THAI NGUYEN UNIVERSITY UNIVERSITY OF AGRICULTURE AND FORESTRY HANNA JOY TILPO ORDANZA GIS-BASED FLOOD RISK MAPPING: A CASE STUDY OF FLOOD RISK ASSESSMENT USING ANALYTIC HIERARCHY PRO
Research Rationale
Natural and human-made disasters pose significant global challenges, with natural calamities like typhoons, earthquakes, and volcanic eruptions causing widespread destruction and loss of life each year These disasters can have immediate effects, such as those from earthquakes, or gradual impacts, like droughts It is crucial to distinguish between a disaster, which directly affects populated areas and results in harm, and a hazard, which occurs in unoccupied regions and poses potential risks Disasters are categorized into three types: natural disasters, which occur without human intervention; man-made disasters, resulting from human activities like pollution; and man-induced disasters, where human actions exacerbate natural events.
Flooding is a significant global disaster that intertwines environmental, community, and social factors The frequency and exposure to floods are expected to escalate, particularly in low-lying areas of the Asia Pacific and Africa This increase is driven by rapid population growth, economic development, and climate change Despite government efforts in disaster management and prevention, flooding remains a critical threat to smaller communities (Luu & Meding, 2018).
In developed countries, the adverse effects of disasters are mitigated through advancing risk reduction strategies Conversely, third world countries continue to suffer significant damage, struggling to compensate for the impacts on their unstable economies and communities.
Vietnam, a developing country in Southeast Asia, experiences distinct climate patterns, with North Vietnam having a temperate climate and South Vietnam being tropical The country is particularly vulnerable to typhoons due to its coastline along the South China Sea, with Da Nang City in Central Vietnam being the most affected Historically, Vietnam has recorded an average of four to six typhoons annually since 1973, with a peak of 23 typhoons in 2019 However, global climate change has led to increasingly severe typhoons, resulting in devastating floods that threaten lives, livelihoods, and the overall stability of the nation.
Tram Tau is a rural district in the mountainous province of Yen Bai, Vietnam, located at coordinates 21° 32' 0" N and 104° 26' 0" E Covering an area of 742 km², it has a population of 33,962, representing 4.2% of the province's total population as of 2019 The district is vulnerable to typhoons and flooding, which pose significant risks to its primary source of income—agriculture As one of the 65 poorest districts in Vietnam, flooding can devastate local livelihoods and endanger residents in flood-prone areas.
In 2017, the province experienced severe devastation, impacting over 900 hectares of crops and vegetation, thousands of livestock and poultry, and 42 hectares of fish farming The total compensation amounted to approximately 700 billion VND (30.8 million USD) Tram Tau reported 13 fatalities and missing persons, seven injuries, and more than 130 houses destroyed due to heavy rains and flooding (Vietnam News Agency, 2017).
Despite the extreme flood risk in Tram Tau district, there is a lack of studies and information on micro-community level flood risk assessment as a disaster risk reduction strategy Effective disaster management and risk preparedness are crucial for government response to inevitable disasters Mitigation, a key phase in the disaster management cycle, emphasizes sustainable development through public education, land use management, and vulnerability analysis updates Currently, there are no published studies or maps predicting flood-prone areas in Tram Tau To address this gap, a novel mapping and modeling technique will be employed, utilizing Geographical Information System and Remote Sensing alongside Multicriteria Decision Analysis This approach aims to provide valuable information to the Tram Tau community for informed decision-making in disaster management.
Research Questions and Hypotheses
This study aims to utilize Geographic Information Systems (GIS) for developing thematic map layers and to apply Multi-Criteria Decision Making Analysis to assign weights based on their significance, facilitating effective flood risk assessment and map production The research questions and hypotheses are formulated from these two key components of the study.
1.2.1 Collecting and processing data using Geographic Information System and Remote Sensing
Question: How are GIS and RS used to make flood risk assessment in a micro-community?
This study proposes the creation of thematic map layers derived from spatial data obtained through remotely sensed satellite images By utilizing GIS software, these layers will be overlaid to generate a comprehensive flood risk map.
1.2.2 Influence of mapping and modeling in disaster management
Question: What is the significance of mapping and modeling in forming mitigation measures?
Hypothesis: GIS-based risk mitigation strategies can help in prioritizing flood-prone areas, predicting future flood hazard in the community, and decision making by the local government unit officials
1.2.3 Effects of varying order of importance of map layers using AHP in creating final flood risk map
Question: What can be the effect of different scenarios where each scenario has different order of importance in producing flood-prone areas?
