Table of Contents Author Guidelines Submit a Manuscript
Advances in Meteorology
Volume 2016 (2016), Article ID 4572498, 13 pages
Research Article

Development and Application of Urban Landslide Vulnerability Assessment Methodology Reflecting Social and Economic Variables

1Division of Earth Environmental System Science, Pukyong National University, 45 Yongso-ro, Nam-gu, Busan 48513, Republic of Korea
2Department of Ocean Engineering, Pukyong National University, 45 Yongso-ro, Nam-gu, Busan 48513, Republic of Korea
3Civil and Environmental Engineering, Cleveland State University, Cleveland, OH 44115, USA
4Department of Environmental Engineering, Pukyong National University, 45 Yongso-ro, Nam-gu, Busan 48513, Republic of Korea

Received 23 December 2015; Revised 29 March 2016; Accepted 7 June 2016

Academic Editor: Philip Ward

Copyright © 2016 Yoonkyung Park et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


An urban landslide vulnerability assessment methodology is proposed with major focus on considering urban social and economic aspects. The proposed methodology was developed based on the landslide susceptibility maps that Korean Forest Service utilizes to identify landslide source areas. Frist, debris flows are propagated to urban areas from such source areas by Flow-R (flow path assessment of gravitational hazards at a regional scale), and then urban vulnerability is assessed by two categories: physical and socioeconomic aspect. The physical vulnerability is related to buildings that can be impacted by a landslide event. This study considered two popular building structure types, reinforced-concrete frame and nonreinforced-concrete frame, to assess the physical vulnerability. The socioeconomic vulnerability is considered a function of the resistant levels of the vulnerable people, trigger factor of secondary damage, and preparedness level of the local government. An index-based model is developed to evaluate the life and indirect damage under landslide as well as the resilience ability against disasters. To illustrate the validity of the proposed methodology, physical and socioeconomic vulnerability levels are analyzed for Seoul, Korea, using the suggested approach. The general trend found in this study indicates that the higher population density areas under a weaker fiscal condition that are located at the downstream of mountainous areas are more vulnerable than the areas in opposite conditions.

1. Introduction

In South Korea, there has been continuous interest in reducing landslide or debris flow damage because about 70% of the Korean territory is covered with mountainous areas. Many researchers have analyzed the factors causing landslides [1, 2]. Landslide susceptibility maps have been built on the mainland across South Korea by Korea Forest Service (KFS) and have been used as the basis for the studies to predict landslides or reduce the damage. The resolution of KFS landslide susceptibility maps is 10 m by 10 m, and the maps are classified into 5 grades according to the probability of landslides. Since the maps were developed to limit the mountain areas, the impacts of landslides on the downstream side are not reflected or evaluated. The Mt. Umyeon landslide that occurred in Seoul, South Korea, in July 2011 suggested that the impact of mountain landslides on the downstream city should be considered in the landslide and debris flow information system. From the lessons of Mt. Umyeon landslide, this study was planned to evaluate the landslide vulnerability to reflect the impact on the urban areas under the mountains.

In view of this, the damage type classification proposed by Smith and Ward [3] provides important insights in the vulnerability assessment framework configuration. They classified the types of damage from natural disasters into direct and indirect damage. Representative examples of the direct damage is the damage of life and property. On the other hand, examples of the indirect damage can be the intangible damage caused by traffic disruptions. In order to evaluate properly the urban vulnerability to natural disasters, various types of damage should be reflected in the vulnerability assessment framework [4].

Although many studies have evaluated the damage of buildings due to a landslide [57], studies evaluating the damage of life or secondary damage are not sufficient to provide a guideline. In recent years, however, some studies on the socioeconomic vulnerability assessment considering the damage of life or secondary damage have been attempted [811]. Those studies mainly focused on evaluating the vulnerability in terms of sensitivity to natural disasters and ability to respond to them.

In this study, a combined methodology for evaluating the urban physical and socioeconomic vulnerability to landslides is proposed. Urban landslide vulnerability is assessed by two categories; physical and socioeconomic aspect. The damage of buildings by landslides is considered to be a proxy variable representing the physical vulnerability. On the other hand, the socioeconomic vulnerability is evaluated using a number of proxy variables that can explain the exposure degree of vulnerable people in landslide disasters, factors causing a variety of secondary damage, and disaster preparedness of local governments.

