Development of a Block-Scale Spatial Flood Vulnerability Index—Case Study: Morelia, Mexico

Abstract

The study of urban floods is increasingly crucial due to their growing frequency and impact on densely populated areas, often characterized by inadequate drainage and located in flood-prone zones. The consequences extend beyond physical damage, significantly affecting economies and livelihoods, necessitating substantial economic resources for recovery and infrastructure rebuilding. Urban planning now must integrate flood risk management, emphasizing not only infrastructural resilience but also comprehensive policies that address environmental and social vulnerabilities to better prepare and protect urban environments against future flood risks. This study addresses the critical issue of urban flood vulnerability through a focused analysis of Morelia, a city known for its susceptibility to flooding due to its geographical and hydrological characteristics and accelerated urban growth. Employing a multifaceted approach that integrates hydrological, socio-economic, and land use data within a Geographic Information Systems (GIS) framework, the research develops a Spatial Flood Vulnerability Index (SFVI). This index is meticulously applied at the urban block level, offering a precise mapping of flood risks across the city. By correlating the SFVI results with historical flood data, the study identifies the most vulnerable areas in Morelia, which are primarily impacted due to their proximity to water bodies, economic density, and infrastructural settings. The methodology not only highlights immediate flood risks but also aids in strategic urban planning to enhance resilience against future flooding events. This paper contributes a novel approach to flood risk assessment, providing a replicable model for similarly affected cities worldwide, aiming to balance structural measures with strategic planning tailored to local needs.

1. Introduction

Floods represent an extensive and varied area of study, whose complexity is due to the different consequences they can have, depending on the characteristics of the region affected by these extreme hydrometeorological events. According to the IPCC [1], catastrophic floods will increase in frequency along with other atypical meteorological events worldwide, and the World Meteorological Organization mentions that water-related disasters since the year 2000 have increased by 134% compared to the previous two decades [2].

In Mexico, floods account for 33% of disasters caused by extreme weather phenomena, and a quarter of the population is exposed to river and stream overflows [3]. From an economic perspective, which is useful for representing all losses due to these extreme events, the National Center for Disaster Prevention (CENAPRED) mentions that in 2016, 86.6% of the amount of economic damage and losses from natural events corresponded to hydrometeorological phenomena; of which, 70.5% corresponded to heavy rains and floods [4].

Floods are considered the most powerful and damaging “natural disaster” [5], and because of this and the complexity of their analysis which is proportional to their dimension, the problem is receiving more attention globally. Although it is true, extreme precipitation and thus extreme runoff occur naturally, anthropization along with climatic, physiographic, land use, economic, access to services, and political aspects are responsible for the impacts of floods.

The identification, analysis, evaluation, and spatial zoning of any potential threat is crucial when seeking to reduce the effects of floods. Conducting a vulnerability assessment is recognized as a necessity, which once met will help overcome the challenges that overflows cause. Moreover, there is a range of non-structural solutions that can help solve the problem, including efficient water use, avenue management, strengthening of the insurance system, generation, and dissemination of information about flood-prone areas in buying and selling processes, new construction methods, trash control, urban planning, environmental impact studies, respect for land uses, clearing of channels, among others. But to reach these solutions, it is necessary to catalog the available information and define the degree of vulnerability of the area [6,7].

Vulnerability is considered as the degree of damage that can be expected under certain conditions, of exposure, susceptibility, resilience, and state of knowledge; specifically in floods, it is the degree of fragility of a system, which marks the ability or inability to cope, recover, or adapt [8].

Faced with the insufficiency of programs and actions that allow mitigating this problem, its devastating effects are generally reflected in the most vulnerable population and infrastructure and usually cause severe damage to heritage, communication routes, urban and hydro-agricultural infrastructure, fauna, as well as various economic activities, and can even cause loss of human lives. If the population is in a state of high vulnerability, with high a probability of suffering a potential disaster, the amount of funds invested in damage repair will be very high and in turn means that the possibilities for development are limited [9].

This is why the vulnerability index is used, which must be designed to produce information for specific areas and must provide information to counteract the different hazards faced by society, in this case, floods [10]. In turn, an index is a measure that combines many individual pieces of information through a precise mathematical formula. Indices are considered useful because they allow specific objectives to be met, enable monitoring of change, comparisons between different areas of interest in space and time, help to recognize alternative dimensions of well-being, and quickly convey complex issues [11,12,13].