Hypothesis: Difference in the order of importance of each layer will noticeably affect the final flood risk map.
Research objectives
To effectively mitigate risk, it is essential to analyze both the hazards and vulnerabilities involved, as well as to document these risks through necessary mapping.
1.3.1 General objective of this study
This research aims to highlight the critical role of Geographic Information Systems (GIS) in reducing flood hazards in Tram Tau, a region still recovering from devastating floods four years ago By employing a multi-criteria analysis (MCA) method, the study seeks to enhance flood risk assessment and mapping As the demand for GIS applications grows, its potential for advanced modeling and simulation becomes increasingly important, especially in the context of climate change Neglecting micro-communities could lead to heightened vulnerability to natural disasters, underscoring the need for effective GIS strategies.
Flood hazard mapping is a crucial strategy for assessing flood-prone areas in Tram Tau, Yen Bai, Vietnam This assessment enables the identification of suitable sites for vegetation and urbanization through multi-criteria analysis, ultimately helping to mitigate the adverse effects of typhoons Such studies are vital for sustainable development, especially in light of the increasing frequency of extreme weather events due to climate change.
1.3.2 Specific objective of this study
The study aims to assess flood hazard and vulnerability through specific objectives derived from research questions, focusing on significant tasks essential for this evaluation.
To collect topographic data using Remote Sensing
To process datasets using GIS software
To produce thematic map layers
To weigh the importance of each map layers based on their potential of flooding occurrence with consistency
To understand the significance of multi-criteria decision analysis
To raise awareness to micro-community by producing maps
To contribute to the remote community of Tram Tau in disaster risk mitigation
Significance of the Study
This study will be significant to the:
The methodological framework that was used and the analysis that was produced can be utilized for deeper and further studies related to Geographic Information System and flood risk assessment
This research was conducted in a community level Thus, maps produced will be helpful for the micro-community that lacks of information that can contribute to disaster management
The findings and forecasts enable local government authorities and staff to make informed decisions and prioritize actions in flood-prone areas By utilizing these maps, the community can enhance its resilience and safety against typhoons.
Scope and Limitations
This study utilizes data from the GeoInformatics Research Center (GIRC), which was processed and published by the author between May and September 2021 It examines the topographical and social factors within the Tram Tau community The findings rely heavily on satellite imagery and other data inputs from the research center, meaning that certain thematic layers may not be included or acquired Additionally, the perceptions and knowledge of local residents were not part of this study's scope Despite achieving its objectives, the research acknowledges inherent limitations.
Several important criteria were excluded from the study due to a lack of available data in the area The weighting of these criteria was intended to be derived from experts and students of ESM; however, only three decision-makers were found to provide AHP judgments Additionally, the validation of results and findings through GPS-based field surveys could not be conducted because of the Covid-19 pandemic.
Definition of Terms
To avoid misunderstanding and misinterpretation upon reading this paper, it is necessary to define all important terms mentioned based on the context of the study
Geographic Information System (GIS) is an innovative tool for evaluating and managing various fields related to environmental, socio-economic, social, and industrial characteristics globally.
Remote Sensing (RS) is a method for gathering data and monitoring spatial features and astrological elements from a distance The term "remote" signifies the ability to collect information without direct presence in the area Examples of remote sensing include satellite imagery and aerial photography.
Multi-criteria Decision Analysis (MCDA) is an approach of making decisions or prioritizing one factor over another
Analytic Hierarchy Process (AHP) is a technique of MCDA widely used in most related studies in the world
Flood Risk Assessment (FRA) involves evaluating the likelihood of flooding in a specific area Various methods exist for assessing flood risk, but this study employs a GIS-based approach for a comprehensive analysis.
Decision Makers (DMs) play a crucial role in prioritizing disaster management in flood-prone areas Typically, these stakeholders include local government officials and flood management experts who are responsible for implementing effective strategies to mitigate the impacts of flooding.
Disaster Management is the systematic measure of preventing, preparing, mitigating, and responding to natural disasters
GIS-based approach is an approach of processing and analyzing a certain event in a GIS environment
Spatial Data is datasets related to space or area
Thematic Layers are layers separated by color in different categories based on its value
Hazard is a crucial element in flood risk assessment, encompassing the physical and statistical characteristics of flooding It is influenced by the severity of potential damage to a specific area.
