Policy Research Working Paper 11092 Climate Risk and Poverty in the Middle East and North Africa Chitra Balasubramanian Sandra Baquie Alan Fuchs Poverty and Equity Global Department March 2025 Policy Research Working Paper 11092 Abstract The Middle East and North Africa faces significant climate surveys to provide high-resolution assessments of the expo- challenges, such as increasing temperatures, heightened sure and vulnerability of the region’s population and poor flood risks, frequent droughts, and growing air pollution people to four types of climate shocks. With the data avail- issues. These challenges are compounded by the large pro- able, the paper estimates that almost the entirety of the portion of the population living below the poverty line extreme poor population is exposed to at least one climate in some countries in the region. Indeed, people living in shock. The region hosts climate-poverty hot spots in the poverty are more exposed to poor air quality and natural Republic of Yemen and Morocco, where adaptation to cli- disasters as they disproportionately tend to live in haz- mate change will be crucial to end poverty. The resulting ard-prone areas. They are also more vulnerable as they high-resolution estimates of exposure and vulnerability can may have scarcer resources to cope with shocks. This paper inform the targeting of climate adaptation measures. combines remote sensing, geospatial data, and household This paper is a product of the Poverty and Equity Global Department. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at afuchs@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Climate Risk and Poverty in the Middle East and North Africa Chitra Balasubramanian, Sandra Baquie and Alan Fuchs 0Disclaimer: This work is a product of the staff of The World Bank and the International Monetary Fund (“© 2025 The World Bank and International Monetary Fund”). The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Acknowledgments: We sincerely thank the Poverty and Equity team in the Middle East and North Africa region, the Global Poverty team, and the Global Facility for Disaster Reduction and Recovery (GF- DRR) for their invaluable support and insights throughout this work. We are also grateful to colleagues who peer- reviewed this work, Ernest John Sergenti (Senior Economist, MNACE), Roberta V. Gatti (Chief Economist, MNACE), Volkan Cetinkaya (Senior Economist, HMNHN), Ugo Gentilini (Lead Economist, HMNSP), Samuel Freije-Rodriguez (Lead Economist, ELCPV) and Sailesh Tiwari (Lead Economist, EEAPV) for their thoughtful comments and feedback. This work has been prepared under the direction and guidance of Salman Zaidi (Practice Manager, EMNPV) and Nadir Mohammed (Regional Director, Prosperity MENA). This paper and the underlying analysis have benefitted from support of the Climate Support Facility Whole of- Economy trust fund for the Middle East and North Africa region of the World Bank. 1 1 Introduction Climate change’s impacts disproportionately threaten the poor, as they have fewer resources to avoid or cope with the adverse consequences of weather shocks, such as diseases, water scarcity, food insecurity, and detrimental health outcomes due to air pollution (Hallegatte et al., 2017). This compound threat of poverty and climate risk is significant in the Middle East and North Africa (MENA) region. Indeed, the region is one of the most vulnerable to the adverse effects of climate change and 40 percent of its population (116.1 million individuals) live in multidimensional poverty (Economic, for Western Asia (ESCWA), et al., 2023). As it increases the frequency and severity of adverse weather events, climate change will exacerbate this threat. Therefore, understanding the intersection of poverty and climate risk is critical to improve climate adaptation of the poor and vulnerable people in the MENA region. On the one hand, the MENA region is highly exposed to weather shocks, amplified by climate change. The region is characterized predominantly by arid and semi-arid zones, facing extreme heat and drought, with recorded temperatures reaching up to 56°C and projected reductions in rainfall of up to 50 percent along the Mediterranean coastline. Heat stress has significant implications for the financial sector (Acharya et al., 2022), though impacts vary across industries, with services showing limited productivity decline. In the MENA region, rising Wet Bulb Global Temperature (WBGT) levels represent a critical climate change-related health threat (Zittis et al., 2022), emphasizing the importance of adaptation strategies (Heal & Park, 2013) and early-warning systems using the 32°C threshold to prevent catastrophic outcomes. The MENA region has the lowest rate of water availability per capita globally, and this scarcity is anticipated to worsen with climate change and high demographic growth. Pro- jections indicate that by 2050, the MENA region may need to import as much as 50 percent of its water requirements to adequately