Flood Risk Assessment for Sustainable Transportation Planning and Development under Climate Change: A GIS-Based Comparative Analysis of CMIP6 Scenarios

Abstract

Climate change is causing a range of environmental impacts, including increased flood frequency and intensity, posing significant risks to human populations and transportation infrastructure. Assessing flood risk under climate change is critical, but it is challenging due to uncertainties associated with climate projections and the need to consider the interactions between different factors that influence flood risk. Geographic Information Systems (GISs) are powerful tools that can be used to assess flood risk under climate change by gathering and integrating a range of data types and sources to create detailed maps of flood-prone areas. The primary goal of this research is to create a comprehensive GIS-based flood risk map that includes various climate change scenarios derived from the Coupled Model Intercomparison Project Phase 6 (CMIP6) models. This goal will leverage the Analytic Hierarchy Process (AHP) methodology to better understand the impacts of these climate change scenarios on the transportation network. Furthermore, this study aims to evaluate the existing flood risk map and assess the potential impacts of prospective climate scenarios on the levels of flood risk. The results showed that the northern and coastal regions of the United Arab Emirates (UAE) are at higher risk of flooding, with the majority of the population living in these areas. The projections for future flood risk levels indicate that under the SSP245 scenario, flood risk levels will generally be low, but some areas in the northern and eastern regions of the UAE may still face high to very high flood risk levels due to extensive urbanization and low-lying coastal regions. Under the SSP585 scenario, flood risk levels are projected to be significantly higher, with a widespread distribution of very high and high flood risk levels across the study area, leading to severe damage to infrastructure, property, and human lives. The recent publication of the CMIP6 models marks a significant advancement, and according to the authors’ knowledge, there have been no studies that have yet explored the application of CMIP6 scenarios. Consequently, the insights provided by this study are poised to be exceptionally beneficial to researchers globally, underscoring the urgent necessity for holistic sustainable flood risk management approaches for geography, planning, and development areas. These approaches should integrate both sustainable transportation infrastructure development and risk mitigation strategies to effectively address the anticipated impacts of flooding events within the study region.

1. Introduction

Climate change, which includes long-term shifts in the Earth’s climate patterns, notably rising temperatures, elevated atmospheric carbon dioxide concentrations, and altered precipitation patterns, has profound and intricate implications for transportation systems worldwide [1]. One of its most significant impacts is the transformation of the natural environment. Climate-induced variations in temperature, precipitation, and sea levels can trigger ecosystem shifts, affecting plant growth patterns and animal migration routes. These changes not only threaten species with extinction but also result in a decline in critical ecosystem services, such as pollination, water purification, and carbon sequestration—aspects that directly intersect with the integrity and functionality of transportation infrastructure. The influence of climate change on human societies further underscores its relevance to transportation infrastructure. Rising temperatures and changing precipitation patterns lead to extreme weather, affecting agriculture, water resources, and infrastructure. This impacts transportation planning and asphalt performance, highlighting significant regional weather variations affecting pavement design [2]. Additionally, assessments of UAE’s precipitation through GCMs highlight the need for adaptation due to significant projected rainfall variability [3].

Economic consequences, yet another facet of climate change’s repercussions, hold critical importance for transportation infrastructure. Altered weather patterns can disrupt industries like agriculture, fishing, and forestry, leading to diminished productivity and increased operational costs—factors that ripple through supply chains and impact transportation systems that facilitate the movement of goods and people. Furthermore, the rising sea levels associated with climate change pose a substantial threat to coastal communities and the transportation infrastructure that supports them, translating into property loss and augmented costs for protective measures, making adaptation and resilience a paramount consideration for coastal transportation infrastructure. In essence, climate change’s multifaceted impacts on the natural environment, human societies, and the global economy have intricate and far-reaching consequences for transportation systems. Adaptation and mitigation measures are essential to ensure the resilience and functionality of transportation networks in the face of a changing climate.

The adverse consequences of floods can be exacerbated under climate change, and it is, therefore, essential to understand and assess flood risk in a changing climate. Assessing flood risk under climate change is critical because floods pose a significant threat to human populations, transportation infrastructure, and the environment. As climate change alters weather patterns, the frequency and intensity of floods are expected to increase, making it essential to understand and manage flood risk. Floods can have severe consequences, including loss of life, damage to property and transportation infrastructure, and disruptions to communities and economies. The adverse impacts of floods can be exacerbated under climate change, as rising sea levels, changing precipitation patterns, and more frequent extreme weather events can increase the likelihood and severity of flooding [3,4]. A study analyzed urban expansion in over 6000 cities, tracking impervious surface area, population, and GDP changes, correlating these with economic growth and flood risks, to inform sustainable development strategies [3]. Another study that utilized CoastalDEM, a neural network-enhanced DEM, reveals that 190 million people currently live below projected 2100 high-tide lines under low emissions, tripling past estimates [4].

In the face of accelerating climate change, understanding its impacts on flood vulnerability emerges as a critical scientific and practical issue, particularly within the arid landscapes of the UAE. This research directly confronts the challenges posed by changing climate patterns on flood risk assessments, focusing on the vulnerabilities of transportation infrastructure in the UAE. By identifying and analyzing the specific factors that increase flood risks in this unique geographical context, our study seeks to inform and enhance the resilience of infrastructure planning and management against the backdrop of climate change. The emphasis on the UAE not only addresses a gap in the current literature but also serves as a case study for similar arid regions grappling with the complexities of climate-induced flood risks.