Connor & Hiroki [14] developed a flood vulnerability index (FVI) that considers climate change and management policies. This index is composed of several components that include meteorological, hydrogeological, socioeconomic, and countermeasure aspects, each with several indicators derived from a cause-and-effect diagram. This work promoted a line of work on the improvement of methodologies that allow determining spatial vulnerability to floods, such as the work of Balica et al. [15], which developed an index that assesses this vulnerability to riverine and urban floods, or that of Balica et al. [16] in which the index is analyzed based on climate change and its impacts on coastal cities. Jacinto et al. [17] developed a flood susceptibility index for the continental Portuguese territory, which is part of the national vulnerability index. The study proposes the aggregation of different variables representing the natural conditions of permeability, runoff, and land accumulation to determine flood susceptibility, working at the watershed scale to cover the entire Portuguese territory. Rashetnia & Heerbod [18] assess flood vulnerability in the City of Moreland, Australia, using a fuzzy rule-based index, dividing the study area into 355 grids. This approach considers hydrological, social, and economic aspects to map and assess the vulnerability of the area. The flood vulnerability index (FVI) developed in this study allows for a detailed identification of flood-prone areas.

A feature shared by the mentioned works is the scale at which the analyses are conducted, which ranges from large basins that collectively cover a country, to local scales at the city level. Therefore, the current research focuses on the development of a Spatial Flood Vulnerability Index at the urban block scale (SFVI), allowing for better detail of analysis and, therefore, vulnerable zones can be defined within a city, considering all factors that can have a positive or negative influence in this scenario and categorizing them within the parameter that corresponds to them. The purpose of developing the Spatial Flood Vulnerability Index (SFVI) is to determine flood vulnerability at the urban block scale, making it a crucial element for the future development of flood early warning systems. The traditional FVI cannot be utilized for early warnings due to its spatial scale and the lack of consideration for urban elements that either increase population vulnerability or enhance resilience. Morelia, Mexico, was selected as the study area due to its recurrent flood events that impact extensive areas. Two main rivers, the Grande and Chiquito rivers, traverse the urban zone of Morelia. These waterways have significantly influenced the urban development of the area. Historically, Morelia has faced frequent floods, but in recent years, rapid urban growth and substantial changes in land use have exacerbated these events, impacting a large portion of the population. This situation has motivated the search for a methodology to determine the vulnerability of urban blocks in the city. While the SFVI developed in this study is tailored to a specific study area, it can be replicated in any city experiencing urban flooding.

2. Materials and Methods

2.1. Case Study

Morelia is a city located in the central part of Mexico, and it is the capital of the state of Michoacán. It has an approximate area of 1.2 km² (Figure 1). The area of interest is the city of Morelia, which according to the latest census by the National Institute of Statistics and Geography (INEGI), in 2020 had 13,847 blocks designated as urban space [19]. Given its geographical location, it is an area vulnerable to flooding as it is situated in a valley near the confluence of two rivers, the Grande River and the Chiquito River.

From its origins, Morelia was prone to swamps and marshes, which caused unsanitary conditions and various problems; this was a turning point in the initiation of river channel rectification works and the draining of swamps. Subsequently, the city experienced a population growth, and these areas were urbanized, thus gradually worsening the flooding problem [20].

In 2018, the city of Morelia faced one of its worst scenarios; the city government reported that at least 50,000 people were victims of heavy rains, with most losing their household items in the 40 neighborhoods that were reported as affected. The current administration proposed projects, which have been delayed due to investment issues.

2.2. Methodology

Flood vulnerability mapping is a crucial component of early warning systems. It involves identifying regions that are most susceptible to flooding based on a combination of physical, social, economic, anthropogenic, and service characteristics. This process plays a vital role in the methods used for preventing and mitigating potential flood events in the future. Flood vulnerability mapping is a critical component of land-use planning, especially for flood-prone areas and mitigation strategies [21]. Flood hazard mapping provides easy-to-interpret graphics and maps, allowing planners to identify risk areas and prioritize mitigation activities [22,23,24].