Vulnerability is one of the components of flood risk assessment This is more focused on the responsiveness of the community and susceptibility exposed to human and their income source.
Disaster Management
Disasters, whether natural, man-made, or induced, disrupt normal lifestyles and are characterized by their sudden, unpredictable, and extensive nature They result in significant human impacts, including fatalities, injuries, and health issues, as well as economic consequences such as damage to properties, infrastructure, and industries Additionally, disasters affect a community's basic amenities, including housing, food, clothing, and healthcare The scientific perspective on disasters emphasizes the need to manage and mitigate their adverse effects on human lives and the environment As the frequency and destructiveness of disasters increase, stakeholders—including government, business, and community members—must take responsibility for preparing, mitigating, responding to, and recovering from these events.
Disaster management encompasses a comprehensive set of measures taken during each phase of a disaster, aiming to prevent, mitigate, and respond to its effects The three key phases include: (i) the pre-disaster phase, which focuses on prevention, mitigation, and preparedness to safeguard society; (ii) the warning phase, involving monitoring and forecasting hazards; and (iii) the post-disaster phase, which entails rescue and response efforts for affected populations Effective disaster management requires collaboration among government and local organizations, as it aims to overcome both natural and man-made disasters While it does not eliminate threats, enhancing mitigation and preparedness can significantly reduce risks and minimize socio-economic damage Additionally, managing disaster risk involves integrating both systemic and non-systemic measures to effectively address potential hazards.
―innovative mechanisms‖ or modern systems out of machineries and equipment
Despite advancements in disaster management, accessibility remains a challenge, particularly in developing countries, which are more susceptible to the severe impacts of disasters (Auzzir et al., 2014) Records and documentation from various global agencies highlight this vulnerability, underscoring the need for improved strategies to support these regions (Ritchie, 2004, as cited in Abdullah).
Modern disaster management technologies are lacking in developing countries, leading to slower recovery rates compared to developed nations (Othman, 2015) The challenges of flooding in these regions are exacerbated by persistent poverty and insufficient expertise in flood management (Ologunorisa & Adeyemo, 2005) Lucini (2014) emphasizes the importance of multicultural approaches to disaster resilience, as definitions of resilience vary across nations Vietnam, as a developing country, faces significant challenges in disaster management, particularly due to its population living near rivers and coastal areas, which makes them vulnerable to climate-related hazards (Luu & Meding, 2018) By adopting modern approaches, Vietnam can enhance its disaster risk mitigation and management through multidisciplinary strategies.
Flood Risk Assessment and its Frameworks
Floods are among the most devastating natural disasters, leading to significant loss of life and severe socio-economic consequences globally (Cai et al., 2021) This complex phenomenon intertwines the natural environment with societal systems (Luu & Meding).
The interaction between humans and the environment is profoundly significant, particularly in the context of natural events like flooding, which humans struggle to control (Khan & Atta-Ur-Rahman, 2005) Moreover, human activities are contributing to an increased likelihood of flood occurrences (Vojtek & Vojtekova, 2018).
2016) and intensify its impacts The International on Large Dams (International
A survey conducted by the International Commission on Large Dams (ICOLD) in 2003 identified floods as the most significant natural disaster, accounting for 65% of cases in 20 countries that host 80% of the world's largest dams The findings revealed that Asian countries bear the highest number of flood victims, highlighting the severe socio-economic impacts of such disasters It is important to note that this analysis only considers floods resulting from rivers and rainfall, excluding those caused by surges and tropical storms Additionally, risk is defined as the combination of the probability of an event and its adverse consequences, as outlined by the United Nations Office for Disaster Risk Reduction (UNISDR).
The most widely accepted definition of "risk," as outlined in the ISO/IEC Guide 73 (2009), highlights two key implications In everyday language, risk often refers to the possibility of misfortune, while in a technical context, it focuses on the potential outcomes, particularly concerning probable costs or damages.
Flood risk is defined as the potential for loss, influenced by three key factors: hazard, vulnerability, and exposure This relationship can be visualized as a triangle, where the area represents risk and the sides correspond to the three factors If any one side is absent, the risk becomes nonexistent For instance, minimizing exposure can effectively lower risk (Crichton, 1999) Additionally, flood risk encompasses both qualitative and quantitative aspects of hazards, alongside the vulnerability of stakeholders and properties to flooding and their susceptibility to damage (Cutter, 2003, as cited in Ali, Koirala, & Bajracharya, 2016) Various methodological frameworks are employed to assess these concepts of flood risk.