Assessing flood risk under climate change involves understanding the exposure, vulnerability, and hazards associated with flooding [5]. This allows for the identification and mapping of flood-prone areas and the estimation of the number of people, buildings, and critical transportation infrastructures located in these areas. It also involves understanding the characteristics of the assets at risk of flooding and the likelihood and severity of flooding in a given area. A key challenge in assessing flood risk under climate change is the uncertainty associated with climate projections [5]. Future climate conditions are subject to uncertainty due to various factors, including the complex nature of the climate system and the limitations of climate models [6]. As such, it is essential to use a range of climate projections and consider a range of possible outcomes in flood risk assessments. Another challenge is the need to consider the interactions between different factors that influence flood risk. For example, land-use changes and changes in precipitation patterns can influence the exposure and vulnerability of assets to flooding [7]. It is, therefore, important to consider these interactions when assessing flood risk. To mitigate the impacts of flooding under climate change, various strategies can be employed. These include structural measures such as flood walls, levees, and drainage systems, as well as non-structural measures such as land-use planning, early warning systems, and insurance [7]. The most effective strategy will depend on the local context, the level of flood risk, and the available resources.

Geographic Information Systems (GISs) are powerful tools that can be used to assess flood risk under climate change. GIS can be used to identify and map flood-prone areas; estimate the number of people, buildings, and critical infrastructures located in these areas; and model the potential impacts of flooding [7]. GISs can be used to gather and integrate a range of data types and sources to assess flood risk. These can include data on topography, land use, hydrology, climate, and infrastructure [8]. By integrating these data types, GISs can provide a comprehensive understanding of flood risk in a given area. One of the key strengths of a GIS in assessing flood risk is its ability to visualize and analyze spatial data. GISs can be used to create detailed maps of flood-prone areas and identify the assets located in these areas. This can assist decision-makers in prioritizing mitigation efforts and efficiently allocating resources.

Floods stand as among the most destructive natural calamities, inflicting substantial harm on transportation infrastructure, claiming lives, and disrupting routine activities. GISs have proven to be effective tools for identifying and mapping areas at high risk of flooding. Many studies have utilized GISs to assess flood risk in different parts of the world. For instance, Kourgialas and Karatzas utilized a GIS to assess flood risk on the Mediterranean island of Crete, Greece [9]. Six factors are integrated within a GIS framework to pinpoint the areas most susceptible to flooding, delineating five distinct flood risk levels that gauge the likelihood of a flooding occurrence within a single hydrological year. Rubio et al. evaluated flood risk in Metro Manila, the Philippines, employing a series of indicators tailored to urban environments, encompassing physical, social, economic, and ecological aspects [10]. The AHP method was used to determine optimal weights for each indicator. They then created spatial maps showing five levels of flood risk ranging from very low risk to very high risk using a GIS. Vanolya and Jelokhani-Niaraki investigated flood hazard maps in Mazandaran, Iran, using subjective and objective weights [11]. This study evaluates the accuracy of these maps based on historic flood occurrences within them and their assessment ratio. The results show that both types of weights produce similar results for very high-class floods but differ for other strategies with better values shown by subjective–objective weights. Osei et al. used a GIS-based approach with Digital Elevation Models (DEMs) to delineate flood-prone zones within the Tarkwa mining area [12]; morphological characteristics such as local slope and streamflow network were constructed using ArcGIS Pro 3.0. These features were subsequently employed to create a flooding susceptibility map by combining the slope map with the stream network maps.

While most of the studies focused on developing risk maps based on the current situation, only a few considered how climate change will impact the areas affected by flood. Oubennaceur et al. examined the influence of climate change on flood risk in two municipalities situated in southern Quebec, Canada [13]. The study employed a modeling methodology to forecast forthcoming river flows and simulate flood hazard maps based on anticipated climate scenarios for the years 2050 and 2080. The results show that while peak-flow frequencies are expected to decrease slightly over time, depth, inundation areas, and monetary losses will also experience minor decreases. While the studies discussed in the conclusion demonstrate the effectiveness of GISs in assessing flood risk and developing flood hazard maps, there is still a research gap in understanding the social and economic impacts of floods on communities. Most of the studies mentioned in the conclusion focused on developing flood risk maps based on physical factors such as topography and hydrology. However, it is important to understand how floods affect communities, including the impact on infrastructure. Further exploration is essential to understand how climate change may affect flood hazards moving forward, especially in regions where infrastructure faces the greatest risk. To the best of our awareness, the application of CMIP6 scenarios in this context has not been previously explored. Thus, the primary aim of our study is to create Geographic Information System (GIS)-powered maps that depict flood risks under various climate change projections. The detailed goals of this study include the following:

• Collection of historical climate databases from local weather stations;
• Obtaining historical and future climate change scenario data from GCMs models;
• Adopting and applying bias correction and downscaling methods to produce future climate projections;
• Suggesting a structured framework for evaluating the effects of climate change on flood vulnerability;
• Creation of a flood risk map utilizing GIS under various climate change scenarios using CMIP6 projections.

2. Study Area and Observed Data

The United Arab Emirates (UAE), located at the eastern end of the Arabian Peninsula, is a member of the Gulf Cooperation Council (GCC), situated from 22°50′ to 26° north latitude and 51° to 56°25′ east longitude, encompassing 83,600 km². It boasts a coastline of about 1318 km along the Arabian Gulf and the Gulf of Oman. The southern shoreline of the Arabian Gulf is distinguished by large salt flats that reach into the interior. The country’s landscape is primarily dry, featuring extensive sand deserts, dunes, oases, rugged mountains, valleys, wetlands, mangrove forests, and salt flats.

