Nga Without their valuable guidance,suggestions and comments, it could be impossible to complete my thesis,

1 also would like to express my sincere thanks to “Investigation, assessment and

‘warning zonation of landslides in mountainous regions of Vietnam” project for kindly

providing me all the essential landslide inventory data, as well as technical supports infield experiment and data processing I would like to extend my sincere thanks to

et Management Center for funding me to finish my thesis

Special thanks to all my colleagues of Vietnam Institute of Geosciences and

‘Mineral Resources for their continuous support and encouragement to overcome thedifficulty in study times

also wish to give my sincere thanks to my family, especially to my wife and

my son, for their love and inspiration they have been giving and unconditionalassistance through the whole period of studying

Without the help from all of you, this study would have been impossible

Nguyen Hoang Ninh

Thereby declare that is the research work by myself under the supervisions of Dr Ngo

Le An and Dr Pham Thi Thanh Nga The results and conclusions of the thesis arefidelity, which are not copied from any sources and any forms The reference

the thesis has cited and recorded as prescribed, Thedocuments relevant sour

results of my thesis have not been published by me to any courses or any awards,

Nghe An Province located in North Central Vietnam where is the beginning of TruongSon range, The topography of Nghe An is very complex and strongly dissected by

Truong § n range of mountain and river systems Besides the landform factor,

from the East Sea are blocked by the Truong Son mountain range, c ising heavy rains

in the province These factors result in the frequent occurrence of landslide, debrisflow and flashflood in Nghe An

RUSLE model is used to predict the average annual rate of erosion But erosion notonly Loses soil but also forms unstable slope lead to increasing of the development ofthe mass movement, Moreover, this model based on four main factors: land use,rainfall, topography, and soil property These factors are also taken into account in thecurrent landslide assessment methods Therefore, the potential soil erosion map seems

to affect the landslide susceptibility assessment By combining Remote Sensinganalysis and Geo Information System analysis to assess the landslide susceptibility, theresult shows that the area very high and high susceptible for landslide cover 26% and25%, respectively These rest ts reflect the real occurring in Nghe An province Thisresull is useful for administrators, decision maker, ind planner

Revised Universal Soil Loss Equation

Geographical Information System

Remote Sensing

Geological Hazard Information System

Spatial Muli Criteria Evaluation

Analytical Hierarchy Process

Digital Elevation Model

‘The Shuttle Radar Topography Mission

United States Geological Survey

ULL.L Landslide Process

TLL.2 Landslide types and triggering factors

112 Apply Remote Sensing and GIS in landslide susceptibility assessment

112.1, Remote Sensing

112.2 GIS

CHAPTER IIL

‘Study area and data collection

TIL Study area

TILL.1, Natural features related to landslide

TL 1.2 Human activities related to landslide

IIL2 Data Collection

IV.1, Rainfall factor (R)

IV.2 Soil Erodibility factor (K)

IV.3 Topography Factors (LS)

IV.4, Crop and Management Factor (C)

IVS Conservation Practice Factor (P)

IV.6, Correlation between soil erosion map and landslide occurrence

IV7 Methodology Flowchart

CHAPTER V

Results

VAL RS analysis

VAL Lansat 8 pre-proc sing

V1.2, Normalized Difference Vegetation Index:

V.1.3 Transformed Value Range

30

31

V2 GIS Analysis

V2.1 R factor

V.22 K factor

V.23.LS factor

V.24, Average Annual Soil Loss

V3, Verification of Erosion Soil Loss Map

V4, Distribution of the susceptibility zones for landslide

V4.1 The Very high susceptibility zone for landslide

‘V4.2 The high susceptibility zone for landslide

‘V4.3 The average susceptibility zone for landslide

V4.4, The low susceptibility zone for landslide

LIST OF FIGURESFigure 1: The tracks of all ropical cyclones in the Northwest Pacific Ocean from 1980t0 2005, 1

Figure 2: Statistic of topical storm affected to Vietnam from 1959 o 2009 2

Figure 3: The damage curves due to nature disaster in Vietnam 2

Figure 4: The relationship between the force and the slopes 5

Figure 5: GIS processing for landslide suseepibilty assessment 9

Figure 6: Administrative map of Nghe An province "

Figure 7: The landslide on 7h highway Is

Figure 8: The soil erodibility nomograph ma

Figure 9: Methodology of research 28

Figure 10; Flow chart of Remote Seasing analysis to caleulate C fator 2»