support its population. However, in some places, rainfall intensity may be high over a small number of days, leading to floods. For instance, the floods in Libya in 2023 affected more than 800,000 lives. Deteriorating climate conditions are leading to reduced agricultural yields and heightened risks to natural habitats, further intensifying existing environmental strains. Finally, the region’s heavy reliance on fossil fuels is a major factor in its poor air quality, which exacerbates environmental and public health challenges. On the other hand, approximately 40 percent of the MENA population lives in multi- dimensional poverty. A recent poverty assessment in the Republic of Yemen shows that about 49 percent of the population lives below the national poverty line due to ten years of conflict and crisis, which have had dire effects on living conditions (World Bank, 2024b). In Lebanon, monetary poverty has more than tripled. The prevalence of climate shocks and poverty in the MENA region poses the risk of compound shock. Indeed, poor people tend to live in places where weather shocks occur more frequently and have fewer resources to cope with them (World Bank, 2024a). This analysis contributes to the literature in three ways. First, despite the recognized threat of compound shock between poverty and climate shocks, a critical gap exists in the assessment of the number of poor and vulnerable people exposed to climatic shocks in the MENA region. This study addresses it by providing a high-resolution assessment of the 2 populations exposed to various climate-related hazards, particularly focusing on the poor and vulnerable, using the IPCC Framework (Adler et al., 2022). First, it quantifies exposure to four types of extreme hazards, extreme heat, floods, droughts, and extreme air pollution,1 on a 10x10 km grid. Second, it locates and estimates the number of people living in poverty (using income thresholds of $2.15, $3.65, and $6.85 per day) and exposed to these hazards. Third, it assesses their vulnerability, i.e., their capacity to cope with or recover from shocks. Doing so addresses the following research questions: 1. Hazards: Where are extreme hazard events most likely to occur? 2. Exposure: Are populations, particularly the poor, disproportionately exposed to these hazards? 3. Vulnerability: What are the factors contributing to physical susceptibility, and how capable are these populations of coping with or recovering from shocks? Second, the current poverty estimates for the MENA region are often outdated or lack local representation. The analysis introduces a high-resolution poverty estimation methodology that combines publicly available data with household estimates to address this issue. Building on approaches by Doan et al. (2023), this methodology refines poverty estimation beyond what conventional surveys allow, providing more precise, locally representative data. Specifically, the Relative Wealth Index (Chi et al., 2022), available at a high resolution of 2km-by-2km, is used as a weighting factor adjusted to match average poverty shares recorded by the Global Subnational Atlas of Poverty (GSAP) database, which are typically at a subnational level. This adjustment yields a more detailed measure of poverty at the local level. The study then estimates the number of poor individuals exposed to each hazard by multiplying this poverty share by the total population exposed to the respective hazards. Third, we propose an approach that broadens the estimation of vulnerability in the MENA region. The existing literature often defines vulnerability as the likelihood of severe loss due to a shock and the inability to cope or recover afterward. For example, Doan et al. (2023) include indicators such as access to clean water, electricity, and financial services to gauge vulnerability. Our approach broadens this concept by categorizing vulnerability into three dimensions: Economic Activity, Connectivity, and the Social and Natural Environment. In the Economic Activity dimension, changes in Nighttime Lights (as a proxy for GDP per capita) and yield gaps (developed by FAO) are used to signal economic instability and rural poverty. The Connectivity dimension assesses access to markets, healthcare, and education services. Finally, the Social and Natural Environment dimension examines access to financial services and social protection, forest cover loss and conflict events over the past decade to understand how instability limits populations’ ability to cope with or recover from climate shocks. The subsequent sections of the paper are organized as follows: Section 2 outlines the data sources and methods to measure hazard risk, exposure and vulnerability; Section 3 presents the results; and Section 4 discusses key findings and future directions for research. 