Flood risk assessment (FRA) is an empirical subdiscipline in connection with numerical analysis and evaluation of flood risk (Diez-Herrero & Garrote,
Flood risk assessments typically follow a framework that combines three key components: flood hazard, exposure, and vulnerability (Winsemius et al., 2013; de Moel et al., 2009; Budiyono et al., 2015; Kron, 2005; de Moel et al., 2015; Luu & Meding, 2018) Flood hazard refers to the likelihood of experiencing a specific level of risk in a given area, whether due to human activities or natural causes (Kron, 2005; Field et al., 2012) Exposure is determined by the potential impact of the hazard on the community, its livelihoods, and industries (Crichton, 1999; Field et al., 2012; Luu & Meding, 2018) Finally, vulnerability is assessed based on the community's sensitivity and responsiveness to disasters (Jongman et al., 2012; Kron, 2005; Maaskant, 2009; Kobayashi & Porter, 2012; Luu & Meding, 2018).
Various studies have employed different methodologies to assess flood risk For instance, Luu and Meding (2018) utilized a framework that enabled decision-makers to evaluate each factor, while Phong et al (2008) focused on residents' knowledge in Thua Thien Hue province, gathering data on community risk awareness Other research, such as Nguyen et al (2020), examined spatial factors influencing flood severity, including slope, elevation, rainfall, and population density, with a specific emphasis on relative slope length for flood hazard mapping Additionally, Dang and Kumar (2017) opted for hydrologic modeling to assess flood risk in urban areas like Ho Chi Minh City, diverging from the use of spatial factors and the previously mentioned components.
Multi-Criteria Decision Analysis in Flood Risk Assessment
Risk assessment is essential for understanding the extent and spatial distribution of flood risks, yet many assessments primarily focus on economic factors while neglecting environmental, societal, and cultural risks Multi-criteria Analysis (MCA) addresses these diverse risks, as highlighted by Meyer et al (2007) and Abdullah & Othman (2015), who emphasize the need for a multidisciplinary approach to tackle the complex issues posed by water-related disasters (WRD) MCA, originally introduced as Multi-criteria Decision Making (MCDM) at a 1972 international conference, has evolved in European contexts to emphasize "analysis" over "decision making" to clarify roles among stakeholders (Costa et al., 1997) For consistency, this paper will refer to the methodology as Multi-criteria Decision Analysis (MCDA).
Decision-making in environment-related research requires a careful balance of socio-political, ecological, and economic factors, which cannot be prioritized equally Multi-Criteria Decision Analysis (MCDA) has emerged as a recognized methodology that effectively utilizes technical information and stakeholder values to support rational decision-making in various fields, particularly in environmental contexts (Huang et al., 2011) This method adeptly navigates the complexities of decision-making by integrating diverse perspectives from stakeholders, making it a valuable tool for group decisions The versatility of MCDA has been demonstrated across multiple sectors, including engineering, healthcare, and industry For instance, Kechagias et al (2020) proposed a process reference model for supply chain operations to address urban client needs while managing limitations In healthcare, Adunlin et al (2014) applied MCDA to evaluate breast cancer treatment options, focusing on optimizing health outcomes and reducing costs Additionally, Cardenas (2018) advocated for a formal MCDA-based approach to quantitatively assess software products, enabling evaluators to make informed and documented decisions.
In flood risk assessment, Multi-Criteria Decision Analysis (MCDA) integrates various geospatial layers to create informative maps (Malczewski & Rinner, 2015, as cited in Luu & Meding, 2018) This methodology addresses the limitations of human ability to evaluate complex, multi-faceted information (Kiker et al., 2005) Prominent MCDA methods include the Analytic Hierarchy Process (AHP), developed by Saaty in 1980, which utilizes a mathematical algorithm to convert a matrix into a vector of relative criteria weights (Yahaya, 2008) Additionally, the Pairwise Comparison Method helps determine criteria weights, while the Ranking method allows decision-makers to prioritize criteria based on their preferences (Yahaya, 2008) Another technique, the Preference Ranking Organization Method for Enrichment Evaluation (PROMETHEE), incorporates "votes" from various elements to assess the weights of different alternatives (Huang et al., 2011).