2. Study Area and Scale

The study area is Seoul, the capital of South Korea, which has a complicated socioeconomic infrastructure and a high population density. In countries most of the critical national infrastructure is highly concentrated in one single city (such as South Korea), and the vulnerability evaluation methods applied to many cities by previous studies are not suitable to the current study area. In order to properly evaluate the vulnerability of a large metropolitan area, it should be evaluated in much higher spatial resolution.

Korea National Statistical Office (KNSO) has set up “Census Output Area (COA)” as a minimal spatial scale to publish national standard statistical data in South Korea [12]. The size is determined by the amount of people and the social homogeneity resulting in an average population size of about 500 people per COA. Therefore the COA level was considered the basic mapping unit for the vulnerability assessment in this study. Seoul is composed of 16,230 COAs. The area of COAs ranges from 0.00012 km2 to 10.10 km2 (average 0.037 km2). Figure 1 shows the locations of Seoul with 25 boroughs and COAs constituting the Gangnam region in the boroughs. Other boroughs also show similar COA distribution to Gangnam.

Figure 1: Study area, Seoul, South Korea, with Census Output Area.

3. Urban Landslide Vulnerability Assessment

The proposed urban landslide vulnerability assessment process consists of three steps. The first step is to identify potential landslide hazard areas, which can be carried out by combining landslide susceptibility maps and Flow-R model [13]. The second step is to assess separately physical and socioeconomic vulnerability based on identified potential landslide susceptibility areas. As a final step, the urban landslide vulnerability is generated by combining physical and socioeconomic vulnerability.

3.1. Potential Landslide Areas Identification

Vulnerability assessment under landslide disaster requires information on the propagation extent and intensity by debris flows. Such information on debris flow susceptibility mapping can be obtained by regional scale runout simulation. Since landslides are caused by various natural factors, landslide analysis in regional scale is difficult. Horton et al. [13] developed a distributed empirical model, Flow-R, for flow path assessment of gravitational hazards at a regional scale. The model is available free of charge at One advantage of this model is that the amount of data required for the simulation is small. A Digital Elevation Model (DEM) may be sufficient to identify potential landslide source areas and to process the propagation. Data such as slope, curvature, and flow accumulation are required for more accurate analysis.

This study applies Flow-R to obtain two landslide characteristics: spreading probability and impact pressure. DEM and predefined sources are used for Flow-R simulation. Predefined source is potential grids where landslide may occur. The predefined source used in this study is obtained by using landslide susceptibility maps. The maps developed by KFS are displayed by classifying the landslide probability to five ranks. Rank 1 areas in the maps have the highest probability of landslide occurrence.

In other words, landslide event is the most frequent in the area. This study is the first step in South Korea about urban landslide vulnerability. It is necessary to study the extreme cases. “Predefined source” is defined by rank 1 area in the map to apply extreme case (red grids in Figure 2). The resolution of DEM and predefined source areas is 10 m by 10 m.

Figure 2: Identification of predefined source areas using landslide susceptibility map.

Flow-R provides a variety of algorithms to analyze debris flow. Table 1 shows the algorithms applied to analyze landslide disaster in this study. Spreading algorithms control the path and the spreading of debris flow. Friction law determines the runout distance. All algorithms in Flow-R are described in detail by Horton et al. [13]. The parameters for simulating Flow-R have to be calibrated and verified using actual data. But such data set is not provided in Seoul. So this study has no choice but to use the reference values from Horton et al. [13].

Table 1: Selected landslide disaster algorithms of Flow-R.

The result of Flow-R simulation is the total area that can be potentially propagated by debris flows with an associated susceptibility value and kinetic energy. The resolution of the output is the same as that of the input data. The susceptibility value is used to extract COAs which are influenced by landslide runout. Equation (1) is used to estimate the impact pressure (p, kPa) in each grid [15]:where is the kinetic energy calculated by Flow-R internally and is the bulk density of debris flow (2,239.17 kg/m3 was used in this study). If the impact pressure of landslide exceeds about 34 kPa, houses built with bricks or woods would be destroyed completely [6].

