Landslides susceptibility assessment using GIS and remote sensing method

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: Disaster Management Major codes: SUPERVISORS: 1.Supervisor 1: Dr Ngo Le An Supervisor 2: Dr Pham Thi Thanh Nga Department : Faculty of Hydrology and Water Resources Ha Noi, 2017 ACKNOWLEDGEMENT First of all, I would like to give my deepest appreciation to my supervisors Dr Ngo Le An and Dr Pham Thi Thanh Nga Without their valuable guidance, suggestions and comments, it could be impossible to complete my thesis I 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 in field experiment and data processing I would like to extend my sincere thanks to Disaster 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 the difficulty in study times I 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 unconditional assistance through the whole period of studying Without the help from all of you, this study would have been impossible Nguyen Hoang Ninh i DECLARATION I hereby 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 are fidelity, which are not copied from any sources and any forms The reference documents relevant sources, the thesis has cited and recorded as prescribed The results of my thesis have not been published by me to any courses or any awards ii ABSTRACT 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, storms from 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 landslide, debris flow and flashflood in Nghe An RUSLE model is used to predict the average annual rate of erosion But erosion not only loses soil but also forms unstable slope lead to increasing of the development of the 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 the current landslide assessment methods Therefore, the potential soil erosion map seems to affect the landslide susceptibility assessment By combining Remote Sensing analysis and Geo Information System analysis to assess the landslide susceptibility, the result shows that the area very high and high susceptible for landslide cover 26% and 25%, respectively These results reflect the real occurring in Nghe An province This result is useful for administrators, decision maker, and planner iii Abbreviation WB World Bank VND Vietnam Dong RUSLE Revised Universal Soil Loss Equation GIS Geographical Information System RS Remote Sensing GHIS Geological Hazard Information System SMCE Spatial Multi Criteria Evaluation AHP Analytical Hierarchy Process DEM Digital Elevation Model SRTM The Shuttle Radar Topography Mission USGS United States Geological Survey iv Table of Contents Chapter I Introduction I.1 General I.2 Objective of study I.3 Research question Chapter II Literature review II.1 Landslide mechanism II.1.1 Landslide Process II.1.2 Landslide types and triggering factors II.2 Apply Remote Sensing and GIS in landslide susceptibility assessment II.2.1 Remote Sensing II.2.2 GIS CHAPTER III 11 Study area and data collection 11 III.1 Study area 11 III.1.1 Natural features related to landslide 12 III.1.2 Human activities related to landslide 15 v III.2 Data Collection 16 III.2.1 Topographic data 16 III.2.2 Satellite images data 16 III.2.3 Soil data 17 III.2.4 Rainfall data 17 III.2.5 Landslide inventory data 17 CHAPTER IV 18 Methodology 18 IV.1 Rainfall factor (R) 19 IV.2 Soil Erodibility factor (K) 20 IV.3 Topography Factors (LS) 21 IV.4 Crop and Management Factor (C) 22 IV.5 Conservation Practice Factor (P) 23 IV.6 Correlation between soil erosion map and landslide occurrence 27 IV.7 Methodology Flowchart 28 CHAPTER V 29 Results 29 V.1 RS analysis 29 V.1.1 Lansat pre-processing 29 V.1.2 Normalized Difference Vegetation Index 30 V.1.3 Transformed Value Range 31 vi V.2 GIS Analysis 32 V.2.1 R factor 32 V.2.2 K factor 34 V.2.3 LS factor 37 V.2.4 Average Annual Soil Loss 39 V.3 Verification of Erosion Soil Loss Map 43 V.4 Distribution of the susceptibility zones for landslide 46 V.4.1 The Very high susceptibility zone for landslide 46 V.4.2 The high susceptibility zone for landslide 47 V.4.3 The average susceptibility zone for landslide 47 V4.4 The low susceptibility zone for landslide 47 Conclusion 49 References 50 vii LIST OF FIGURES Figure 1: The tracks of all tropical cyclones in the Northwest Pacific Ocean from 1980 to 2005 Figure 2: Statistic of tropical storm affected to Vietnam from 1959 to 2009 Figure 3: The damage curves due to nature disaster in Vietnam Figure 4: The relationship between the force and the slopes Figure 5: GIS processing for landslide susceptibility assessment Figure 6: Administrative map of Nghe An province 11 Figure 7: The landslide on 7th highway 15 Figure 8: The soil erodibility nomograph 21 Figure 9: Methodology of research 28 Figure 10: Flow chart of Remote Sensing analysis to calculate C factor 29 Figure 11: Pre-processing of Landsat 30 Figure 12: NDVI map 31 Figure 13: C factor map 32 Figure 14: Flowchart of GIS analysis to calculate R factor 33 Figure 15: R factor map 34 Figure 16: Flow chart of GIS analysis to calculate K factor 35 Figure 17: K factor map 37 Figure 18: Flow chart of GIS analysis to calculate LS factor 38 Figure 19: LS factor map 39 viii 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 45 Figure 24: The statistic graph of the distribution of landslide susceptibility zone in all districts of Nghe An province 46 ix Table 7: Classification of soil erosion caused by rain based on Vietnam Standards for soil quality Vietnam Erosion Intensity Annual Soil Loss Description Standard Level (t.ha-1.year-1) TCVN I 200 Very High Erosion I 50 Very High Erosion 5299-1995 TCVN 5299-2009 In this study, the erosion intensity classification is proceeded in both version In the classification based on 1995 version, the distribution is variable with the following proportion of each level: 42% low level, 7% moderate level, 25% high level and 26% very high level 41 Figure 21: Erosion Intensity map based on Vietnam standard for soil quality in 1995 (TCVN 5299-1995) with landslides distribution In the 2009 classification, the distribution is variable with the following proportion of each level: 39% no erosion, 2% low level, 1% moderate level, 7% high level and 51% very high level 42 Figure 22: Erosion Intensity map based on Vietnam standard for soil quality in 2009 (TCVN 5299-2009) with landslides distribution V.3 Verification of Erosion Soil Loss Map A frequency ratio statistical analysis was used to compute the verification process This method is based on the relationships between spatial distribution of landslides and soil erosion prone zones The frequency ratio is the ratio of the probability of a landslide event to a non-event for a given soil erosion zone [3], [5], [6] To calculate the frequency ratio, the frequency distribution of landslide points was calculated for each erosion intensity zone of the study area Additionally, the area ratio for each erosion intensity zone to the total area was computed Finally, the frequency ratio for each soil erosion zone was calculated by dividing the frequency of slope failure events to the area ratio.[6] 43 According to Pradhan[5] and Rozos [6], frequency ratio value present the relationship between landslide event and the given erosion intensity zone The frequency ratio value of is an average value The frequency ratio values greater than 1, indicate strong relationship and this values smaller than 1, indicate poor relationship Table 8: Frequency ratio values of landslides occurrences into each zone in the erosion Intensity map based on Vietnam standard for soil quality in 1995 Erosion intensity zone Pixel Pixels rate Landslide occurrences Landslide Frequent rate ratio Low 857920 42% 433 34% 0.8 Moderate 136323 7% 70 5% 0.82 High 513016 25% 346 27% 1.07 Very high 518029 26% 427 33% 1.31 Total 2025288 100% 1276 100% Table shows that the high and very high erosion intensity zones are strongly correlated with landslide occurrences In these two classes, the frequency ratio was >1 indicating a high probability of the landslide occurrences Regarding the average, the low erosion intensity zones, the frequency ratio values were 1 indicating a high probability of the landslide occurrences Regarding the high, the average, the low intensity zones, the frequency ratio values were