Abu Dhabi, one of the emirates in the UAE, hosts most of the nation’s oases, well known for their dense date palm populations. Situated in a region with a hot, arid climate, it experiences average temperatures ranging from 26 °C to 33.5 °C. The period from December to March marks the rainy season in the UAE, with annual precipitation typically falling between 140 and 200 mm, though areas with elevation may receive as much as 350 mm. Severe dust storms, or shamal winds, significantly impact visibility across UAE airports, influencing transportation safety. Other studies exploring the negative influences of adverse weather on driver behavior collectively point to the expansive consequences of climate change on transportation networks. This body of work stresses the necessity for improved predictive tools and safety protocols to lessen these impacts and bolster the resilience of transportation infrastructure [14,15]. Figure 1 illustrates the geographical and infrastructural features of the UAE. The climate data utilized in this analysis were collected from a variety of weather stations throughout the UAE, supplied by the National Center of Meteorology. In addition to climate data, this study utilized high-resolution land cover data from the Sentinel-2 10-m Land Use/Land Cover dataset provided by ESRI [16]. This comprehensive dataset allowed for detailed analysis of land cover variations across the UAE, crucial for assessing the impact of land use on flood risk. The Sentinel-2 data, with its 10-meter resolution, encompasses a wide range of land cover types, including urban fabric, agricultural lands, natural vegetation, water bodies, and barren areas. Following the acquisition of climate and land cover datasets, topographic data were sourced from the Shuttle Radar Topography Mission (SRTM) [17]. The SRTM data form a Digital Terrain Model (DTM) with a resolution of 1 arc/second (~30 m), designed to provide a detailed representation of the Earth’s bare surface. This resolution is particularly suited for flood risk assessment as it allows for the precise delineation of elevation contours. The SRTM data, which were collected during the 11-day Space Shuttle Endeavour mission in February 2000, have since become a benchmark for global topographic data used in various geospatial analyses.

3. Global Climate Model Data

The Coupled Model Intercomparison Project Phase 6 (CMIP6) is a multi-model initiative that provides updated climate projections for a wide range of scenarios. Shared Socioeconomic Pathways (SSPs) are a set of scenarios that describe different socioeconomic conditions and emissions pathways for the future. In this paragraph, we will discuss the CMIP6 data under different SSPs. The CMIP6 data provides valuable information on how the Earth’s climate will respond to different socioeconomic conditions and emissions scenarios [18]. The UAE study maps temperature changes using GCMs, predicting significant increases, thus aiding decision-making for climate change mitigation strategies [19].

Under the SSP1-2.6 scenario, which represents a future with rapid decarbonization and sustainable development, the global mean surface temperature is projected to remain below 2 °C above pre-industrial levels, which is the target of the Paris Agreement. However, under the SSP5-8.5 scenario, which represents a future with high greenhouse gas emissions and limited climate policies, global mean surface temperature is projected to increase by over 4 °C by the end of the century, resulting in significant impacts on ecosystems and human societies. In addition to temperature projections, the CMIP6 data also provide information on other climate variables, such as precipitation, sea level, and ocean acidity. Under the SSP1-2.6 scenario, precipitation is projected to increase in many regions, particularly in the tropics and high-latitude areas, while sea level rise is expected to be limited to around 0.5 m by 2100. However, under the SSP5-8.5 scenario, precipitation is projected to decrease in many regions, particularly in the subtropics and mid-latitudes, while sea level rise is expected to exceed 1 m by 2100. Furthermore, the CMIP6 data also show that the impacts of climate change will vary depending on the region and sector. For example, under the SSP1-2.6 scenario, the Arctic is expected to experience significant warming, leading to the widespread melting of sea ice and permafrost. This could have significant impacts on the region’s ecosystems and indigenous communities. In contrast, under the SSP5-8.5 scenario, the tropics are expected to experience more frequent and intense heatwaves, which could have significant impacts on agriculture, human health, and transportation infrastructure.

In this study, two Shared Socioeconomic Pathway (SSP) scenarios, specifically SSP2-4.5 and SSP5-8.5, were chosen for two CMIP6 models, CanESM5 and MIROC6, spanning both near-future (2021–2050) and far-future (2051–2100) timeframes [20,21]. The advantage of utilizing SSP scenarios lies in the ability to integrate estimated changes into flood risk with socioeconomic projections under varying SSP scenarios to explore flood risk dynamics. Presently, these two SSP scenarios incorporate a total of 19 CMIP6 General Circulation Models (GCMs). Each climate model encompasses daily precipitation data for both historical (2000–2020) and future (2021–2100) periods.

4. Methodology

Figure 2 presents the approach used to estimate flood risk vulnerability in the context of climate change scenarios, which consists of two primary phases: first, assessing the accuracy of CMIP6 global climate models by comparing them to historical data; second, using CMIP6 model outputs to forecast future flood risks, incorporating the evaluation and prioritization of factors through the Analytic Hierarchy Process (AHP). The preliminary analysis of CMIP6 GCMs involves a comparison of model-generated rainfall data against actual precipitation figures from a reference period in the past. The consistency between the modeled and real-world data acts as a measure of the CMIP model’s proficiency in mirroring climate patterns pertinent to this research.