3.2. Physical Landslide Vulnerability Assessment

When a landslide occurs, the physical characteristics such as velocity, depth, and impact pressure are determining the level of damage of buildings. Many studies have been conducted to find the relationship between physical characteristics and the damage of various structures [1618]. These previous studies mainly focused on creating a vulnerability curve by correlating the relationship between landslide characteristics and the damage of buildings. Based on those findings, this study developed vulnerability curves for two different structure types in the study area, and then the curves were used to assess physical vulnerability. Figure 3 shows the procedure of the physical vulnerability assessment.

Figure 3: Procedure of physical vulnerability assessment.

The first stage of the physical vulnerability assessment is to identify COAs affected by landslides using Flow-R simulations and to calculate the spatially averaged impact pressure in identified COAs. The second stage is to classify the buildings in an identified COA into two categories: nonreinforced concrete (non-RC) frame and reinforced-concrete (RC) frame. Finally, the physical vulnerability assessment is to calculate the physical vulnerability by linking the spatially averaged impact pressure in the identified COA and vulnerability curves.

Two kinds of vulnerability curves, which have been developed by Kang and Kim [6], are applied in accordance with each of the two building structure types (Figure 4). The physical vulnerability therefore ranges from 0 (the lowest vulnerability) to 1 (the highest vulnerability) as shown in Table 2. The highest vulnerability (value equals 1) means buildings are broken down completely when landslide event happens.

Table 2: Vulnerability functions [6].
Figure 4: Vulnerability curves depending on impact pressure [6].
3.3. Socioeconomic Vulnerability Assessment

Socioeconomic vulnerability aims to evaluate the social and economic factors that may be affected by landslides. Socioeconomic vulnerability should be evaluated by a variety of factors with various perspective. This can be carried out using an index-based model [8]. In this study, the modified version to match the conditions in Korea based on the socioeconomic vulnerability assessment framework of landslides developed by Safeland [19] is proposed. The proposed model can evaluate demographic factors, economic factors, and landslide disaster preparedness and response capabilities. The potential proxy variables are identified as age distribution, population density, housing type, personal assets, risk perception, the presence of disaster warning system, and so on. Those variables are mainly selected by a literature review and expert interviews. The procedure of socioeconomic vulnerability assessment is shown in Figure 5.

Figure 5: Procedure of socioeconomic vulnerability assessment.

The model is composed of a total of three subindexes: Demographic and Social Index (DSI), Secondary-Damage-Triggering Index (STI), and Preparation and Response Index (PRI). DSI is evaluated by six population-related and social variables that may be affected by natural disasters. For example, “age distribution” is classified into vulnerable people group. The children or elderly people are more vulnerable than young people. It is the reason to select “age distribution” as proxy variable in DSI (Table 4). “Population density” influences vulnerability. If landslide event happens in high population density area, it will caused heavy casualties. Therefore, “population density” can be used as proxy variable to assess socioeconomic vulnerability.

STI is an index to evaluate the indirect damage caused by natural disasters. For instance, when roads are malfunctioning, damage such as traffic jam and destruction of many life lines is caused [20]. Public office becomes control tower in emergency situation. Damage in public office causes secondary damage due to absence of control systems in emergency situation such as landslide event. For this reason, “the number of public offices” is selected as proxy variable of STI (Table 5).

PRI assesses the ability to prevent and respond to natural disasters. All used proxy variables and subindexes are listed in Table 3. Statistical data that used in this study are obtained from KOSIS (Korea Statistical Information Service). Statistics used in this study were compiled in 2010 as a base year, and the spatial resolution of the data is COA. However, the resolution of the proxy variables for PRI is a borough (Table 6). Application of PRI is suitable in borough scale because disaster response system of South Korea is the local government.

Table 3: Proxy variables and their weights in socioeconomic vulnerability assessment model.
Table 4: Criteria for ranking of proxy variables of DSI.
Table 5: Criteria for ranking of proxy variables of STI.
Table 6: Criteria for ranking of proxy variables of PRI.