Being able to represent flood vulnerability spatially and numerically is the key element for flood risk assessment, and over time, various methodologies have been developed to obtain these parameters. The applied methodology should prove effective in representing vulnerability and the impact it has on mitigation and adaptation to the problem [25]. Urban areas are densely populated, making them vulnerable to the impacts of floods; vulnerability is the relationship between exposure multiplied by susceptibility and the result divided by resilience. To define the SFVI for the city of Morelia, the following steps were taken.

As mentioned by Connor & Hiroki [14] and Balica et al. [15,16], the first step in developing a vulnerability index is to gather as much information as possible, to build the various indicators and subsequently the components. Based on the framework established by Karmaoui and Balica [8], information was cataloged to determine which indicators could be obtained at the block scale, resulting in the indicators shown in

Table 1. Relationship of components and indicators cataloged in exposure, susceptibility, and resilience.

Components	Indicators
	Exposure	Susceptibility	Resilience
Climatic (SFVIc)	Block in flood area (Bfa)	Communication (C)	Dams, storage capacity (Dsc)
Physiographic (SFVIph)	Cultural heritage (Ch)	Drainage system (Ds)	Evacuation roads (Er)
Land use (SFVIlu)	Disabled population (Dp)	Industries (In)	Green area (Ga)
Anthropogenic (SFVIan)	Marginalization (M)	Population growth (Pg)	Health facilities (Hf)
Economic (SFVIec)	Population density (Pd)	Unemployment (Un)
Access to services (SFVIas)	Population in flood area (Pfa)
	Proximity to river (Pr)
	Topography (T)
	Torrential rainfall (Tr)

.

Each of these indicators, represented numerically and measured in different units, necessitates standardization of information. The primary purpose of standardization is to address repetitive situations and unify criteria, thereby facilitating the comparison of data. There are various types of standardization depending on the needs; in this case, scaling was used, which is expressed in Equation (1).

$$C_i={\displaystyle \frac{C_{scale}}{C_{max}}}$$

(1)

where $C_i$ is the standardized value of one of the components, $C_{scale}$ represents the range of original values of the component, and $C_{max}$ is the maximum value of the component in question. It is important to note that each of these processes was carried out within a Geographic Information Systems (GIS) software (ArcGIS Desktop v10.8), and each characteristic was obtained for each of the 13,847 urban blocks. Once all the indicators were standardized, it was necessary to obtain the values by components; these equations are based on the models proposed by Karmaoui & Balica [8], listed from Equation (2) to Equation (7). Indicators related to exposure and susceptibility increase the SFVI, so they are placed in the numerator; however, indicators related to resilience decrease this value, so they are placed in the denominator.

$$SFV{I}_c=\left[Tr\right]$$

(2)

$$SFV{I}_{ph}={\displaystyle \frac{\left[T+Pr\right]}{\left[Dsc+Er\right]}}$$

(3)

$$SFV{I}_{lu}={\displaystyle \frac{\left[Pr+In+Pd+Ds\right]}{\left[Ga\right]}}$$

(4)

$$SFV{I}_{an}={\displaystyle \frac{\left[Bfa+Pfa+Ch+M+C+Pd\right]}{\left[Er+Hf\right]}}$$

(5)

$$SFV{I}_{ec}={\displaystyle \frac{\left[Pr+Un+M\right]}{\left[Dsc\right]}}$$

(6)

$$SFV{I}_{as}={\displaystyle \frac{\left[Dp, C\right]}{\left[Er,Hf,\right]}}$$

(7)

In the analysis of the components, it is understood that not all affect or compromise to the same extent; however, assigning a random weighting without any theoretical or numerical basis does not adequately meet the needs of the SFVI. Therefore, the decision was made to calculate the dimensionality of each component, both numerator and denominator, for each component. This analysis was conducted using Principal Component Analysis (PCA), and it determines what percentage of the dimension each indicator represents. This percentage is then multiplied by each value to subsequently obtain a single value for each component. These values, in turn, go through the same process of dimensionality calculation to ultimately obtain a single SFVI value (Equation (8)) for each of the 13,847 urban blocks in the city of Morelia.

$${SFVI}_{Total}=\sum SFV{I}_i$$

(8)

Vulnerability intervals were standardized on a scale from 0 to 1 using the natural breaks (Jenks) classification, where classes are based on the natural groupings inherent in the data. Class breaks are created so that similar values are grouped more effectively, maximizing the differences between classes [26]. Entities are divided into classes whose boundaries are established where there are significant differences between data values [27]. These intervals are shown in

Table 2. Designation of ranges from 0 (Not Vulnerable) to 1 (Highly Vulnerable).