Luu and Meding (2018) developed an AHP-based questionnaire for local staff to assess flood risk factors, incorporating sub-criteria for a weighted analysis and utilizing a Weighted Sum approach to generate maps of flood hazard, exposure, and vulnerability Similarly, Phong et al (2008) employed AHP, enhancing their study with field verification by consulting both decision-makers and households Nguyen et al (2020) compared two cases using integrated AHP and GIS, revealing that the case with six criteria produced a more detailed flood hazard map than the one with five criteria Finally, Cabrera and Lee (2019) identified three MCDM methods—AHP, WR, and ratio weighting (RW)—concluding that AHP is the most accurate method for evaluating flood hazards.
GIS-based Approach in Flood Risk Assessment
Flood risk assessment can be conducted through various approaches, including the historic approach, which relies on documented past flood events, utilizing data from maps, aerial photos, and reports The geomorphological approach studies landform changes due to flooding, aiding in predicting future disaster extents and intensities The modeling approach employs GIS software and hydro-related data for flood simulations, offering large-scale assessments quickly, though it may be hindered by cloud cover Lastly, the GIS and Remote Sensing approach is optimal for flood risk mapping, as it can gather data even in the absence of hydrologic information through satellite imagery While each method has its advantages, the choice of approach should align with the available data This study will utilize a combination of Modeling and GIS and Remote Sensing approaches, given its GIS-based focus.
Flooding types, such as flash floods, river floods, and coastal floods, vary in their occurrence frequency, magnitude, and duration, as noted by Van Westen (2000) While satellite data has proven useful for flood risk management, it falls short in flood rescue operations due to low spatial resolution and cloud cover Despite these limitations, remote sensing (RS) remains a reliable source for flood risk mapping Compared to traditional mapping methods, which are time-consuming and labor-intensive, RS facilitates easier data collection and integration with Geographic Information Systems (GIS) for efficient data management and analysis (Sharma & Joshi, 2010).
Luu and Meding (2018) evaluated flood risk in Quang Nam province, Vietnam, by combining demographic data, satellite data, and GIS They collected demographic data through field surveys and obtained satellite data, including the Digital Elevation Model (DEM), from reputable sources The data was processed and analyzed using spatial MCDA and GIS software The study concluded that remote sensing data alone is insufficient for creating local maps, GIS lacks value without accompanying data and field surveys, and traditional mapping methods are inadequate for effective disaster risk management.
Digital Elevation Models (DEMs) provide a three-dimensional representation of topography, which can be obtained by downloading files from sources like NASA and USGS or by collecting data for interpolation using GIS software Once a DEM is acquired, it facilitates the analysis of hydrologic data Key parameters for flood hazard mapping include rainfall data, distance to streams, slope, land cover, drainage density, and soil type Additionally, some studies incorporate local socio-economic factors, such as population growth rates and the diversity of ethnic groups, to enhance the analysis Notably, local knowledge has proven to be a valuable asset in flood hazard mapping, as demonstrated in Vietnam, where community insights combined with modern technology effectively identified flood-prone areas and mitigated potential risks.
Natural disasters are inevitable, and complete recovery from their devastation is often unrealistic However, the risk can be significantly minimized through the implementation of early warning systems, preparedness strategies, and effective disaster management plans.
Conclusion
Tram Tau district serves as an ideal location for testing GIS and RS capabilities in flood risk assessment This study addresses the lack of literature focused on micro-community levels, providing valuable insights for remote areas that are particularly vulnerable to flood hazards Additionally, the GIS spatial information method is highlighted as a tool for evaluating risk on a global scale and identifying local regions most at risk.
Collecting sufficient data for flood risk assessment using GIS and MCA in unfamiliar localities can be challenging; however, established criteria and indicators can facilitate the creation of accessible flood risk maps and simulations for local citizens Despite geographical challenges faced by disaster experts, the rapid advancement of GIS technology has enhanced the ability to integrate geographic methods for effective disaster risk management within communities (Thomas et al., 2007) By identifying areas at high or low flood risk, communities can implement mitigation strategies and plans, thereby enhancing safety While it may not be possible to prevent natural hazards and their impacts, collaboration with decision-makers and conducting such studies can raise awareness, promote risk preparedness, and strengthen disaster risk mitigation efforts in the community.
This chapter highlights 15 selected publications that focus on flood risk assessment (FRA), showcasing a variety of approaches While some studies utilized Multi-Criteria Decision Analysis (MCDA) and incorporated local knowledge, others concentrated solely on Hazard, Exposure, and Vulnerability, with some omitting exposure entirely An overview of the content from each study is provided in Table 1.