1In the context of this study, the term ”hazard” refers to natural or human-induced physical events that have the potential to negatively impact vulnerable and exposed elements. We distinguish ”hazard” from ”risk,” with hazard being a component of risk rather than synonymous with it. 3 2 Measuring Hazard Risk, Exposure, and Vulnerability The disaster risk assessment in this study begins with estimating the extent of each hazard. Using the hazard data summarized in Table 1 of the Appendix, two key decisions are made: (a) selecting the appropriate data source and aggregation method, and (b) determining the threshold for defining an “extreme” event for each hazard. Various sources of satellite imagery were used to assess heat stress, drought risk, flood risk, and air pollution. The literature employs diverse indicators for hazard assessment, including metrics such as degree days, the number of rainy or dry days, and maximum average temperature to quantify aspects of heat stress and drought risk. A review of the existing research highlights that impact functions often demonstrate the non-linear effects of heat on various outcomes, as documented by studies such as Guo et al. (2014), Burgess et al. (2017), Hsiang and Narita (2012), Roberts and Schlenker (2013), Hidalgo et al. (2010), Graff Zivin and Neidell (2014), Auffhammer and Schlenker (2014), Burke et al. (2015), Deryugina and Hsiang (2014), Baylis (n.d.), S. M. Hsiang and Meng (2015) and Rentschler et al. (2022). This variability implies that the choice of data source and the definition of thresholds for identifying hazards are context dependent. Thus, the thresholds employed in the Middle East and North Africa region differ from ones applied in South Asia or any other region. Moreover, prior studies suggest that extreme climate and weather events, despite being rare, are often critical in explaining disaster-related losses. Consequently, this study focuses on assessing disaster risk by examining the extent and severity of extreme hazard events and their impacts on populations. The analysis estimates the number of people exposed to extreme events, including heat stress, flooding, severe drought, and air pollution. The thresholds for each extreme event are selected based on the literature to ensure consistency and accuracy in identifying the most impactful events: 1. Extreme heat: Wet Bulb Globe Temperature (WBGT) above 32°C 2. Flooding: Water depth of 50 cm or more 3. Drought: Areas where 35% of arable land is affected more than 20% of the time 4. Air pollution: PM2.5 levels exceeding 15 µg/m³ Binary hazard layers are created for each hazard using the selected thresholds. Grid cells exceeding the threshold are assigned a value of ”1,” while those below the threshold receive a value of “0.” The analysis draws upon data from two categories. First, remote sensing data, crucial for assessing natural hazards and some of the vulnerability dimensions. Second, geospatial data from non-profit platforms, such as poverty and wealth estimates from the Global Subnational Atlas of Poverty (GSAP) (World Bank, 2023) and Data for Good (Chi et al., 2022), and conflict data from the Armed Conflict Location and Event Data Project (ACLED) (Raleigh et al., 2023). We also use data on roads, schools, and hospitals, sourced from the open data platform OpenStreetMap. These data sources are all publicly accessible, facilitating replication and further analysis. 4 Table 1 in the Appendix provides a comprehensive summary of the various data sources utilized in the analysis. This table methodically lists and explains each source according to the three dimensions of disaster risk: hazards, exposure, and vulnerability. 2.1 Measuring Hazard Risk 2.1.1 Extreme Heat The study employs a Wet Bulb Globe Temperature (WBGT) threshold of 32°C as a crucial benchmark for assessing heat stress risks in outdoor environments, particularly relevant in the MENA region as seen in Figure 12. This comprehensive measure accounts for multiple environmental factors and helps identify high-risk conditions before they reach fatal levels, which can occur beyond 35°C WBGT. When temperatures approach this threshold, there is a significant increase in heat-related hospitalizations and deaths. This impact is particularly severe for vulnerable populations, such as outdoor workers and elderly individuals, who can experience dangerous health effects even at WBGT levels below 33°C. Research indicates that extreme heat could have substantial economic ramifications by the century’s end, potentially impacting GDP growth (S. Hsiang et al., 2017). These ef- fects manifest through increased energy demands, reduced electricity production efficiency, and decreased labor productivity, particularly in high-risk industries like construction and mining. The situation is further complicated by heightened wildfire risks, which can lead to property damage and financial instability (Jones et al., 2022). 