A high degree of concordance suggests that CMIP models are instrumental in exploring the effects of climate change scenarios on flood events. The reference period for historical climate model simulation data spanned from 2000 to 2020, with the periods 2021–2050 and 2051–2100 designated as the near and far future, respectively, based on SSP scenarios in line with the IPCC Sixth Assessment Report’s guidelines. The calculation of mean annual precipitation was carried out for both the historical and future periods using each of the CMIP6 global climate models. The calculated baseline period values were subsequently incorporated into the authors’ model for assessing national susceptibility to rainfall-induced flooding, which considers the impact of mean annual precipitation.

4.1. Statistical Downscaling and Global Climate Model Evaluation

Climate models at both global and regional scales are pivotal for projecting future climate variations. Numerous studies have applied these models, often incorporating bias correction techniques and real-world measurements to rectify consistent discrepancies in the prediction of weather elements such as temperature and precipitation. In the context of this study, bias correction was performed using the delta method, a strategy celebrated for its straightforwardness, efficacy in correcting biases, widespread use in the scientific literature, and ability to withstand comparison with other top methods. The delta change technique allows for the application of General Circulation Model (GCM) projections in hydrological modeling and studies at the watershed level, though in a somewhat indirect manner. It operates on the principle of applying change factors or ratios derived from comparing future projections to historical averages. These factors, when applied to historical observational series, adjust the data to reflect future climate scenarios. For the purpose of determining delta correction factors in this research, data on average monthly precipitation collected over a two-decade period (2000 to 2020) from 21 meteorological stations throughout the UAE were analyzed. Corrections were made for data under various Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5) employing the delta method. The computation of the delta factor was based on the following equation [22]:

$$\Delta \mathrm{m}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{h} \mathrm{i} = \frac{\mathrm{F}\mathrm{u}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{e} \mathrm{a}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e} \mathrm{m}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{l}\mathrm{y} \mathrm{f}\mathrm{o}\mathrm{r} \mathrm{m}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{h} \mathrm{i} (\mathrm{f}\mathrm{r}\mathrm{o}\mathrm{m} \mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l} \mathrm{u}\mathrm{s}\mathrm{e}\mathrm{d})}{\mathrm{H}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l} \mathrm{a}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e} \mathrm{m}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{l}\mathrm{y} \mathrm{f}\mathrm{o}\mathrm{r} \mathrm{m}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{h} \mathrm{i} (\mathrm{f}\mathrm{r}\mathrm{o}\mathrm{m} \mathrm{a} \mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l} \mathrm{u}\mathrm{s}\mathrm{e}\mathrm{d})}$$

Given the varied resolution of CMIP6 GCM data and the notable spatial diversity in the UAE’s landscape, which influences precipitation and temperature projections, the technique of climate data interpolation was utilized to align accurately with other critical variables within the flood risk vulnerability model. Climate models provide data as a function of latitude and longitude along predefined grid lines covering the entire Earth’s surface. Adjusting the data for UAE weather stations necessitated interpolation or extrapolation for each model. Interpolation entails predicting data from available datasets, while extrapolation involves making predictions beyond the dataset. Three types of interpolation were utilized to determine precise rainfall for each area based on the station’s location within each model’s grid. Interpolation was required when a station fell between two points, and interpolation between four points was necessary when it was situated in the middle of a square formed by those points.

4.2. Flood Risk Analysis

Flooding poses a critical danger to the safety of both public and private transportation networks [23], impacting worldwide transportation safety and infrastructure development efforts. In evaluating the spatial variation in flood risk across the study zone, five unique flood risk levels (RLs) were analyzed, identified as very high, high, moderate, low, and very low. These RLs were defined based on the likelihood of a flood occurring within a given hydrological year. Zones identified with a very high flood risk corresponded to a 2% probability of a flood event within a hydrological year (akin to a 50-year recurrence interval), while high-flood-risk areas had a 1% chance (matching a 100-year recurrence interval). A moderate flood risk was associated with a 0.5% chance (corresponding to a 200-year recurrence interval). Areas under the very low flood risk category showed a likelihood of less than 0.2% (surpassing a 500-year recurrence interval), and those deemed to have a low flood risk presented a probability of 0.2% or less.

In the evaluation of flood risk factors, special attention was given to the role of elevation in determining flood susceptibility. Our analysis associated higher elevation levels with an increased risk of flash flooding. This decision was grounded in the specific geographical context of our study area, where steep terrains are prevalent, and the hydrological dynamics can lead to rapid water flow and accumulation during heavy rainfall events, thus precipitating flash floods. This phenomenon is not only theoretically supported but also evidenced by historical flood events within the region [24,25].

This classification considered the elements that determine the occurrence of floods. The final flood risk map was generated by combining seven separate theme map components that are closely related to flooding incidents. Based on a linear algebraic function that assigned relative weights to each element, this combination was determined. Within the ArcGIS environment, four separate thematic map-factors, including land cover (L), rainfall (R), elevation (E) and slope (S), were generated to estimate the final flood risk map.

The influence of each factor was categorized into five specific flood risk levels (RLs): very high, high, moderate, low, and very low. These distinct flood risk categories for each factor were established using Jenk’s Natural Breaks method. Subsequently, a numerical score was allocated to each factor, quantifying the flood risk levels (RLs) in this manner: very high (RL) scores 10, high (RL) scores 8, moderate (RL) scores 5, very low (RL) scores 2, and low (RL) scores 1.