The structure of the socioeconomic vulnerability is modified to be suitable in South Korea using the result of them [21].

Tables 4, 5, and 6 show the suggested socioeconomic vulnerability assessment model with proposed proxy variables and criteria for ranking of the variables. These variables have been used in many previous researches (see citations in Tables 4, 5, and 6).

Finally, the socioeconomic vulnerability index is calculated by weighted average of quantified proxy variables ranks (using (2)). The weights are determined after extensive expert survey based on the analytic hierarchy process developed by Saaty [22, 23]. The weights can be found in Table 3. Consider

To assess the socioeconomic vulnerability to landslide disasters, exposure information of landslide should be combined. By applying the proposed methodology to the COAs identified from previous Flow-R simulation, the vulnerability is evaluated for each COA. The vulnerability of the COAs identified as nonaffected areas is assigned to be 0.

4. Results and Discussion

4.1. Physical Vulnerability Assessment

Landslide affected COAs were identified by Flow-R simulation, and the corresponding impact pressure for each identified COA was estimated. Figure 6 shows the spatial distribution of identified landslide affected regions and the corresponding impact pressure. The impact pressure was not calculated for the COAs that are not affected by landslide disasters. The number of identified COAs is 2,249 and the maximum average impact pressure is estimated to be 37 kPa.

Figure 6: Results of Flow-R simulation.

The physical vulnerability index is calculated by combining the impact pressure presented in Figure 6(b) and vulnerability functions (Table 2) developed by Kang and Kim [6]. In this study, two patterns of vulnerability functions were applied to two building structure types, non-RC and RC. It is noted that the average physical vulnerability is estimated to be 0.96 in the case that the impact pressure 37 kPa is applied to non-RC, while it is 0.20 for the RC case. Figure 7 shows the landslide physical vulnerability assessment map calculated by the proposed methodology.

Figure 7: Physical vulnerability to landslide disaster.

If physical vulnerability in a certain COA is estimated to be 1, it means that the buildings in the area are completely destroyed when landslide event happens. Therefore, a high physical vulnerability range (e.g., from 0.841 to 0.960 in Figure 8) means that the buildings in the area are damaged severely by landslide event.

Figure 8: Result of socioeconomic vulnerability assessments.

It can be observed that the COAs located below the mountainous areas with denser residential areas are more vulnerable to landslide disasters in physical aspect than the nonmountainous or less dense.

However, there are clear limitations to the use of the regional scale Flow-R model for obtaining quantitative output that can be used in combination with physical vulnerability curves. The Flow-R is an empirical model, which does not take into account source volume, entrainment, or rheology. In further research, a regional scale model that can calculate the volume of landslide should be applied [27, 28].

4.2. Socioeconomic Vulnerability Assessment

Socioeconomic vulnerability is evaluated by three detailed subindexes (Table 3) and the results of each vulnerability assessment are presented in Figure 8.

DSI is a vulnerability assessment subindex related to the population. In general, it can be seen that COAs where the population density is high are more vulnerable (Figure 8(a)). STI is a vulnerability assessment subindex related to the impact on other areas of the occurrence of a disaster event at a COA. In general, it can be seen that COAs including major urban areas are more vulnerable (Figure 8(b)). PRI is a vulnerable assessment subindex related to preparedness and ability to respond to disasters in a borough (Figure 8(c)). This vulnerability index is inversely proportional to the fiscal health of local governments. Therefore, the boroughs in a weak fiscal condition (northwestern and southeastern regions of the study area, Seoul) are classified as a lack of preparedness and ability to respond to disasters. Socioeconomic vulnerability assessment calculated by the weighted average of three detailed subindexes is shown in Figure 8(d). Overall, the southern region in Seoul is more vulnerable against the disaster in a socioeconomic perspective. The results shown in Figure 8(d) assume that natural disasters have occurred uniformly throughout the whole study area, Seoul.

To evaluate the vulnerability to landslides specifically for COAs, Figures 8(d) and 6(a) should be combined. The combined landslide vulnerability assessment in a socioeconomic aspect is shown in Figure 9 as a result. It is suitable to apply quantitative characteristics of landslide events. But this study overlaid susceptibility map from Flow-R due to lack of quantitative data about landslide event. If the quantitative assessment is able to analyze landslide in regional scales [27, 28], more accurate socioeconomic vulnerability about landslide can be obtained.