2.1.2 Flood Risk The study employs a 50cm water depth threshold for assessing flood exposure, based on recent global flood risk studies that consider pluvial, fluvial, and coastal flooding (Doan et al., 2023); (Rentschler et al., 2022). While focusing solely on fluvial flooding might underestimate total flood impact, this analysis uniquely considers expected flooding across all return periods, providing a comprehensive view of the probabilistic flood risk curve rather than relying on single-period assessments. The research utilizes Fathom-Global 2.0 dataset, which offers global coverage at 3 arc- second resolution (approximately 90 x 90 meters at the equator) and simulates flood events for return periods ranging from 5 to 500 years. This high-resolution dataset, based on the latest terrain and hydrographic models, enables detailed flood extent and depth assessments across all countries, with comprehensive information available in the Appendix. While there is no universal standard for flood depth in the literature, and even 0.15- meter depths can cause significant disruptions (Rentschler et al., 2022). The study adopts a 50cm threshold to maintain consistency and enable meaningful comparisons. This choice of threshold significantly influences the number of people classified as “exposed” to flooding risk. 2.1.3 Severe Drought Risk The study employs the Agricultural Stress Index (ASI) (FAO, 2023), developed by the FAO in 2023, to assess drought risk in cropped areas. This index evaluates the percentage of arable land affected by drought within administrative areas, utilizing historical drought data 5 from 1984 to 2022. The methodology involves assessing dry period intensity and duration during crop cycles at 1-kilometer resolution, and calculating the percentage of grid cells in arable areas where the Vegetation Health Index (VHI) falls below 35 percent. The analysis process involved aggregating the initial 1-kilometer resolution data to 10- kilometer grid cells, focusing exclusively on cropland regions. An area is classified as having high drought risk when more than 20 percent of its arable land experiences severe drought conditions according to the ASI measurements. The MENA region’s vulnerability is particularly noteworthy, as it contains just 2 percent of global renewable water resources and includes 12 of the world’s most water-scarce coun- tries. With average water availability of 1,200 cubic meters per person annually—six times below the global average—the region faces intensifying challenges due to rising temperatures and projected 20 percent decrease in precipitation. 2.1.4 Air Pollution To evaluate air quality in the MENA region, this study utilizes annual global satellite-based estimates of fine particulate matter (PM2.5), as provided by van Donkelaar et al. (2021). The dataset, updated for 2023, provides surface-level PM2.5 concentrations at a 1.1 km res- olution, covering the period from 1998 to 2023. Further details on this dataset can be found in the Appendix. Given that PM. concentrations in MENA countries significantly exceed the WHO annual standard of 5 µg/m³—often due to regional dust storms and naturally high ambient dust levels—a less stringent daily threshold of 15 µg/m³ offers a more practical benchmark for analysis. First, it aligns better with commonly recognized daily standards (including the WHO’s recommended 24-hour guideline) and thus captures meaningful short-term fluctu- ations that the stricter annual target might mask. Second, it helps distinguish areas with extremely high pollution levels from those that are merely above average, providing clearer variation in maps and statistical analyses. Third, this more moderate threshold acknowledges the natural dust contributions prevalent in the region, reducing the bias against locations that experience persistent desert influences. Finally, adopting a 15 µg/m³ standard enables researchers and policy makers to identify and address anthropogenic sources of pollution more effectively, rather than conflating them with the baseline dust levels. Overall, this approach yields a more nuanced and actionable understanding of PM pollution across the MENA region. 2.2 Measuring Exposure 2.2.1 Total Population Exposed The analysis involves overlaying the binary hazard layers to estimate the number of people exposed to each hazard at the 10km-by-10km grid level. Population data was sourced from WorldPop (WorldPop & Center for International Earth Science Information Network, 2018), available at a 1km resolution and aggregated to a 10km-by-10km resolution for each country in the MENA region. Figure 1 illustrates the population density (per km2) across the 6 Figure 1: Population Density per Square Kilometer in the MENA Region for the Year 2020, WorldPop region for the year 2020. It shows that the region has major population centers within each country. It is also important to note that population estimates can vary significantly across different datasets, such as WorldPop, Global Human Settlement Layer, and official country estimates, we recommend cross verifying these population figures with official data at the lowest available administrative level. 