Table 1 Chosen FRA-related publications
Study Area
The study area is illustrated in Figure 1, which includes a map of Vietnam, a detailed map of Yen Bai Province, and an elevation map of Tram Tau District This information has been prepared by the author and is sourced from the administrative files of Vietnam available on Diva GIS's website.
Tram Tau is a rural district in Yen Bai province, located in the Northeastern region of Vietnam at coordinates 21° 32' 0" North and 104° 26' 0" East Covering an area of 742 km², it has a population of approximately 33,962, representing 4.2% of the province's total population as of 2019.
Tram Tau is 30km away from Nghia Lo Town in the southwest It is also bounded from west to south by the neighboring province of Son La A town of
Bac Yen in Son La is 60km away from Tram Tau and the road is almost not passable by the locals given its hidden nature beauty
Tram Tau features a diverse topography characterized by high mountains and deep gulfs, making agriculture and livestock the primary sources of income for its residents Situated at an average elevation of 800 meters, the district's lowest point is 390 meters above sea level The climate is predominantly warm throughout the year, with temperatures occasionally dropping below freezing The wet season typically spans from late August to early November The district is home to 11 ethnic groups, with the Hmong minority representing 77% of the population, alongside other groups such as the Thai, Kinh, and Khmu.
Methodological Framework
The Tram Tau flood risk assessment methodology, illustrated in Figure 2, comprises two main sections: the establishment of a multi-criteria system and the analysis procedure using GIS The first section features three layers: the object layer representing the Flood Risk Assessment (FRA) of Tram Tau District, the criteria layer encompassing Hazard (H) and Vulnerability (V), and the sub-criteria layer Within the Hazard index, there are seven conditional and triggering indicators, while the Vulnerability index includes three sub-criteria, resulting in a total of ten indicators in the sub-criteria layer.
The study utilizes a GIS-based approach to preprocess data from 10 indicators collected from conventional datasets and remotely sensed images The AHP method is employed in MCDA to assign weights based on the importance of each indicator These criteria weights are then overlaid using Raster Calculator tools in ArcMap software Subsequently, three scenarios are developed for the flood hazard index and flood vulnerability index, each featuring a distinct ranking of the most critical indicators The resulting hazard and vulnerability indices are computed to create a flood risk map, intended to serve as a valuable reference for decision-makers in Tram Tau district, aiding in the prioritization of flood-prone areas.
Figure 2 Methodological Framework Chart of Flood Risk Assessment in Tram Tau
Data Source and Description
Majority of data gathered in this study is provided by the GeoInformatics Research Center (GIRC) affiliated in Thai Nguyen University of Agriculture and
The rainfall data utilized in this study is sourced from the Forestry (TUAF) and verified websites such as the U.S Geological Survey (USGS) This data is in vector format, having been pre-processed by GIRC A detailed overview of the data sources is provided in the table below.
DEM (30 m) USGS - EarthExplorer Raster data/TIFF image LandSat 8 imagery 10 m USGS – EarthExplorer Visible/Optical
Administrative Area Diva GIS Vector data/polygon
Networks Land Use Map Vector data/polygon
In this part, each data collected from the sources will be deciphered; the definition of the data and the relevance to the flood risk assessment
The seven indicators used in the hazard index are crucial for evaluating flood-prone areas, primarily focusing on the site's topographical and geomorphological characteristics Each indicator is discussed in detail to highlight its significance.
Slope (Sl) - The slope is considered as one of the most critical factors that can trigger flooding As the slope gets lower, the risk of flooding rises
The slope significantly affects water movement in drainage channels and watersheds, with steeper slopes leading to increased runoff and higher water discharge Consequently, slope serves as an important indicator of flood hazard in a given area.
Rainfall (R) is the most significant indicator of flood occurrence among the ten studied indicators (Subbarayan, 2020) High levels of rainfall can lead to flooding, especially when the soil's capacity to absorb water is exceeded.
Elevation (E) plays a crucial role in flood risk, as low-lying landforms are more susceptible to flooding This factor significantly influences flow accumulation, making it a key indicator of potential flood events Water flows from higher elevations to lower terrain, highlighting the importance of understanding elevation in flood management (Kazakis et al., 2015).
The Topographic Wetness Index (TWI) is a key physical characteristic of flood-prone regions, indicating the potential for runoff formation in these areas (Kirkby, 1975; Feloni, Mousadis, & Baltas, 2019).