Figure 9: Socioeconomic vulnerability to landslide disaster.

The biggest difference between Figures 8(d) and 9 is observed in southwestern Seoul. In fact, the southwest area of Seoul is a very high likelihood of a flood disaster area. Therefore, although this region has higher vulnerability to common natural disasters, the region may not have high vulnerability to landslide.

4.3. Urban Landslide Vulnerability

Urban characteristics were reflected in vulnerability assessment under landslide disaster by combining physical and socioeconomic vulnerabilities. For the combination of the two vulnerability assessments, the physical vulnerability region (Figure 7) was multiplied by the socioeconomic vulnerability region (Figure 9) and the region values were normalized between 0 and 1 as one urban vulnerability to landslide. The urban vulnerability index can be classified into five categories as very low, low, moderate, high, and very high. The result of the integrated urban landslide vulnerability assessment is shown in Figure 10.

Figure 10: Urban vulnerability to landslide disaster.

Figure 10 indicates that 50% or more of COAs affected by landslides (Figure 7) have an urban vulnerability of less than 0.3 (urban vulnerability is very low or low). This means that these COAs are likely to be in a landslide damage but relatively less vulnerable. The size of these COAs varies (0.00046 km2–7.52 km2, average 0.11 km2). On the other hand, the size of very vulnerable COAs where the urban vulnerability is at least 0.6 is small (0.0047 km2–2.24 km2, average 0.046 km2). The size of a COA is mainly determined by the population and their socioeconomic homogeneity. When the size of a COA is very small, the COA has a very high population density and is highly urbanized. If landslides occurred in these COAs, the damage caused by the landslides would be amplified. Therefore, these COAs are extremely vulnerable to the landslide disaster, and disaster mitigation measures must be taken first in these priority areas.

5. Summary and Conclusion

For suitable vulnerability assessment of natural disasters including landslides, this study showed that both direct damage and indirect damage should be considered in the vulnerability assessment processes. In this study, vulnerability assessment for landslides among various natural disasters was conducted. In particular, the proposed methodology properly reflects the urban characteristics in terms of vulnerability to natural disasters. Therefore, the results of this study can be directly utilized for setting the priority of various landslide damage reduction projects in urban areas.

In the proposed methodology, the possible landslide source areas were first identified using existing landslide susceptibility maps. Flow-R simulation provided the extents and impact pressure of debris flow routed from the identified sources areas of landslide. The outcomes were served as landslide exposure information for physical and socioeconomic vulnerability assessment.

Direct damage due to landslide disasters was considered by physical vulnerability. Vulnerability functions quantifying the degree of damage of buildings were applied to evaluate the physical vulnerability. By applying vulnerability functions to building information and impact pressure of each affected COA, the physical vulnerability index was estimated.

Socioeconomic vulnerability was applied to assess indirect damage caused by landslides. An index-based model was used to assess the socioeconomic vulnerability. The model was comprised of three subindexes and 17 proxy variables. Three subindexes were the demographic-social index, the secondary-damage-triggering index, and the preparation-response index. The socioeconomic vulnerability index calculated by the index-based model is applicable to common natural disasters. Although this study only demonstrated an example that the landslide exposure information obtained by Flow-R simulation is combined to assess the socioeconomic vulnerability of landslide disasters, the socioeconomic vulnerability of flood disasters can be assessed by combining with flood exposure information.

Finally, the urban landslide vulnerability was assessed by combining the physical vulnerability and socioeconomic vulnerability. The combined vulnerability therefore reflects both direct and indirect aspects of damage caused by landslides. The general trend of the findings in this study reveals that the higher population density areas under a weaker fiscal condition that are located at the downstream of mountainous areas are more vulnerable than the areas in opposite conditions. The framework and results of this study are expected to be used directly to prioritize the landslide damage reduction plans for the Seoul metropolitan area in South Korea.