2.2.2 Number of Poor Exposed The analysis calculates the number of people in poverty who are exposed to hazards by estimating poverty rates in each 10km-by-10km grid cell using the relative wealth index where available. This index indicates the proportion of the population that is ‘relatively wealthier’ compared to the rest of their country. The national or regional poverty rates are disaggregated by inferring the ‘relative poverty’ of each 10km-by-10km grid cell compared to the national or regional level from the relative wealth index. The Appendix provides detailed methodology for inferring poverty rates at the grid cell level. Figures 2 and 3 illustrate the distribution of poverty at the 10km-by-10km grid level, based on the $2.15/day poverty threshold. Regions with the highest concentrations of poverty are highlighted in red and purple hues, while countries without data are shown in white. The results reveal significant disparities across the region, with Yemen having the highest number of individuals below the $2.15/day threshold at 19,908,012 people, while Malta has the lowest count among countries with available data, with fewer than 2,000 individuals below the threshold. The spatial distribution of poverty is then overlaid with hazard risk maps to estimate the number of poor people exposed to each hazard. This overlay analysis enables the iden- 7 Figure 2: Share of Poor (per 100 km2) at the $2.15/day Figure 3: Number of Poor (per 100 km2) at the $2.15/day tification of areas where high poverty rates intersect with significant hazard risks, providing crucial information for targeting interventions and resource allocation. 2.3 Measuring Vulnerability The analysis evaluates vulnerability across three dimensions (Economic activity, Connectiv- ity, and Social and Natural Environment) at a 10km-by-10km grid resolution, incorporating variables such as nighttime lights and yield gaps for economic activity, travel time to ur- ban centers and essential facilities for connectivity, access to social protection and financial services, forest cover loss and conflict events for social and environmental factors. The vulnerability rate for each grid cell is calculated as the proportion of vulnerability variables exceeding specified thresholds relative to the available dimensions, with Figure 4 highlighting areas where vulnerability rates surpass 40%, notably in regions of Morocco, Algeria, the Arab Republic of Egypt, 8 Figure 4: Share of Vulnerability Saudi Arabia, and parts of the Levant. This 75th percentile threshold of the vulnerability rate distribution serves as a regional benchmark for identifying the ”most vulnerable” areas, though local conditions may warrant different thresholds across countries. 2.3.1 Aggregate Yield Achievement Ratio This analysis evaluates populations at risk of food insecurity by examining yield gaps be- tween potential and actual crop yields as a proxy for food scarcity or rural poverty. Using FAO’s Global Agro-Ecological Zones (GAEZ version 4) portal, the study analyzes yield gaps at approximately 10-kilometer resolution for 22 crops under both rain-fed and irrigated con- ditions. A yield achievement ratio (actual yield/potential yield) below 5 in a 10 km-by-10 km grid area indicates vulnerability to food insecurity or rural poverty, with this threshold determined based on data distribution patterns. Several countries in the MENA region demonstrate significant vulnerability, particularly the Islamic Republic of Iran, Algeria, Morocco, and Yemen. The Islamic Republic of Iran’s Khorasan region shows the highest concentration of vulnerable grid cells, corroborating findings from an NIH study that indicated 42% of the Islamic Republic of Iran’s population faces food insecurity, especially in Khorasan (Omidvar et al., 2013). Yemen’s Hadhramaut region and Morocco’s Sud region, affected by water scarcity and droughts, also display notably low yield ratios, indicating severe food insecurity challenges. 9 Figure 5: Agricultural Yield Achievement Ratio, FAO 2.3.2 Change in Nighttime Lights between 2017 and 2019 NASA’s Black Marble Nighttime Lights Suite (Rom´ a n et al., 2018) provides high-resolution measurements of nighttime lighting, serving as a proxy for human activity and economic conditions in regions with limited economic data. After excluding gas flares from the oil industry, the data is aggregated into 10 km-by-10 km grids for analysis over the period 2017-2021. Regions showing declining nighttime lighting during this period are identified as potentially vulnerable, as decreased lighting may indicate economic or social challenges that could increase susceptibility to climate-related risks. Several sub-regions within MENA demonstrated significant vulnerability, particularly in the Islamic Republic of Iran and Saudi Arabia, where notable reductions in nighttime lighting occurred between 2017 and 2021.2 In the Islamic Republic of Iran, affected regions include Khorasan, Khuzestan, Fars, Sistan-o-Baluchistan, Kerman, and East Azerbaijan, while in Saudi Arabia, the provinces of Riyadh, Makkah, Eastern, and Madinah showed substantial decreases in lighting intensity as this coincided with COVID-19. In most urban areas, physical distancing policies caused a significant abrupt reduction in NTL and associated urban activities. For instance, the timing of lockdowns in late March of 2020 corresponded to decreases in nighttime radiance of at least 5% in a large majority of urban areas in Saudi Arabia (82%) (Stokes & Rom´ a n, 2022). These findings align with recent studies that have similarly used nighttime lighting data to identify areas lacking illumination, particularly noting reduced light intensity in specific regions of the Islamic Republic of Iran. 2Gas flaring data are from the World Bank repository. 