The Analytic Hierarchy Process (AHP) was used to assign weights to the significant elements since not all investigated factors have the same effect on flood generation. The AHP is a useful technique for many decision-making procedures in numerous disciplines, including finance, transportation, engineering, business, education, and politics. The AHP is the most often used method for multi-criteria decision-making (MCDM). Numerous researchers have recreated the AHP due to its numerous advantages in making vital decisions [26,27,28,29]. The Analytic Hierarchy Process (AHP) is a comprehensive mathematical framework that facilitates the assessment of both tangible and intangible elements in decision-making scenarios, taking into account the preferences of individuals or collective bodies. This method unfolds across various decision-making phases, during which values are attributed and alternative assessments are made according to the selected criteria. The procedure is initiated by establishing a hierarchical framework and creating a comparison matrix. This matrix is subsequently converted into a priority vector, followed by the computation of the consistency ratio using a specific random index value. Figure 3 displays the diagrammatic depiction of a three-tier hierarchical structure applied to a multi-criteria decision-making (MCDM) issue, as detailed by [30].

At the topmost tier is the goal of the decision, succeeded by the criteria at the intermediary level, and, when relevant, the bottom tier features the options under consideration. These alternatives are delineated at the most fundamental level. To maintain accuracy in the pairwise comparisons, it is crucial to clearly specify both the total number of criteria and the individual details of each criterion. Criteria ought to be categorized according to their common attributes, facilitating the application of AHP across a broad spectrum of criteria. Within the AHP, the significance of each criterion is established through pairwise comparisons once the hierarchical framework is in place. The comparative significance of each criterion is denoted by a scale from 1 to 9, based on pairwise evaluations, as shown in

Table 1. Scales for pairwise comparison in the AHP.

Importance Intensity (Scores)	Definition
1	Equal importance
3	Moderate importance of one over another
5	Strong importance of one over another
7	Very strong importance of one over another
9	Extreme importance of one over another
2, 4, 6, 8	Intermediate values
Reciprocals	Reciprocals for inverse comparison

, and following discussions with decision-makers. Consequently, the pairwise comparison matrix for evaluating specific criteria is constructed according to the formula presented:

$$A=\left[\begin{array}{ccc}{a}_{11}& \cdots & a_{1n}\\ ⋮& \ddots & ⋮\\ a_{n1}& \cdots & a_{nn}\end{array}\right]=\left[\begin{array}{ccc}{a}_{11}& \cdots & a_{1n}\\ ⋮& \ddots & ⋮\\ {1/a}_{n1}& \cdots & a_{nn}\end{array}\right]$$

To obtain the normalized matrix, the process involves calculating the totals of the elements within every column of the comparison matrix. Subsequently, every element within the comparison matrix is divided by the sum of its corresponding column. The creation of the weight vector follows, determined by averaging the values in each row of the normalized matrix. Furthermore, the priorities matrix is formed through the multiplication of the comparison matrix with the weight vector, executed as follows:

[AWi] = [A][Wi]

The maximum eigenvalue (λmax) is computed with the following:

$$\mathsf{\lambda }\mathrm{m}\mathrm{a}\mathrm{x}=\frac{1}{\mathrm{n}}\sum _{\mathrm{i}=1}^{\mathrm{n}}\frac{\mathrm{A}{\mathrm{W}}_{\mathrm{i}}}{{\mathrm{W}}_{\mathrm{i}}}$$

Within these formulas, “n” stands for the count of criteria, “A” stands for the matrix for pairwise comparisons, and “W” stands for the vector of weights. The method employed to extract the weight vector out of the pairwise comparison matrix is referred to as the principal right eigenvector method (EM) [23]. Given that decision-makers may not always provide completely consistent pairwise evaluations, it is recommended that matrix A for pairwise comparisons demonstrate a satisfactory consistency level. The consistency of the matrix can be evaluated through the consistency ratio (CR), ensuring the reliability of the comparisons [29]:

$$\mathrm{C}\mathrm{I}=\frac{(\mathsf{\lambda }\mathrm{m}\mathrm{a}\mathrm{x}-\mathrm{n})}{(\mathrm{n}-1)}$$

$$\mathrm{C}\mathrm{R}=\frac{\mathrm{C}\mathrm{I}}{\mathrm{R}\mathrm{I}}$$

Here, CI stands for the consistency index, and RI stands for the random inconsistency index, which is determined from

Table 2. RI values depending on criteria numbers (n = 1–15).

n	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
RI	0.00	0.00	0.58	0.90	1.12	1.24	1.32	1.41	1.45	1.49	1.51	1.48	1.56	1.57	1.59

based on the value of “n” obtained from Wang et al. [29].

Should the consistency ratio (CR) be less than 0.10, the comparison matrix is considered to possess adequate consistency. Otherwise, the decision-making process is repeated to achieve the necessary level of consistency. A CR of 0.00 indicates optimal consistency [31].

Table 3. The weightings and ratings of the considered factors.