SFVI Value	Designation
0.00–0.27	Low
0.28–0.33	Medium Low
0.34–0.42	Medium
0.43–0.63	Medium High
0.64–1	High

.

3. Results

With the applied methodology, it was possible to spatially represent areas that were previously recognized as vulnerable only empirically or through experience. Below are the results of the calculated components; maps of the indicators are shown in the annex.

3.1. Climatic Component (SFVIc)

The climatic component identifies areas most vulnerable to precipitation. It considers low-lying and runoff areas where water moves more slowly due to topography and urban settlement (Figure 2).

3.2. Physiographic Component (SFVIph)

The analysis of the physiographic component positively influences the issue, particularly in aspects such as identifying areas equipped to serve as evacuation routes. It also highlights vulnerable zones, such as those close to rivers. Understanding the topography of the study area helps identify at-risk areas located in low-lying zones (Figure 3).

For this component, the Topography (T) indicator was derived from the Digital Elevation Model with a 1:10000 scale and 5 m resolution, obtained from the INEGI database (https://www.inegi.org.mx/temas/topografia/ (accessed on 20 June 2024)), to determine the maximum slopes in the study area. The Proximity to river (Pr) indicator considers urban blocks that are directly adjacent to the river or drains, with a buffer of twenty meters, making them highly vulnerable to flooding.

The Proximity to river (Pr) indicator accounts for blocks that are directly adjacent to the river or drainage systems, with a setback of twenty meters, identifying them as areas highly vulnerable to flooding. For the Dams, storage capacity (Dsc) indicator, which is considered a resilience indicator, a buffer zone of 5 m downstream from the dam located in the study area is used. This zone acts as a buffer for extraordinary releases from the dam.

The Evacuation roads (Er) indicator was derived using data collected by the brigade from the Municipal Planning Institute of Morelia (IMPLAN) in 2019. This survey assessed the presence or absence of such routes and the condition of the roadways in Morelia; the survey included five types of surfaces: concrete pavement, asphalt pavement, cobblestone pavement, stone pavement, dirt roads, pedestrian paths, and roads under construction.

3.3. Land Use Component (SFVIlu)

Land use involves the actions, activities, and interventions that people perform on a certain type of surface to produce, modify, or maintain it. The results of this component are displayed on the map in Figure 4.

To obtain the Industries (In) indicator, information was used from the National Statistical Directory of Economic Units (https://en.www.inegi.org.mx/app/mapa/denue/default.aspx (accessed on 15 July 2024)), which provided data on the different business scales present in the city. These businesses are categorized by various aspects, including the number of people they employ, which was used to construct this indicator. Data ranged from a minimum of 0 to 5 people, but for this case, the information on industries employing between 11 and 30 people was used, as the loss of these places would represent a negative impact on many people.

For the Disabled population (Dp) indicator, a population growth map was made. However, it was not considered suitable to generate the indicator because, although the population growth has been considerable, it has been generated towards the periphery of the city, and in the case of the city center, instead of growth, there has been a decrease. It is believed that this is because Morelia is a student, tourist city, and a city where the population tends to migrate to other states. This causes constant movement in the population distribution, and the change in urban blocks is considerable because more blocks are generated, some are divided, and others are joined, so measuring population growth per block is a task that does not adequately represent an indicator that can be considered. Therefore, it was considered to work with the population density indicator, since it only considers the amount of population among the area of the block.

The Drainage system (Ds) indicator was constructed from information from the INEGI 2020 census.

The Green area (Ga) indicator was generated from information provided by IMPLAN (https://sigmorelia.gob.mx/ (accessed on 15 July 2024)), which consists of points that they consider as green areas and with the help of the Google Earth tool, where areas that can also allow infiltration due to their low density of housing are observed.