Figure 12: NDVI map 31

Figure 13: factor map 2

Figure 14: Flowchar of GIS analysis to calculate R factor 3B

Figure 15:R factor map a

Figure 16: Flow chart of GIS analysis to calculate K factor 38

Figure 17: K factor map mr

Figure 18: Flow chat of GIS analysis to calculate LS factor 38

Figure 19: LS factor map, 39

Figure 20: Average Annual Soil Loss Map 40

Figure 21: Erosion Intensity map based on Vietnam standard for soil quality in 1995(TCVN 5299-1995) with landslides distribution 42

Figure 22: Erosion Intensity map based on Vietnam standard for soil quality in 2009(TCVN 5299-2009) with landslides distribution 43

Figure 23: Graphs of frequency ratio analysis 4

Figure 24: The statistic graph of the distribution of landslide susceptibility zone in all

districts of Nghe An province 46

LIST OF TABLES

‘Table I: Type and attributes of landslide 6

‘Table 2: The distribution of the slope in Nghe An province 12

‘Table 3: The percentage of forest cover in Nghe An province from 2008 to 2011 14

‘Table 4: Information of used Landsat 8 images 16

‘Table 7: Classification of soil erosion caused by rain based on Vietnam Standards for

CHAPTER I INTRODUCTION

1.1 General

Landslides are reported as one of the most dangerous natural disaster in the world [1]

Annually, landslides are responsible for a lot of damage to human and monetarylosses Vietnam located in the region with high incidence of storm Intensive rainformed by storm is the trigger factor causing a lot of landslide in the mountainous area,

‘Together with this, human activities such as deforestation, mining, construction andtransport works promote physical landslide processes The: result in the frequent

‘occurrence of landslides in the mountainous area of Vietnam,

Figure 1: The tracks of all tropical eyclones in the Northwest Pacific Ocean from 1980

10 2005 (Source: Wikipedia)

1

= The wopical storm lended on Vietnam

‘= The woplcal storm in East sea ‘= The topical storm affected to Vietnam

Figure 2: Statistic of tropical storm affected to Vietnam from 1959 to 2009 (Source.Vietnam's Institute of Meteorology, Hydrology, and Environment/2010)

‘These landslides would not only damage infrastructures, and in social and economicaspects but also kill a human being In the World Bank report, landslide and other

natural disasters have resulted in a total loss of life of 13,035 people, with damage to

residential housing, agriculture, and infrastructure valued at VND 91 trillion (USS 6.4

billion)[2] As predicted in this report, the damage and loss will be continued to

increase because of negative climate change in the future Thus, research on landslidedisaster management and risk mitigation is a topic to be considered,

Figure 3: The damage curves due to nature disaster in Vietnam (Source: WB/2010)

2

Nghe An Province located in North Central Vietnam where is the beginning of Truong

‘Son range, The topography of Nghe An is very complex and strongly dissected by

‘Truong Son range of mountain and river systems Besides the landform factor, stormsfrom the East Sea are blocked by the Truong Son mountain range, causing heavy rains

in the province These factors result in the frequent occurrence of the landslide, debrisflow and flash flood in Nghe An,

In order to mit landslide risk, the first step is the zonation of the differentlandslide susceptibility area, Landslide risk and damage assessment is next step toevaluate influences on community and eco-social activities Finally, measures of riskmitigation are determined by cost and benefit method The process including the threesteps was popularly applied in the world This research is focused on landslidesusceptibility zoning to find out the high potential landslide hazard in Nghe An

province based on Remote Sensing and GIS technology.

‘Many factors trigger slope failure including exogenic, endogenic and man-madefeatures[1| But in this study, the focused triggering factor is heavy precipitation.Because it is ma triggering factor caused landslide in Vietnam, especially Nghe Anprovince To evaluate landslide triggered by rain, there are several researchers whichapplied Revised Universal Soil Loss Equation (RUSLE)[3}{6] RUSLE model wasdeveloped in 1960 and improved in 1978 by Wischmeier and Smith of the UnitedStates Department of Agriculture At RUSLE model is used to predict the

average annual rate of erosion, But erosion not only loses soil but also forms unstableslope lead to increasing of the development of mass movement{6} Moreover, thismodel based on four main factors: land use, rainfall, topography, and soil property

‘These factors are also taken into account in the current landslide assessment methods