10 Figure 6: Change in Nighttime Lights (2017 – 2021), NASA BlackMarble 2.3.3 Travel Time by Car to Urban Centers, Health Centers, and Educational Institutions The analysis examines access to three critical services: educational facilities, healthcare providers, and marketplaces. Access to education impacts long-term economic mobility and resilience, while healthcare access is crucial for managing both chronic conditions and emergency responses during extreme events. Market access affects economic opportunities and food security, with limited access often correlating to higher costs for basic necessities and reduced economic participation. The combination of these access metrics provides a comprehensive view of service accessibility challenges that can compound vulnerability to climate hazards. The analysis measures accessibility through travel time calculations using OpenStreetMap data, with a one-hour threshold to the nearest educational facility, healthcare provider, or marketplace serving as a vulnerability indicator for each 10 km-by-10 km grid cell. This methodology reveals significant accessibility challenges in regions such as Tamanrasset, East- ern Province in Saudi Arabia, Adrar in Algeria, and Al Wadi Al Jadid in Egypt, where travel times frequently exceed the one-hour threshold, indicating areas of high vulnerability due to limited access to basic services. 2.3.4 Conflict Events within the Past Five Years Climate change particularly impacts refugees and conflict-affected populations, with approx- imately 90% of refugees originating from nations already struggling with climate impacts or lacking adequate adaptation capacity. The UNHCR has highlighted the compound vulner- ability of those fleeing conflict who must simultaneously cope with climate impacts, while conflict-affected regions face intensified challenges including food insecurity, rapid urbaniza- tion, and competition for scarce resources, particularly water (S. M. Hsiang et al., 2013). Analysis of Armed Conflict Location and Event Data Project (ACLED) (Raleigh et al., 11 Figure 7: Travel time to nearest educational facility by car (10km – by- 10km) Figure 8: Travel time to nearest health facility by car (10km-by-10km) 12 Figure 9: Travel time to the nearest market (airport, seaport or city center) by car (10km- by-10km) 2023) data from 2019 to 2023 identifies Yemen, the Syrian Arab Republic, Iraq, and Islamic Republic of Iran as the most conflict-affected countries, with Aleppo and Hasaakeh in Syria showing particular vulnerability. Specific conflict hotspots include Yemen’s Taizz, Marib, and Al Jawf regions, Iraq’s Ninevah and Arbil regions, and Islamic Republic of Iran’s Sistan-o-Baluchistan area. Grid cells recording any conflict incidents in the past five years are classified as vulnerable, with the spatial distribution of these incidents over the past decade illustrated at 10 km resolution. Figure 10: Total Conflict Events in the Last 10 years, ACLED 13 2.3.5 Change in Forest Cover between 2000 and 2022 The study examines forest loss from 2000 to 2022 using data from the Global Land Analysis and Discovery (GLAD) laboratory (Hansen et al., 2013) at the University of Maryland and Global Forest Watch (GFW), aggregating 30-meter resolution data to 10-kilometer grids for consistent vulnerability assessment. The analysis reveals significant forest cover changes across the MENA region, with substantial losses highlighted in dark brown areas according to the time-series imagery from Landsat satellites. The most severe forest loss is concentrated in specific regions of Morocco (Northwest and Centre North), the Islamic Republic of Iran (Mazandaran, Gilan, Khorasan, East Azerbaijan, and Golestan), Algeria (Tizi Ouzou, Tlemcen, Skikda, and Bejaia), and Syria (Aleppo). This widespread forest loss amplifies regional vulnerability by reducing biodiversity, disrupting water regulation, and increasing soil erosion susceptibility, thereby intensifying the environmental challenges faced by communities dependent on these natural resources. Figure 11: Forest Cover Loss During 2000 – 2022, Global Land Analysis and Discovery (GLAD) 2.3.6 Access to Financial Services and Social Protection Doan et al. (2023) utilized the Atlas for Social Protection Indicators of Resilience and