Factors	Flood Risk Levels	Weight of Effect (RL)	Value	Total Weight (%)
Land Cover	Very High	5	Bare Ground, Built Area	14.5
	High	4	Crops
	Moderate	3	Rangeland
	Low	2	Vegetation, Trees
	Very Low	1	Water
Rainfall (mm)	Very High	5	>22	38
	High	4	11–22
	Moderate	3	7.6–11
	Low	2	6–7.6
	Very Low	1	0–6
Elevation	Very High	5	>143	10.4
	High	4	117–143
	Moderate	3	94–117
	Low	2	46–94
	Very Low	1	−6–46
Slope	Very High	5	0–1.27	37.1
	High	4	1.27–5.6
	Moderate	3	5.6–14.5
	Low	2	14.5–30
	Very Low	1	>30
Sum	100

outlines the assigned weights and ratings for the factors and their respective categories. In the evaluation of these factors, the unique hydrological characteristics of the UAE, which lacks significant rivers or lakes, were taken into account. As a consequence, the selection and verification of parameters in our model were tailored to the region’s conditions. Historical flood event data were analyzed, and expert consultations further substantiated the relevance of the selected factors to validate the weightings. This comprehensive approach was essential in a region where typical hydrological models may not apply, and it provided a foundation for an accurate and context-specific flood risk assessment [32].

The final proportion of each element’s influence on flood risk was determined by dividing the individual weight of each factor by the total combined weight of all factors (that is, the aggregate of the weights for all six factors). Through the use of the weighted linear combination technique, each factor was then multiplied by its respective percentage weight. This process led to the creation of the ultimate flood risk map for the UAE by aggregating all the factors according to the equation provided below:

$$\mathrm{R}=\sum {\mathrm{x}}_{\mathrm{i}}{\mathrm{w}}_{\mathrm{i}}=\mathrm{F}{\mathrm{w}}_{\mathrm{L}}+\mathrm{R}{\mathrm{w}}_{\mathrm{R}}+\mathrm{E}{\mathrm{w}}_{\mathrm{E}}+\mathrm{S}{\mathrm{w}}_{\mathrm{S}}$$

where

• X_i = the map of each parameter i;
• W_i = the weight of each parameter i;
• L = land cover, R = rainfall, E = elevation, and S = slope.

Using geoinformatics and field measurement techniques, the aforementioned thematic maps/factors were determined and digitized. A Digital Elevation Model (DEM) of the study area was created using topographic maps. Using a DEM in a GIS environment, a 3D Analyst tool was used to generate a slope map. The occurrence of floods is inversely proportional to topographic characteristics such as slope and elevation. The slope of land is a critical factor in determining how water flows over the surface. If land is steeply sloped, water will flow downhill quickly and with greater force, leading to more rapid runoff and potential flooding. On the other hand, if land is relatively flat, water will flow more slowly and be absorbed into the soil more easily, reducing the risk of flooding. Elevation refers to the height of the land above sea level. Low-lying areas are more prone to flooding than areas at higher elevations, as water naturally flows downhill to lower elevations. In addition, areas close to bodies of water, such as rivers or lakes, are at a higher risk of flooding due to their proximity to the water source. In addition, the type of land cover in an area can also impact flooding. Natural vegetation, such as forests and wetlands, can absorb and slow down rainwater, reducing the amount of runoff and preventing flooding. However, when natural vegetation is replaced by impervious surfaces like concrete and asphalt, rainwater cannot be absorbed into the ground and instead runs off rapidly into rivers and streams, increasing the likelihood of flooding. Rainfall is one of the primary factors contributing to the risk of flooding. When rainfall occurs, the amount of water on the ground increases, and depending on the intensity and duration of the rainfall, the water may accumulate and cause flooding.

5. Results and Discussion

Figure 4 displays the present flood risk map of the UAE, categorized into five levels—very high, high, moderate, low, and very low. The map accurately represents the flood risk across the country. The map indicates that the northern region of the UAE is at a higher risk of flooding than the southern region, with the majority of the areas in the northern part of the country being classified as either very high or high risk. This is likely due to the fact that the northern part of the UAE receives more rainfall than the southern part, particularly during the winter months. The map also shows that the coastal areas of the UAE, particularly along the Arabian Gulf and the Gulf of Oman, are at a higher risk of flooding than the inland areas. This is due to a combination of factors such as high tides, storm surges, and heavy rainfall, which can lead to flash floods and the inundation of low-lying areas. These areas are characterized by high population density, extensive infrastructure, and large-scale development, which make them particularly vulnerable to flooding events. It is important to note that the majority of the population of the UAE lives in coastal cities, which means that they are particularly vulnerable to the impacts of flooding. In contrast, the inland regions of the UAE are generally at a lower risk of flooding, with most areas classified as being at moderate to very low risk. These regions are largely dominated by desert and semi-arid landscapes, with fewer population centers and less transportation infrastructure development. However, even these regions are not immune to the risks of flooding, particularly in the case of extreme weather events such as cyclones and severe thunderstorms.

Figure 5 and Figure 6 illustrate the flood risk projections for the UAE across two different Shared Socioeconomic Pathways (SSPs) for both near- and far-future scenarios, using the CanESM5 and MIROC6 climate models, respectively. Under SSP245, indicative of a sustainable path with reduced greenhouse gas emissions, the overall flood risk across the UAE is anticipated to be low for both timeframes. Nonetheless, specific locales in the UAE’s northern and eastern parts are expected to exhibit high to very high levels of flood risk in the long-term future.

These areas are predominantly characterized by low-lying coastal regions and extensive urbanization, which increases the surface runoff and reduces the groundwater recharge. The potential impact of these high to very high flood risk levels can lead to severe damage to transportation infrastructure, property, and human life. Conversely, within the SSP585 scenario, indicating elevated greenhouse gas emissions, the flood risk levels exhibit notable disparities compared to SSP245, particularly in the far-future timeframe. The map presented shows a broad distribution of very high and high levels of flood risk across the UAE, with the most significant risks highlighted in the northern and eastern areas of the country. These areas are projected to experience an increase in precipitation intensity, duration, and frequency, resulting in severe flash floods and the inundation of coastal areas. The projected impact of these high flood risk levels can lead to devastating consequences for the UAE’s economy, transportation infrastructure, and human life. Overall, flood risk levels in the near-future period were observed to be comparatively lower than those projected for the far-future period under both the SSP245 and SSP585 scenarios. Nevertheless, the severity and coverage of flood risk levels in the far-future period, particularly under SSP585, were notably greater than in other scenarios.