2019) High wetness index in the landscape shows high risk of flood accumulation due to low slope This indicator provides solid quantitative information to assess flood risk
Drainage Density (Dd) refers to the total length of streams within a specific area, as defined by Elkhrachy (2015) A denser drainage system is linked to increased flood hazards, as higher surface runoff makes the area more susceptible to flooding (Subbarayan, 2020).
Distance to river (Dr) – The distance to river is significant to indicate flooding The regions with the nearest distance to river (Rincon et al.,
2018) and flow accumulation course (Islam & Sado, 2000) are most susceptible to floods
Land Cover (Lc), often referred to as land use, describes the allocation of land in a specific area This includes various types of coverage such as vegetation, water bodies, and residential zones Understanding land cover is essential as it serves as a key indicator of whether an area is facilitating water flow on the ground (Khosravi et al.).
The vulnerability index comprises three key indicators that focus on the risk of flooding and its impact on the socio-economic characteristics of a region However, the inclusion of indicators is significantly constrained by the availability of data, as the area is classified as a micro-community.
Population density (Pd) - (Zahran et al., 2008) stated that zones that are denser in population and incompetent to stand flooding indicate more experiences in critical injuries and fatalities
Increased road density (Rd) in an area elevates the risk of flooding due to the low infiltration rate of water on road surfaces Additionally, roads facilitate rapid water runoff, leading to flash floods Furthermore, a higher density of roads often correlates with a denser population, compounding the potential impact of flooding events.
Thus, making the area more vulnerable to flooding
Distance to road (Dro) serves as a crucial indicator of accessibility during rescue and relief operations Areas closer to roads are more easily accessible for saving lives, highlighting that proximity to roads reflects the vulnerability of both residents and their livestock.
Methods
This study focuses on a GIS-based method for assessing flood risk, detailing the procedures undertaken to achieve the desired results The process is divided into two main sections: spatial data pre-processing and the delineation of maps, along with multi-criteria decision analysis.
3.4.1 Spatial Data Pre-processing and Maps Delineation
Spatial data is essential for developing flood risk maps (Adjei-Darko, 2017) It is crucial to ensure that the Digital Elevation Model (DEM) is georeferenced to Vietnam's spatial reference, specifically WGS 1984 UTM zone 48N Additionally, all data should maintain a uniform cell size of 30m x 30m This section outlines the key methods for data acquisition.
The slope is derived from a Digital Elevation Model (DEM) using the Slope spatial analyst tool in ArcMap software The resulting map (Figure 3) is categorized into five classes (Table 3) using the Natural Breaks (Jenks) method Slope measurements are expressed in degrees, with the lowest slope angles indicating a significantly high risk of flooding.
Figure 3 Slope Map (a) default slope map, (b) reclassified slope map
Elevation is represented by the Digital Elevation Model (DEM), which is processed using the Reclassify tool in ArcMap to categorize terrain into five distinct classes Higher elevations are generally less susceptible to flooding, with measurements expressed in meters (m) The accompanying map illustrates the classification of terrain from lowest to highest elevation This thematic elevation map serves as a foundation for analyzing flow accumulation, flow direction, and stream networks.
Figure 4 Elevation Map (a) default elevation map, (b) reclassified slope map
Rainfall data is typically collected as point data and interpolated using various methods, with the Kriging Method being ideal for less dense datasets In small and remote communities like Tram Tau, collecting accurate rainfall data poses significant challenges Fortunately, GIRC has provided a rainfall map for Tram Tau, detailing the annual rainfall in millimeters per year (mm/yr) Areas with higher annual rainfall are more vulnerable, and the rainfall map categorizes the data into five distinct classes.
Figure 5 Rainfall Map (a) default rainfall map, (b) reclassified rainfall map
Drainage density is derived from hydrologic data, specifically by processing the Digital Elevation Model (DEM) of Tram Tau after filling the sinks This data is used to determine flow direction and flow accumulation, ultimately producing a stream network through the Raster Calculator tool The stream network is then converted into a polyline, which serves as input for the Line Density spatial analyst tool The output is a drainage density map, measured in meters per square kilometer (m/km²) Areas with a higher concentration of drainage channels are more susceptible to flood accumulation, as detailed in the classification provided in Table 6.
Table 6 Drainage density map classification
Figure 6 Drainage Density Map (a) default drainage density map, (b) reclassified drainage density map
The buffering tool creates buffer zones around specific areas using the polyline of river or stream networks as input This data is then combined with the study area's polygon through the union tool in the geoprocessing bar The resulting map is categorized into four classes, with distances to the river measured in meters (m) Areas closest to the river are more vulnerable to flooding, increasing the risk of water overflow into nearby homes.