Equity (ASPIRE) and the Global Financial Inclusion database (Global FINDEX) to analyze social protection and financial services globally. This analysis seeks to replicate their findings specifically within the MENA region. ASPIRE provides insights into the proportion of individuals in each income quintile covered by social assistance, while Global FINDEX, drawing on data from the Gallup World Poll, measures the prevalence of financial account ownership among individuals. In the MENA region, the 2021 Global FINDEX survey covers 11 countries, excluding Djibouti and Yemen, and provides data crucial for calculating indicators of access to essen- tial services and poverty. Using this data, a welfare indicator for 2019 is constructed, which 14 is then merged with ASPIRE data based on the 2019 welfare estimates. From this combined dataset, the average number of household deprivations, including lack of electricity, water, and infrastructure, and access to a financial account, is calculated. Details on the conver- sion of the household data into aggregate region-level estimates are provided in the Appendix. 3 Results 3.1 Saudi Arabia Shows Stark Temperature Contrasts (21-41.5°C WBGT), Yemen and Egypt Face Critical Flood Risks, and Air Pollution Exceeds WHO Limits across All Areas The Dammam region in the Eastern Province of Saudi Arabia records the highest Wet-Bulb Globe Temperature (WBGT) in the region, exceeding 41.5°C. It is important to note that WBGT is an extreme measure of heat stress, different from standard temperature readings. This province, the largest by area in Saudi Arabia and the third most populous after Riyadh and Mecca, has a population of approximately 4.9 million. Over a third of this population resides in the Dammam metropolitan area, which has an estimated population of 1.53 million. This region, highly popular among tourists, is strategically located near key ports and serves as a major hub for the kingdom’s oil production and exports. Saudi Arabia experiences both the highest and one of the lowest Wet-Bulb Globe Tem- peratures (WBGT) in the region when comparing the 10km grid cells with each other. While the Eastern Province records the highest WBGT temperature, Jizan, the capital of the Jizan region in the southwest corner of the country, records one of the lowest, averaging 21°C. Jizan, with an estimated population of 1.4 million, is strategically positioned on the Red Sea coast. It serves as the central hub for a significant agricultural area that supports approximately 170,000 people. In the MENA region, the areas with the highest flood risk, based on flood extents and water depths without considering population, include Abd Al Kurwi Island and Al Hudayah in Yemen, as well as parts of Al Bahr Al Ahmar (Red Sea Governorate) in Egypt. Abd Al Kurwi Island has fewer than 100 inhabitants, while Al Hudayah is Yemen’s fourth-largest city and principal port on the Red Sea, with an estimated population of 735,000 in 2023. This port is crucial for Yemen’s humanitarian aid efforts and handles approximately 70 percent of its commercial imports. The analysis also indicates that Al Bahr Al Ahmar, located between the Nile and the Red Sea in southeastern Egypt, is among the top five areas in the region for high flood risk. This governorate has experienced significant population growth in recent decades and is a popular tourist destination. It also has a large offshore fishing industry and is rich in natural resources, including phosphates. Ras Gharib, a municipal division within this region, produces 70% of Egypt’s oil. Human-driven desertification and climate change are intensifying air pollution across the MENA region through increased atmospheric dust and sand, air stagnation, and elevated temperatures. While coastal areas like Morocco’s shoreline show lower pollution levels due to wind patterns that disperse pollutants, even these areas exceed the WHO’s safe limit of 5 µg/m³ for PM2.5 concentrations. 15 Figure 12: Heat Stress using Wet Bulb Global Temperature (WBGT) Figure 13: Average Projected Water Depth from FATHOM Flood Data (2019) Across All Return Periods, WB Calculations 16 Figure 14: Annual PM2.5 Concentration Estimates by Van Donkelaar et al. (2021) The health and economic toll is severe, with air pollution reducing life expectancy by an average of 4.1 years in the most affected areas and causing GDP losses of up to 5.5% across the region. The impact is particularly stark when considering that exposure to 22 µg/m³ of PM2.5—common in many MENA locations—equals smoking one cigarette daily, highlighting the urgent need to address emissions from vehicles, waste burning, and industrial activities. 