Table 4. The percentage of UAE areas under different flood risk levels.

Flood Risk Level	Current	CanESM5				MIROC6
		SSP245	SSP245	SSP585	SSP585	SSP245	SSP245	SSP585	SSP585
		Near Future	Far Future	Near Future	Far Future	Near Future	Far Future	Near Future	Far Future
Very Low	34%	25%	13%	13%	23%	13%	25%	13%	13%
Low	19%	17%	29%	29%	15%	29%	17%	29%	29%
Moderate	14%	22%	21%	22%	22%	22%	22%	22%	22%
High	23%	21%	22%	22%	23%	21%	22%	22%	22%
Very High	10%	14%	15%	15%	18%	14%	14%	14%	14%

,

Table 5. The percentage of UAE infrastructure under different flood risk levels.

Flood Risk Level	Current	CanESM5				MIROC6
		SSP245	SSP245	SSP585	SSP585	SSP245	SSP245	SSP585	SSP585
		Near Future	Far Future	Near Future	Far Future	Near Future	Far Future	Near Future	Far Future
Very Low	36%	26%	13%	14%	23%	14%	26%	14%	14%
Low	17%	17%	28%	28%	16%	29%	17%	29%	28%
Moderate	15%	23%	21%	22%	22%	23%	22%	22%	22%
High	20%	18%	19%	19%	20%	18%	19%	19%	19%
Very High	12%	16%	18%	17%	20%	16%	16%	16%	17%

,

Table 6. The percentage of UAE major roads under different flood risk levels.

Flood Risk Level	Current	CanESM5				MIROC6
		SSP245	SSP245	SSP585	SSP585	SSP245	SSP245	SSP585	SSP585
		Near Future	Far Future	Near Future	Far Future	Near Future	Far Future	Near Future	Far Future
Very Low	44%	27%	12%	12%	22%	13%	27%	13%	12%
Low	13%	20%	34%	34%	22%	34%	20%	34%	34%
Moderate	17%	17%	18%	17%	18%	17%	17%	17%	17%
High	9%	20%	20%	20%	20%	20%	21%	20%	21%
Very High	17%	16%	17%	17%	18%	17%	16%	17%	17%

and

Table 7. The percentages of UAE minor roads under different flood risk levels.

Flood Risk Level	Current	CanESM5				MIROC6
		SSP245	SSP245	SSP585	SSP585	SSP245	SSP245	SSP585	SSP585
		Near Future	Far Future	Near Future	Far Future	Near Future	Far Future	Near Future	Far Future
Very Low	37%	26%	13%	14%	23%	14%	26%	14%	14%
Low	17%	17%	29%	29%	16%	29%	17%	29%	29%
Moderate	15%	23%	21%	22%	22%	23%	22%	23%	22%
High	20%	18%	19%	19%	20%	18%	19%	18%	19%
Very High	12%	16%	18%	18%	20%	17%	17%	17%	17%

display the proportion of the UAE’s land, transportation network, and major and minor roads exposed to flood risks under various climate change projections. Currently, 34% of the UAE’s territory is categorized as having a very low level of flood risk. Under the CanESM5 SSP245 scenario for the near future, this figure declines to 25%, and drops further to 13% in the far-future scenario. The MIROC6 SSP245 scenarios for both the near and far future indicate a reduction in areas with very low flood risk to 13% and 25%, respectively. With the CanESM5 SSP585 scenarios for the near and far future, the percentages of areas classified under very low flood risk fall to 13% and 23%, respectively. Under the MIROC6 SSP585 projections for the near and far future, the area of the UAE facing very low flood risk is expected to decrease to 13% and 18%, respectively. The portion of the UAE at “low” flood risk is currently 19%. The CanESM5 SSP245 scenario for the near future predicts a reduction to 17%, but this increases to 29% in the far future. The MIROC6 SSP245 scenarios for both the near and far future forecast a rise to 29%. The CanESM5 SSP585 scenarios predict a fall to 15% in the near future and a rise to 29% in the far future. The MIROC6 SSP585 scenarios anticipate drops to 29% and 22% for the near and far future, respectively. The levels of moderate and high flood risk remain fairly constant across all scenarios, fluctuating from 21% to 23% and from 14% to 22%, respectively. The very high flood risk level sees an escalation to 14% in both the near and far future under the CanESM5 SSP245 and MIROC6 SSP245 scenarios. For the CanESM5 and MIROC6 SSP585 scenarios, very high flood risk escalates to 18% and 14% for the near and far future, respectively.