3.4. Anthropogenic Component (SFVIan)

Anthropogenic components refer to outcomes and processes resulting from human actions, which are necessary for life as it is known. However, while these processes are helpful and necessary, they also cause impacts. This is why it is considered necessary to perform this analysis to identify the flood vulnerability index. This component is shown in Figure 5.

With the results from the hydraulic modeling performed by the National Water Commission (CONAGUA) in 2016 for a 100-year return period, it was possible to construct the Block in flood area (Bfa) indicator, as it allows for the intersection of maps to identify which urban blocks could be affected in the event of extreme precipitation.

To generate the Population in flood area (Pfa) map, like the previous case, the CONAGUA hydraulic modeling with a 100-year return period was used. The difference is that in this case, the percentage of the population that would be affected in the event of extreme precipitation and, therefore, the flooding event was calculated.

For the map of the Cultural heritage (Ch) indicator, the database from the IMPLAN was also used, which includes an updated list of centers that can be considered cultural, from theaters to libraries. The percentage of this value was calculated in relation to the distance from the block to the nearest cultural center.

The Communication (C) map was created using data from the 2020 INEGI census, which considers the number of people who own a cell phone out of the total population, for each block. Despite being in an advanced technological era, the proportion of the population with access to a cell phone is not as high as expected.

The information needed to perform the Health facilities (Hf) indicator was obtained from the Unique Key of Health Establishments Catalog (CLUES) (http://www.dgis.salud.gob.mx/contenidos/intercambio/clues_gobmx.html (accessed on 15 July 2024)), the official directory at the national level, which is mandatory for inclusion and use in the technological tools and projects of all institutions belonging to the National Health System, as well as by the central areas of the Ministry of Health. Therefore, no more than 500 m was considered adequate to count as optimal access to a health facility. It is important to mention that a purification of these centers was carried out to only include clinics and not administrative buildings.

3.5. Economic Component (SFVIec)

Economic components are those that influence and ensure the proper functioning of a society’s wealth. Indicators such as unemployment and the marginalization index increase the flood vulnerability index. In contrast, indicators that safeguard the integrity of structures where all kinds of activities such as economic, tourism, and even daily life are conducted are those that prevent the SFVI from rising excessively (Figure 6).

The Unemployment (Un) indicator was created using data from the 2020 INEGI census. The variables used to calculate this ratio were the employed population versus the economically active population. From this operation, the data for the indicator and the respective map were obtained.

3.6. Access to Services Component (SFVIas)

Access to services is a critical component, recognized as both a necessity and a priority. Local governments are judged on their ability to provide quality services to citizens. However, this task is complicated by the rapid changes in urban development, demographics, and the environment. Therefore, it is essential to start by identifying each of the indicators that are part of both the problem and the solution, and to promote improvements that make the city increasingly less vulnerable in general terms, and specifically less vulnerable to flooding in this case.

The information to create the Disabled population (Dp) indicator was obtained from the INEGI 2020 Census. Since the data were generally worked by urban blocks, the density of the disabled population was calculated by dividing the number of disabled individuals by the total population. The Access to Services Component is shown in Figure 7.

The previously obtained components result in the Spatial Flood Vulnerability Index (SFVI) when applied to each of the 13,847 blocks using Equation (8), producing the map displayed in Figure 8.

Figure 8 illustrates that areas adjacent to drains are recognized as the most vulnerable. However, the vulnerability ranges and the degree of importance that these areas should have in the event of an extreme precipitation event are also evident. Each of the indicators that make up this index is examined, and an analysis of the areas that can be improved to decrease the SFVI value is feasible.

In 2018, the city of Morelia faced one of the worst floods recorded in recent years, during the presence of Hurricane Willa and Tropical Storm Vicente. Numerous neighborhoods and inhabitants were affected as both the Grande River, which runs through the city, and the drains along the city were overwhelmed.

Following this event, IMPLAN flew a drone over a specific area to capture an orthophoto of images that showed the magnitude of the flood. This aerial photograph was used to validate the obtained Flood Vulnerability Index by overlaying the orthophoto on the map from Figure 8 and confirming that the areas marked in red within the image are the same ones that are flooded. It is even possible to see the distinction with the area where more green spaces are visible, and the color representing their vulnerability is in shades of lower value (Figure 9).