‘Therefore, the potential soil erosion map seems to affect the landslide susceptibility

assessment,

1.2 Objective of study

‘The overall objective of this study is to evaluate the application of the RUSLE model

to zoning of area susceptible to landslide

Besides that, the specific objectives of this study include

1 Select and parameterize the factors related with RUSLE model

2 "Assess soil erosion and generate the potential soil erosion map

3 Select statistic analysis method to correlate with landslide events/locations inthe reality to verify the accuracy of map

13 Research question

Based on the objectives of this study, the research questions are as following:

1 Do we have the correlation between the potential soil erosion map and realitylandslide location?

2 Can the potential soil erosion map use as the landslide susceptibility map?

CHAPTER II LITERATURE REVIEW

ILL Landslide mechanism

I11.1 Landslide Process

I is assumed that the entire terrain surface is a combination of different slopes Onthose terrains, surface gravity tends to shift any material down to the center of the

earth, Therefore, the material on a certain slope is affected by the following forces:

Figure 4: The relationship between the force and the slopes (M represents the material

‘on the slopes, with mass is m)

P=m gis gravity force

is the driving force or the tractive force of M mass This is the main force causinglandslides because it tends to pull mass to the foot of slopes

P; is the normal force or the pressure of M mass to the sliding surface

P, is the resisting force or the frictional force caused by friction between the ssurface and M

‘Through forces analysis, landslides occur when the driving forces equal or exceed the

resisting forces, Factors causing landslides fall into two categories:

1) Factors increasing driving forces: seismic shaking, inereasing the steep of slope

and incre wing the weight of the mass

ing resisting forces: rain (adding water to the slope cause

ire and reduce frictional strength), weathering (reduce

material hardness) and endogenous process (increase the amount of joint)

Landslide types and triggering factors

Based on the type of movement and the type of material, landslides ean be divided into

the different types According to Highland et al [7], the material of landslides is either

idslides shiftsoil or rock or both of them The type of movement describes how the la

such as; fall, rotation, translation and so on, The following table shows the movementtypes of landslide and the triggering factors causing each type

Table 1: Type and attributes of landslide

‘Type | Material Velocity ‘Triggering factors

Fall Rock Rapid, extremely Earthquake shaking,

rapid undercutting of slope, excavation

Topple Rock, soil | From extremely Different weathering,

slow to extremely excavation, stream - erosion,rapid excavation

Rotation | Rock, soil | From extremely Intense rainfall or - mpid

slow to rapid snowmelt

‘Translation | Rock, soil | From — slow to Intense rainfall, rise of ground

extremely rapid water, undercutting

Spreads | Soil Slow to rapid Liquefaction of lower weak layer

by earthquake shaking,Saturation of underlying weakerlayer due to rainfall, snowmelt,and (or) ground-water changes

Flow Rock, soil | Rapid to extreme intense surface-water flow due to

(debris rapid heavy precipitation or rapid flow) snowmelt combine with

previously occurred landslide

Debris Rock, soil | Rapid to extreme | Volcanic earthquake, earthquake

avalanche rapid shake, weathered slopes collapse

and move down with rapid

velocity

Earthflow | Soil Slow to very rapid | Rainfall, snowmelt

1.2 Apply Remote Sensing and GIS in landslide susceptibility assessment

In order to assess landslide hazards, L Highland et al [7] found 2 types of evaluation:direct observation and the use of technical tools In which, using technical tools areactually derived from a theory: “The past is the key to the future" It means that thepast landslide area has the same physical sitations such as geological, geomorphic.land cover, hydrologic situations with the future landslide area, Based on this, experts,

apply techn cal tools to combine these situations to find the landslide potential areas.Remote Sensing and Geo Information System are currently tools used most inlandslide hazards assessment Therefore, there are many researches to generatelandslide susceptibility map using RS and GIS approach [8]-[15]

1.2.1 Remote Sensing

In all reviewed article, the researchers applied Remote Sensing in two ways: landcovers/land uses classifying and landslides mapping, But data and methods used in thestudies are variety

In his research, Effat [9] used SPOT-4 data to classify the study area to 5 different

classes, Five land use classes included bare land, new cities, cities and village,cultivated land and natural vegetation He showed that the more vegetation cover, themore stable the area is, Therefore, the bare land was presented the highest landslidesusceptibility area