For the CanESM5 SSP245 near-future scenario, the percentage of infrastructure under very low flood risk decreases to 26%, which is a significant decrease compared to the current scenario. In contrast, the MIROC6 SSP245 near-future scenario shows a slight decrease to 14%. In the far-future scenarios for both models, the percentage of infrastructure under very low flood risk decreases further. The CanESM5 SSP245 far-future scenario predicts that only 13% of infrastructure will be under very low flood risk, while MIROC6 SSP245 far-future scenario shows a decrease to 13%. In terms of the low flood risk level, the CanESM5 SSP245 near-future scenario predicts that the percentage of infrastructure under this level will remain the same as the current scenario, while MIROC6 SSP245 near-future scenario shows an increase to 29%. In the far-future scenarios for both models, the percentage of infrastructure under low flood risk level increases, but the CanESM5 SSP245 far-future scenario shows a more significant increase to 28% compared to the MIROC6 SSP245 far-future scenario, which shows an increase to 29%. Both models predict that the percentage of infrastructure under moderate and high flood risk levels will remain relatively stable in all scenarios, ranging from 21% to 23% and from 14% to 22%, respectively.

However, in terms of the very high flood risk category, both models project an increase in both near- and far-future scenarios. Across the CanESM5 SSP245 scenarios for the near and far future and the MIROC6 SSP245 scenarios for both timeframes, there is a consistent climb to 16% and 18%, respectively. Under the CanESM5 SSP585 scenarios for the near and far future, there is a rise to 17% and 14%, respectively, while in the MIROC6 SSP585 scenarios for both periods, there is an increase to 16% and 17%, respectively.

For major roads, CanESM5 SSP245 and SSP585 predict a decrease in the percentage of roads in the very low flood risk level, while MIROC6 SSP585 predicts an increase. In contrast, MIROC6 SSP245 predicts an increase in the percentage of major roads in the very low flood risk level. For low flood risk levels, all models predict an increase in the percentage of major roads in this category, except for the CanESM5 SSP245 far-future scenario, which predicts a decrease. For moderate flood risk levels, all models predict relatively consistent percentages across current and future scenarios. For high flood risk levels, the CanESM5 SSP245 near- and far-future scenarios predict an increase in percentage, while the other models predict a relatively consistent percentage across current and future scenarios. For very high flood risk levels, all models predict an increase in the percentage of major roads in this category.

For minor roads, the same trend can be observed in terms of an overall increase in flood risk levels predicted by all models. Similar to major roads, CanESM5 SSP245 and SSP585 predict a decrease in the percentage of minor roads in the very low flood risk level, while MIROC6 SSP585 predicts an increase. MIROC6 SSP245 predicts an increase in the percentage of minor roads in the very low flood risk level. For low flood risk levels, all models predict an increase in the percentage of minor roads in this category, except for CanESM5 SSP245 for future scenarios, which predicts a decrease.

For moderate flood risk levels, all models predict relatively consistent percentages across current and future scenarios. For high flood risk levels, CanESM5 SSP245 near- and far-future scenarios predict an increase in percentage, while the other models predict a relatively consistent percentage across current and future scenarios. For very high flood risk levels, all models predict an increase in the percentage of minor roads in this category.

The tables indicate that climate change is poised to elevate flood risk in the UAE. Across all climate scenarios, a notable pattern emerges: a rise in the percentage of areas, infrastructure, and roads facing moderate and high flood risk, coupled with a decline in the percentage of areas, infrastructure, and roads facing very low and low flood risk. Of particular concern is the CanESM5 SSP245 far-future scenario, revealing a substantial uptick in the percentage of areas facing very high flood risk. This emphasizes the critical need to adopt efficient strategies for managing flood risks and implementing adaptation measures in the UAE to mitigate the potential consequences of climate change on the country’s infrastructure and population.

6. Conclusions

In conclusion, the flood risk map of the UAE accurately depicts the country’s flood-prone regions, with the northern and coastal regions being at higher risk. The majority of the population lives in coastal cities, making them particularly vulnerable to flooding events. The future flood risk map projections for both the CanESM5 and MIROC6 models indicate that under the SSP245 scenario, flood risk levels will generally be low across both time periods. Nevertheless, specific areas within the northern and eastern regions of the UAE may continue to face heightened levels of flood risk, ranging from high to very high, particularly in the far-future period, due to extensive urbanization and the presence of low-lying coastal zones. The SSP585 scenario predicts significantly higher levels of flood risk compared to SSP245, with widespread occurrences of very high and high flood risk levels throughout the UAE. The potential consequences of these increased flood risk levels could include substantial damage to transportation infrastructure, property, and human lives. It is expected that the proportion of areas classified as having very low flood risk will decrease in future scenarios, while the percentages of areas with moderate and high flood risk levels will remain relatively stable. Additionally, a significant reduction in the proportion of infrastructure exposed to very low flood risk is anticipated in future scenarios. The recent publication of the CMIP6 models marks a significant advancement, and according to the authors’ knowledge, there have been no studies that have yet explored the application of CMIP6 scenarios. This study highlights the global need for sustainable flood risk management in transportation and planning, advocating for integrated approaches to mitigate flood impacts. It calls for combining sustainable geography with infrastructure development and risk strategies, aiming to address future flooding challenges effectively. Based on the findings of this study, opportunities for further research have been identified, as outlined in the following recommendations:

• Diversify the use of GCMs: Future studies should consider incorporating a wider array of GCMs to capture a broader range of climate scenarios. This approach will help in obtaining more robust and comprehensive results.
• Implement an Ensemble Approach: Adopting an ensemble approach with multiple GCMs can provide a more accurate representation of climate uncertainties. This method allows for the aggregation of different model outputs, leading to more reliable predictions.
• Develop More Advanced Downscaling Techniques: To ensure that global climate model outputs are more applicable at regional or local scales, future research should invest in the development of sophisticated downscaling methods. These techniques should be capable of translating large-scale climate information into finer, more detailed local climate data.