The previous studies about landslide in South Korea were focused on mountainous areas. On the other hand, the influence of landslide in urban areas has not been analyzed. This study is prototype about assessing vulnerability in urban areas. The study area is Seoul, in South Korea. Therefore, this study has some limitations. One of limitations is that this study did not use quantitative output about landslide process. It is not the accurate landslide vulnerability assessment. However it is good enough as the first step for high quality landslide disaster vulnerability assessment in urban area including mountainous area.

Competing Interests

The authors declare that they have no competing interests.


This research was supported by a grant (13SCIPS04) from Smart Civil Infrastructure Research Program funded by Ministry of Land, Infrastructure and Transport (MOLIT) of Korea Government and Korea Agency for Infrastructure Technology Advancement (KAIA).


  1. S. Lee, J.-H. Ryu, J.-S. Won, and H.-J. Park, “Determination and application of the weights for landslide susceptibility mapping using an artificial neural network,” Engineering Geology, vol. 71, no. 3-4, pp. 289–302, 2004. View at Publisher · View at Google Scholar · View at Scopus
  2. K.-Y. Song, H.-J. Oh, J. Choi, I. Park, C. Lee, and S. Lee, “Prediction of landslides using ASTER imagery and data mining models,” Advances in Space Research, vol. 49, no. 5, pp. 978–993, 2012. View at Publisher · View at Google Scholar · View at Scopus
  3. K. Smith and R. Ward, Floods: Physical Processes and Human Impacts, John Wiley & Sons, Chichester, UK, 1998.
  4. S. L. Cutter, B. J. Boruff, and W. L. Shirley, “Social vulnerability to environmental hazards,” Social Science Quarterly, vol. 84, no. 2, pp. 242–261, 2003. View at Publisher · View at Google Scholar · View at Scopus
  5. M. Galli and F. Guzzetti, “Landslide vulnerability criteria: a case study from Umbria, Central Italy,” Environmental Management, vol. 40, no. 4, pp. 649–664, 2007. View at Publisher · View at Google Scholar · View at Scopus
  6. H. Kang and Y. Kim, “The physical vulnerability of different types of building structure to debris flow events,” Natural Hazards, vol. 80, no. 3, pp. 1475–1493, 2016. View at Publisher · View at Google Scholar
  7. Z. Li, F. Nadim, H. Huang, M. Uzielli, and S. Lacasse, “Quantitative vulnerability estimation for scenario-based landslide hazards,” Landslides, vol. 7, no. 2, pp. 125–134, 2010. View at Publisher · View at Google Scholar · View at Scopus
  8. M. L. Carreño, O. D. Cardona, and A. H. Barbat, “A disaster risk management performance index,” Natural Hazards, vol. 41, no. 1, pp. 1–20, 2007. View at Publisher · View at Google Scholar
  9. A. Dwyer, C. Zoppou, O. Nielsen, S. Day, and S. Roberts, Quantifying Social Vulnerability: A Methodology for Identifying Those at Risk to Natural Hazards, Australian Government, Geoscience Australia, Symonston, Australia, 2004.
  10. Y. Park, S. Jeong, and S. Kim, “Natural disaster vulnerability assessment at boroughs and census output areas in seoul focusing on socio-economic perspective,” Journal of Korean Society of Hazard Mitigation, vol. 14, no. 6, pp. 439–449, 2014. View at Publisher · View at Google Scholar
  11. T. H. Siagian, P. Purhadi, S. Suhartono, and H. Ritonga, “Social vulnerability to natural hazards in Indonesia: driving factors and policy implications,” Natural Hazards, vol. 70, no. 2, pp. 1603–1617, 2014. View at Publisher · View at Google Scholar · View at Scopus
  12. Statistical Geographic Information Service (SGIS), 2015,
  13. P. Horton, M. Jaboyedoff, B. Rudaz, and M. Zimmermann, “Flow-R, a model for susceptibility mapping of debris flows and other gravitational hazards at a regional scale,” Natural Hazards and Earth System Science, vol. 13, no. 4, pp. 869–885, 2013. View at Publisher · View at Google Scholar