‘Saro Lee [3] applied an unsupervised classification method to class from Landsat TMimage to land use groups To interpret landslide masses places, he used air photos thatcover the 3-year period from 1996 to 1999 The Iandslide masses places weredetermined by identifying a break in the forest canopy, bare soil or geomorphiccharacteristic of scars and body of landslide masses They were collected to check thesoil loss map against landslide hazard map,

For the historic landslide inventory, Legorreta [11] collected two sets of aerial

‘orthophotographs that cover from 1994 ata scale 1:20.000 to 2008 at a scale 1:10.00,

and eight variable to analyze the features of each landslide mass Combining with thefield survey, a hundred and s fen landslides were mapped and classified into 6different types: shallow landslides, debris flows, debris slides, deep-seated landslides,earthflows, and rock falls Based on the parameters associated with landslide, eachlandslide was attributed such as landslide size, landslide activity, landslide parts (headscarp, evacuation zone, and deposit) ete

Like other researchers, Magliulo {15] carried out the establishment of landslide

inventory map based on the field survey and geomorphic analysis Each landside wasinterpreted by the analy s from aerial photographs and 1:5000 scale topographie mapsand were divided into landslide detachment zone and landslide body Besides, the

‘aerial photographs were used to create land use map

In addition, many researchers are using radar to study the landslide But within theframework of the subject, to getting radar images is very difficult Thus this study wasfocused on the methods that used free source image, such as Landsat image

1122 GIS

Revently, GIS is increasingly developed and played an important role in the field of

prevention and mitigation of damage caused by natural disasters in general and inparticular landslide hazards, An incomparable advantage of GIS isa tool that manages,analysis, process and modeling spatial data, Additionally, the different types of datacan be shown together to easily find information such as the distribution and properties

of the landslide or the other elements related to the landslide The application of GIS

for evaluation of landslides are management and modeling

‘¢ Management refer to data digitalization, storage, recovery and visualization

* Modeling refers to the analytical capacity of GIS applications Recently, GI

applications can be used to perform complex analysis based on the different

‘mathematic models and determine the area prone to landslides or the volume oflandslide masses Therefore, the output information of GIS application can help

the planner or administrator make a decision.

‘Triggering di Factor Inventory

Lee Saro is one of the experts who used GIS and Remote Sensing to lands idesassessment He proposed and developed GIS-based geological hazard informationsystem (HS) and applied it for landslide analysis by using the probability-frequeneymethod [10] The spatial relationship between the location of landslides and eachlandslide-related factor was derived The factors such as altitude, slope, aspect andcurvature from the topographic database, soil texture, material, drainage, effectivethickness, and toposraphy from the soil database, forest type, forest diameter, and

forest density from the forest map, and land cover data from Landsat TM image, were

jc hazard in GIS, Acharyaextracted from database of GHIS In order to assess landsli

et al [I4] thought that landslides occur because of unstable slope They, therefore, usedtopography, land soil types and rainfall as thematic maps to calculate the slopestability map in three different soil conditions (dry soil, half-saturated soil, and fullsaturated soil, In the research of Ardesen’s land slides susceptibility, Yalcin & Buluthave combined field wip and aerial photo interpretation to determine location and area

of landslides{13) ‘Then they created the landslide inventory map to show the areadistribution and properties of the landslide It is important to find the relationshipbetween landslides and factor layers related landslide

Besides the authors applied different mathematical models in their research, A range

of methods currently available in the literature for the preparation of landslidesusceptibility maps, For example, in the study of Ardesen’s landslides susceplibility,the factor layers were assigned the weight and calculated in Analyt al Hierarchy

Proce s (AHP) method to mapping landslides susceptibility [13] Otherwise, Themethod, such as Spatial Multi-Criteria Evaluation (SMCE) [9], is similar to AHP using

\wise-pair comparison Other approaches, such as Multiple Logistic Regression [11] orUniversal Soil Loss Equation (USLE)I3-I6] is very complex and need the detaileddata, Each method has its advantage and weakness Thus, need to considerationcarefully before choosing to apply for particular area study

10

CHAPTER II STUDY AREA AND DATA COLLECTION

TILL Study area

Nghề An province is located in the center of North Central, cover an area of 16.487

kmỂ, from 18°33" to 20°01" north latitude, from 103'52' to 105748" east longitude It

\vas bordered by Thanh Hoa province to the north, Ha Tinh province to the south andLaos to the west Nghe An divided into 21 district-level administrative units, including