Table 7 Distance to river map classification
Figure 7 Distance to River Map (a) reclassified distance to river, (b) close-up look of the map
Land cover is typically categorized into three main classes: Water bodies, Vegetation, and Residential areas However, this study identifies five distinct classes: Settlements (residential areas), Agricultural Land (where crops are cultivated), Barren and Non-agricultural Land (infertile areas with minimal plant growth), Water bodies (such as rivers and lakes), and Forest and Vegetation (regions covered with trees and plants that do not require cultivation).
The land cover for the study area was derived using the image classification tool in ArcMap and LandSat 8 imagery The map creation process involved manually marking random samples on the image and applying Maximum Likelihood classification, as illustrated in Figure 9 Areas with a high density of settlements are identified as being at significant risk for serious injuries and fatalities, while forested and vegetated regions remain largely unaffected The total area of Tram Tau district is approximately 742 km², with the land cover classification and statistics detailed in Table 8 and Figure 8 It is important to note that the "Others" category, while included in the table and figure, is not represented in the land cover classification on the map The total area represented in the Land Cover map is roughly 739 km².
Table 8 Land Cover Classification and Statistics
Barren and Sparsely Vegetated 205.48 Moderate Risk
Figure 9 Land Cover Map (a) LandSat image of Tram Tau district, (b) classified
Settlements Agricultural Barren and Sparsely Vegetated Water Bodies
The Topographic Wetness Index (TWI) involves two key estimations: the hydrographic location of a cell within a catchment and the slope gradient (Feloni et al., 2019) This index is crucial for identifying areas most susceptible to flooding from rainfall events TWI is calculated using flow direction and flow accumulation derived from a projected Digital Elevation Model (DEM) and is based on a constant equation A higher wetness index indicates an increased likelihood of flooding, categorized into four distinct classes.
TWI = Topographic Wetness Index ln = natural logarithm function
FA scaled = Flow Accumulation Scaled tan slope = tan slope
First, generate slope map and convert slope in degree to slope in radian
Equation 2 Slope in radian formula
This conversion of degree to radian is same as: slope in rad = (slope in deg * pi) /
Next, get the tan slope by this formula:
When the slope is zero, the tangent of zero also equals zero, resulting in an undefined pixel value To address this issue, pixels with a value of zero are replaced with a small value, such as 0.001 or 0.00565, which represents the tangent of a nearly flat terrain with minimal slope angles.
Then, get the Flow Accumulation Scaled through this equation:
FA is a raster map generated using Flow Accumulation spatial analyst tool
In areas of the watershed where there is no flow accumulation, a value of zero is assigned, unless a value of one is added for the border pixels This zero value leads to an undefined result, as it corresponds to ln(0).
Cell size is the size of each cell making up the whole raster data In this data, the cell size is 30m
Finally, calculate the Equation 1 to get TWI
Figure 10 TWI Map (a) default TWI map, (b) reclassified TWI map
The population data for each commune in Tram Tau district is sourced from a research center studying Yen Bai province Population density, illustrated in Figure 11, represents the ratio of the population to the area A higher population density indicates a greater number of individuals at risk of flooding, as detailed in Table 10.
Table 10 Population density map classification
Figure 11 Population Density Map (a) default population density, (b) reclassified population density
The road density is determined by analyzing the polyline of the road network with the Line Density spatial analyst tool, which helps identify areas with concentrated road networks High road concentration leads to reduced water absorption in the ground due to the conversion of soil into asphalt or plastic, and human activities, such as littering, can exacerbate flash flood conditions.
12) is reclassified into four classes (Table 11)
Table 11 Road density map classification
Figure 12 Road Density Map (a) default road density, (b) reclassified road density
The distance to road map can be generated using a buffer tool, similar to the process for rivers; however, the road network is more complex and densely concentrated Due to overlapping information in the attribute table when inputs are near each line, the Euclidean Distance tool in ArcMap is employed This method produces raster data, in contrast to the vector data generated by the buffer tool The results are categorized into four classifications Settlements located farther from roads are anticipated to face greater challenges during rescue or access, highlighting the safety of those residing near roadways during emergency evacuations.
Table 12 Distance to road map classification
Figure 13 Distance to Road Map (a) reclassified distance to road, (b) close-up look of the map