  14. P. Holmgren, “Multiple flow direction algorithms for runoff modelling in grid based elevation models: an empirical evaluation,” Hydrological Processes, vol. 8, no. 4, pp. 327–334, 1994. View at Publisher · View at Google Scholar · View at Scopus
  15. Safeland, “Living with landslide risk in Europe: assessment, effects of global change, and risk management strategies,” Deliverable D7.5 GIS-based training package on landslide risk assessment Work Package 7—Dissemination of project results, 2012.
  16. S. Fuchs, K. Heiss, and J. Hübl, “Towards an empirical vulnerability function for use in debris flow risk assessment,” Natural Hazards and Earth System Science, vol. 7, no. 5, pp. 495–506, 2007. View at Publisher · View at Google Scholar · View at Scopus
  17. E. D. Haugen and A. M. Kaynia, “Vulnerability of structures impacted by debris flow,” in Landslides and Engineered Slopes, Z. Chen, J. M. Zhang, K. Ho, F. Q. Wu, and Z. K. Li, Eds., pp. 381–387, Taylor & Francis, London, UK, 2008. View at Google Scholar
  18. B. Quan Luna, J. Blahut, C. J. van Westen, S. Sterlacchini, T. W. J. van Asch, and S. O. Akbas, “The application of numerical debris flow modelling for the generation of physical vulnerability curves,” Natural Hazards and Earth System Science, vol. 11, no. 7, pp. 2047–2060, 2011. View at Publisher · View at Google Scholar · View at Scopus
  19. Safeland, “Living with landslide risk in Europe: assessment, effects of global change, and risk management strategies,” Deliverable D2.6 Methodology for evaluation of the socio-economic impact of landslides (socio-economic vulnerability), 2012.
  20. Y. Park, J. Kim, D. J. Jo, and S. Kim, “Urban mud and debris flow disaster vulnerability assessment associated with landslide hazard map: application to Busan, Korea,” Journal of Korean Society of Hazard Mitigation, vol. 15, no. 5, pp. 283–289, 2015. View at Publisher · View at Google Scholar
  21. U. M. K. Eidsvig, A. McLean, B. V. Vangelsten et al., “Assessment of socioeconomic vulnerability to landslides using an indicator-based approach: methodology and case studies,” Bulletin of Engineering Geology and the Environment, vol. 73, no. 2, pp. 307–324, 2014. View at Publisher · View at Google Scholar · View at Scopus
  22. T. L. Saaty, “A scaling method for priorities in hierarchical structures,” Journal of Mathematical Psychology, vol. 15, no. 3, pp. 234–281, 1977. View at Google Scholar · View at MathSciNet
  23. T. L. Saaty, “How to make a decision: the analytic hierarchy process,” European Journal of Operational Research, vol. 48, no. 1, pp. 9–26, 1990. View at Publisher · View at Google Scholar · View at Scopus
  24. S. Tapsell, S. Tunstall, C. Green, and A. Fernandez, “Social indicator set,” FLOODsite Report T11-07-01, 2005. View at Google Scholar
  25. A. Ebert, N. Kerle, and A. Stein, “Urban social vulnerability assessment with physical proxies and spatial metrics derived from air- and spaceborne imagery and GIS data,” Natural Hazards, vol. 48, no. 2, pp. 275–294, 2009. View at Publisher · View at Google Scholar · View at Scopus
  26. A. Steinführer, B. De Marchi, C. Kuhlicke, A. Scolobig, S. Tapsell, and S. Tunstall, “Vulnerability, resilience and social constructions of flood risk in exposed communities,” FLOODsite Report T11-07-12, 2009. View at Google Scholar
  27. B. Q. Luna, J. Blahut, C. Camera et al., “Physically based dynamic run-out modelling for quantitative debris flow risk assessment: a case study in Tresenda, northern Italy,” Environmental Earth Sciences, vol. 72, no. 3, pp. 645–661, 2014. View at Publisher · View at Google Scholar · View at Scopus
  28. G. Gaprindashvili and C. J. Van Westen, “Generation of a national landslide hazard and risk map for the country of Georgia,” Natural Hazards, vol. 80, no. 1, pp. 69–101, 2016. View at Publisher · View at Google Scholar · View at Scopus