1 city, 3 towns and 17 districts In which, 11 districts are inthe mountainous area of[Nghe An, including the following districts: Anh Son, Con Cuong, Do Luong, Ky Son,Nghĩa Dan, Que Phong, Quy Chau, Quy Hop, Tan Ky Thanh Chong and Tuong

Duong

"

THL.1.1 Natural features related to landslide

THL1.1.1 Topography features

ue to located in Truong Son range where Eastern slopes of the mountain range reach

4o Gulf of Tonkin, the topography of Nghe An is very strongly dissected by TruongSon range of mountain and river systems The common feature of terrain in Nghe Anprovince is the mountain in the west and the narrow plain in the east Generally, thetopography divided into 3 main parts: the highland is in Northwest part, midland in thecenter and plane to Southeast Nghe An has 2 main river system: Ca river and Hiewriver Slope land greater than 8% cover 80% of the area of the province, especially

over 38% of itis steep greater than 25%, This feature of topography increases the

erosion and landslide in Nghe An

Table 2: The distribution of the slope in Nghe An province

Slope Area (km) % Area

(groups of rock and soil) which the basic composition as following.

‘Alumosilicate enriched metamorphic rock group: distributing in Que Phongdistrict, intersected by Ban Chieng, Dai Loc magma complexes and covered

by MuongHinh formation The composition contains biotite quart schist,mixed quartzite two mica schist; light brown-gray colored, moderate beddedand strongly metamorphism in Bu Khang formation, This group exposes inlarge area in Chau Hank, Chau Hoi, Chau Binh town, Quy Chau district,

2

with the composition of sandstone, siltstone, sericited schist, clayed s

moderate to low bedded mixing together, in some area sandstone and,siltstone set with tens of meter thickness mixing with clayed schist, sericitedlay from one to some meters thickness of Song Ca formation,

Alumosilicate enriched terrigene sediment group: Dong Trau (Tad), QuyLang (T;gl), Dong Do (Ty

distributing as NW-SE lengthen large ribbon in the north of Quy Chau

rid) formation cover by 25% of study area,

istrict, large distributed in Nghĩa Dan district, the composition of

sandstone, gritstone, silty sandstone, moderate to low bedded schist mixingwith erupted acid lens Muong Hình (nh) formation large distributes in thenorth — northeast of Que Phong and Quy Chau district, with the exposed

areas of 600 km", the composition contains conglomerate, sandstone

siltstone, ryolite, porphyr ryolite and its tuf

‘Moderate acidity intrusive rock group: the composition of Dai Loc (G/D dl)and Ban Chieng (G/Ebe) complex’s grabo to granite, forming diversdimension masses The compostion contains gneis-formed granite, weakformed plagiogranite, porphyrited granitebiotte, fluorite containedporphynied alkali felspate granite, It large distributes in Que Phong district,

it is unavailalbe in Quy Chau and Nghia Dan district

Carbonate rock group: The formations of carbonate terrigene sedimentandcarbonate belong to above two formations distributing as plot shaped in QuyChau, Que Phong (small plot) district The composition contains clayedlimestone, silie schist, clayed schist (La Khe formation), thích to moderatebedded limestone (Bac Son formation)

Looxe sediment rock: The Quaternary sediment distributes as uncontinousnarrow ribbon along Hieu, Nam Viec, Quang river, in the south of Quy

Chau and Nghia Dan istrict containing the first and second terrace, alluvial

flat and river bed sediment, The sediment composition contains block,

B

pebbles, grit sand, clay, silt; weak cementation, easy collapse, especially inalluvial and first terrace s ment

IIL1.1.3 Meteorological features

‘The study area is located in the monsoon tropical climate and is divided into twodistinct seasons, Climatic characteristics are generally very harsh Average annualprecipitation is 1600-2000 mm, concentrated in the rainy season A number of rainydays in the year is 150-160 days Annual average humid: 85-86%

‘The rainy season from May to November: The rainy season accounted for > 70% ofthe total annual rainfall The average temperature during the rainy season is from 23,6

to 274°C, sometimes reaches approximately 41°C.