As expected, the areas where vulnerability concentrates are regions of medium-high to very high flood risk. These are characterized as flat areas with little slope, lower elevation, low drainage density, and proximity to the river, all of which are significant conditioning variables for flood hazard mapping. There are also areas of medium-low to low flood risk, primarily located along the upstream section of the basin. These areas are distinguished by their steep slopes, higher elevations, and other factors that reduce their flood vulnerability index, confirming the effectiveness of the map representation obtained.

4. Research Limitations

The primary limitation of this research lies in the availability of spatial information for the components required to determine vulnerability at the urban block-scale. This could pose a challenge when attempting to replicate the methodology in areas where data at the required scale is not available.

In this study, IMPLAN and INEGI, two governmental institutions that publish official data for the city of Morelia and Mexico, respectively, provided the necessary information. The data from these sources are collected and processed by experts in the field, making them highly reliable. The data are originally gathered at the urban block-scale, eliminating the need to modify the scale of the data used in this work. Additionally, the information is collected every five years, ensuring that the data can be considered up to date.

Given that the primary aim of this research is vulnerability assessment, it does not include the evaluation of flood hazard and risk. Consequently, an exhaustive analysis of surface runoff in the rivers was not conducted, nor was the impact of land use and soil type on river flows examined. However, the historical distribution of torrential rainfall in the study area is considered, as illustrated in Figure 2, which highlights a higher frequency of torrential rainfall in the southern zone.

5. Conclusions

Floods represent a persistent issue for the city of Morelia, impacting beyond just the economic aspect. The development of a map displaying the city’s flood vulnerability index serves as a strategic planning tool. This tool can be enhanced to integrate into early warning systems, risk maps, and any plan that boosts the city’s resilience against such challenges.

It was possible to construct and propose a Spatial Flood Vulnerability Index (SFVI) from data processing. This index, validated as mentioned earlier and considering local expert opinions, is deemed reliable as it accurately represents an ongoing issue in the city. The results of the SFVI indicate that while the highest flood vulnerability is indeed concentrated around river areas, other regions are also at risk due to social, economic, and topographical factors. This underscores the importance of not limiting flood mitigation efforts solely to river-adjacent areas but rather conducting comprehensive analyses that encompass the entire city. This broader approach will ensure that all vulnerable areas are addressed, enhancing the overall effectiveness of flood resilience strategies.

The vulnerability map, along with previously created maps highlighting the spatial location of critical sites like hospitals and shelters, industrial areas, and groups such as marginalized populations or those with restricted access to information, offer crucial insights. These maps do not only serve to locate these features but also to understand the needs that should be addressed to reduce vulnerability to flooding, transforming each factor into a resilience indicator rather than just susceptibility.

Flood resilience is not coincidental but results from a society taking proactive measures to prepare for floods, both preemptively and responsively. Flood preparedness might involve evaluating plans and policies, enhancing community response capabilities during a flood, and identifying post-flood recovery strategies. After a flood, resilience efforts can include rapidly communicating information to residents, understanding available community and individual recovery resources, assisting residents with mitigation measures for repairs or reconstructions, or identifying available programs for property buyouts or other mitigation projects.

The SFVI can help reduce the impact of floods to be less costly compared to previous events by enabling the community to plan strategies effectively.

The index serves as a database and is easily updatable due to the methodological approach taken—analyzing every urban block currently in the city.

The data analysis provides a deeper understanding of Morelia’s current situation, which is crucial given the number of floods experienced and the problems caused to the population. Implementing this system offers a considerable advantage by providing valuable time for the timely development of evacuation measures or response actions to potential disasters caused by floods.

It is important to emphasize that this is not a battle against water or nature, but about educating ourselves on how to coexist with this element of our environment, which, in most cases, offers more benefits than challenges. The methodological process followed is replicable for any city facing these issues, due to the adaptability of the indicators used, which can be adjusted or modified depending on local needs. The proposed methodology allows for a detailed evaluation of flood vulnerability in a city by working at the urban block scale, considering the resilience or exposure elements of these small areas.