The dry season is from December to April next year and goes with the NE monsoon

pulled down low temperatures, averaging 16-19°C, sometimes down to November 36C

THỊ 1.1.4 Land cover features

Overall vegetation covers in a large area However, forest area is recently decreased

‘Most of natural jungle has been cleared for logging and for cultivation, The primevalforest only remains around the high mountains where border to Laos or in the Pu MatNational Park, Much of the remaining forest area only reed or small tree or manyplaces are bare

Table 3: The percentage of forest cover in Nghe An province from 2008 to 2011

Year | Natural forest (ha) | Artificial forest (ha)| Total (ha) | Cover (%)

11.1.2 Human activities related to landslide

Human activities have a strong impact on the environment It causes the failures orpromotes the movement arising or make the generation more complex Of which, theexploration, traffic and house construction, deforestation and land use changing are the

‘most significant,

ML1.2.1 Traffic and house construction

48" highway (access Nghe An to Thong Thu international border gate) and 7°

highway (access Westem of Nghe An to Nam Can border gate) are currently upgraded

and expanded by cutting to the slope The applying this construction method in

mountainous generates the high steep taluses where are potential landslide risk.

Besides expansion of roads, the urbanization, as well as establishment of the newresidential areas along valleys and narrow streams, also arises landslide

ML1.2.2 Deforestation

‘The stream valleys in the upstream of Ca river basin and Hieu river basin are verynarrow and steep, ethnic minorities live in the areas with no land to cultivate rice,Hence, they mainly deforest to cultivation, This activity has destroyed a series of

forests in the upper part of the rivers, making the slopes deereased ability to retain

‘water and increased erosion and denudation of soil layer Besides that, deforestation

decreased the coverage of vegetation This thing is the potential cause for debris flow

forming

15

ML1.2.3 Exploration

‘There are some positions of iron exploration in the area study The landfills of theseexplorations were damed up as small mountains The special feature of these man-made mountains is loose material Therefore, the potential erosion is expected to behigh,

ML1.2.4 Land use changing

Due to lack of residential land and understanding of landslides and the damage caused

by landslide, people built their house close to streams and taluses where have beensuffering many impacts of landslides such as Sop Phe village, Na My village in Ky

Son district, Pieng village in Que Phong district These villages have been devastated

by debris flow Some facilities collapsed and must be displaced such as Sop Pheelementary school, Ta Ca hospital, the Luong Minh People's Committee headquarters

THL2 Data Collection

11.2.1 Topographic data

Digital Elevation Model (DEM) was collected from The Shuttle Radar TopographyMission (SRTM) dataset and downloaded via USGS website The spatial resolution ofDEM is 90 meter The data distributed in GeoTIEF format, in decimal degrees anddatum WGS84,

m2 Satellite images data

In the study, Landsat 8 is used to create the C factor map To cover the study area, fourimages was downloaded via USGS website, The data distributed in GeoTIFF format,

ed date were descril

in decimal degrees and datum WGS84, ID and acqui ed in belowtable,

Table 4: Information of used Landsat 8 images

Landsat Scene ID Patch Row Acquired Date

LC§1260472016114LGN00 | 126 47 2016-04-23

16

‘The daily rainfall data is collected from 8 meteorological stations in Nghe An

provine „ including Vinh, Con Cuong, Do Luong, Quy Chau, Quy Hop, Quynh Luu,

“Tay Hieu and Tương Duong station This data was taken from 1975 to 2006,

11.2.5 Landslide inventory data

Landslide inventory data was collected from “Investigation, assessment and warningzonation of landslides in mountainous regions of Vietnam” project Set of dataincluded 1276 landslide points and its attributes which were surveyed from 2012 Thescope of the project covers 11 mountainous districts of Nghe An province, includingAnh Son, Con Cuong, Do Luong, Ky Son, Nghia Dan, Que Phong, Quy Chau, QuyHop, Tan Ky, Thanh Chuong and Tuong Duong district Since the remaining area ismainly flat plain, the investigation was not conducted,

1

CHAPTER IV METHODOLOGY

Rainfall is the most important triggering factor of landslide in Vietnam, There areseveral researches using the rainfall erosivity index to estimate landslide

susceptibility[3]-6] The rainfall erosivity index is the factor used in RUSLI (Revised

Universal Soil Loss Equation) Because erosion process forms unstable and loosegeological materials leading to the generation of extensive mass movements, Thus,application of rainfall erosivity index as an indicator to assess landslide susceptibility

is necessary.

RUSLE, is an empirical model, was developed in 1960 and updated in 1978 byWischmeier and Smith of the United States Department of Agriculture, This modelenables (0 prediet the average annual rate of soil erosion for the specific area, The

model based on several factors which affect erosion There are rainfall, topography,land use practices, soil, and vegetation[ 16] Each factor is represented by the following,

equation:

A=RxKxLxSxCxP

‘Where: A is the average annual soil loss (tons ha” year ”)

Ris the rainfall factor (MJ mm ha ' h' year)

K is the soil erodibility factor (th MF" mm")

Lis the slope length factor (unitless)

Sis the slope steepness factor (unitless)

Cis the crop and management factor (unitless)

is the conservation practice factor (unitless)

18

RUSLE model has been applied in many researches, But six parameters werecalculated by many different formulas in each research, That difference depends oncollected data in each research and particular conditions of a given ease study.

TY.1 Rainfall factor (R)

‘The rainfall erosivity index is a numerical descriptor of the capacity of rainfall to erodesoil The rainfall erosivity index is an indicator of the two variables most critical to astorm’s erosivity: the amount of rainfall and the peak intensity sustained over anextended period R is the average annual sum of all erosive rainfall events (Els)

G Wall [16] state that there are three approaches to parameterize R factor

1, Calculate using measured rainstorm El values Itis suitable if 22 or more years

of rainfall intensity data is available

2 Use equations which rely on an empirical relationship between R

3 Use hourly precipitation records, where available, to predict R

Saro Lee[3] used the following formula to calculate R index The ra

collected from 51 meteorological stations in South Korea

(SE)Bomax

100R

‘Whereas, the empirical equation suggested by Amoldus was applied in Pradhan'sstudy[5] The monthly average rainfall data of 20 years (1985-2004) was collectedfrom the Bayan Lepas weather station fo compute R

Ee)»

‘Where: Pi is the monthly average rainfall (mm) for the month i,

P is the annual average rainfall (mm)

19

Rugg represents the annual average R

According to the equation suggested by Flabouris, R factor was determined in Rozos’sstudy[6l:

R=R xa

‘Where: a is equal to 0.5 for the study area

R, is the mean annual rainfall

In the same way, Zini chose the regression formula was published by Grauso for hisstudy

à

‘Where: ais the total annual rainfall (mm)

iis the annual maximum daily rainfall (mm)

his the annual maximum hourly rainfall (mm)

Thus, there are many formulas to calculate R factor R factor can be directly computed

by the original formula developed by Wischmeier if enough data, If not, the empirical

‘equation which is suitable for each study area is another choice for the scientists

1V.2 Soil Erodibility factor (K)

“The soil erodibility factor is referred the susceptibility/resistance to erosion of soil Inthe field, itis difficult to determine soil erodibility index K factors were measuredfrom over 10000 soil plot in laboratory condition (Wischmeier and Smith, 1978)

ca ‘ulation of a K value is based on five parameters, routinely characterized throughstandard soil profile descriptions and laboratory analyses These five parameters arepercent silt plus very fine sand (0.05 to 0.10 mm), percent sand greater than 0.10 mm,organic matter content, structure, and permeability Finally, the soil erodibilitynomograph was developed from the results of this research, The nomograph provides a

20

‘graphical solution for determining a soils K value and can be used if the percent sand

Figure 8: The soil erodibility nomograph

IV.3 Topography Factors (LS)

‘The topography factors mean the length and the steepness of the slope Both of themaffect the rate of soil erosion, Slope length is defined as the distance from the point oforigin of overland flow to the point where the slope decreases sufficiently for

deposition to occur or to the point where runoff enters a defined channel (wet or dry),

‘The slope steepness is the segment or site slope, usually expressed as a percentage

Slope length and slope steepness strongly influence the transport of soil particles once

the soil particles are dislodged by raindrop impact or runoff, Because the LS-factor can

be defined to be substantially greater than unity, it can have a considerable effect onthe predicted erosion

“The original USLE - LS equation for a uniform slope is

LS = (Flow accumulation x Cell size/22.13)"" x sin (Slope/0 896)! `

Or Rozos used the equation, which was proposed by Morgan[6),

LS = (L/22)"* x (0.065 +0.045 x S + 0.0065 x 8),

‘Where: L is slope length (m)

Sis slope steepness (%)

IV.4 Crop and Management Factor (C)

‘The C factor typically represents the soil erosion confinable degree, which dependingoon the land cover In general, the land cover takes the significant efforts in decreasingthe raindrop’s kinetic energy in the progress of falling down as well as inerea ng the

soil cohesion avoiding sweep away by the runoff

In detail, the land cover has affected the soil erosion regarding two aspects:

‘© The canopy would protect the soil surface from raindrop by decreasing its

kinetic energy - the key Factor of soil erosion

2