3.2 MENA Faces a Triple Environmental Challenge of Widespread Air Pollution, Extreme Heat Stress, and Severe Drought The Middle East and North Africa (MENA) region faces a significant triple environmental challenge: widespread air pollution, extreme heat stress, and severe drought. Although the arid and semi-arid conditions of MENA generally limit the occurrence of widespread flooding, the frequency of flash floods has dramatically increased in recent years, underscoring an emerging risk for some areas. Among the Gulf countries, heat stress is particularly severe, with these nations experi- encing some of the highest temperatures in the region. Severe drought conditions are most prominent in Syria, Iraq, Islamic Republic of Iran, and Morocco, which are identified as the four countries most affected by prolonged periods of drought. Air pollution is another pervasive issue, with almost the entirety of the MENA region experiencing high levels of particulate matter, presenting a major concern for public health. Notably, Iraq and Egypt are among the countries most affected by flooding. Figure 15 illustrates population exposure to the four key environmental hazards: extreme heat stress, severe drought, high levels of air pollution, and flood risk. Air pollution, specifi- cally, has emerged as one of the most critical environmental challenges facing the region, with most countries exceeding the World Health Organization (WHO) guideline of 15 µg/m³ for PM2.5. Additionally, the region is heavily impacted by drought, particularly in rural areas, coupled with high heat exposure across approximately six countries. 17 Figure 15: The Triple Environmental Challenge in MENA: Widespread Air Pollution Expo- sure, Extreme Heat Stress and Severe Drought Impact in Rural Areas 18 3.3 MENA Observes High Levels of PM2.5 Largely from Anthropogenic Sources In the Middle East and North Africa (MENA) region, desert dust is often assumed to be the primary contributor to air pollution. However, recent studies by Heger et al. (2022); Osipov et al. (2022) indicate that a substantial portion of the pollution originates from anthropogenic sources. The health burden associated with air pollution in this region, as measured by excess mortality attributable to PM2.5 exposure, ranges from 6 to 16 percent, whereas in countries such as Germany and the United States, this figure is considerably lower, at 3 to 4 percent. Our analysis reveals significant concentrations of PM2.5 across the MENA region, con- tributing to a major health burden, with estimated economic losses equivalent to 5.5 percent of the region’s combined GDP in 2019. Given the elevated levels of pollution observed, we utilized the World Health Organization’s (WHO) daily PM2.5 threshold of 15 µg/m³ rather than their annual threshold of 5 µg/m³. Among the MENA countries, Tehran Province in Islamic Republic of Iran has the highest number of individuals exposed to elevated air pollution levels, followed by Riyadh in Saudi Arabia. In provinces such as Qatar, Djibouti, and Lebanon, more than 40 to 50 percent of the total population is exposed to high levels of PM2.5, indicating significant exposure across these areas. These sub-regions share common characteristics: they are urban and industrial centers with a high concentration of vehicular emissions, as well as emissions from factories and other industrial activities. In contrast, coastal regions in parts of Morocco, Algeria, Egypt, Libya, and Yemen exhibit the lowest levels of PM2.5, as indicated by the white areas on the map in Figure 16. This observation can be attributed to several factors, including the physical properties of aerosol particles that may influence measurement accuracy, and the effects of coastal breezes, which can result in reduced concentrations of PM2.5, despite the presence of activities such as port operations or offshore oil extraction. Further research is necessary to delineate the contributions of different pollution sources and to understand the temporal patterns that emerge from various emissions across the region. 19 Figure 16: Population exposed to air pollution (PM2.5) in 2020; WB calculations 3.4 Populations Exposed to Both Heat and Air Pollution across the Region Are in Major Urban Centers The Makkah Province in Saudi Arabia records the highest number of individuals exposed to both extreme heat stress and air pollution, comprising approximately 20 percent of the total population. Major urban centers within this province include Jeddah, Mecca, and Ta’if. Additionally, Makkah attracts approximately 2 million pilgrims annually during the Hajj. Jeddah serves as an economic hub, with a major port and significant oil facilities contributing to pollution levels. Following Makkah, the Giza District in Egypt, the third-largest city after Cairo and Alexandria, also experiences high exposure levels. The combination of rapid urbanization and a semi-arid climate result in approximately 8 percent of Egypt’s total population being exposed to both extreme heat stress and elevated PM2.5 concentrations. Other administrative regions, such as Giza (Egypt), Baghdad (Iraq), Khuzestan (Islamic Republic of Iran), and Dubai (United Arab Emirates), are major urban centers with significant industrial activities, resulting in roughly 20 to 50 percent of their populations being exposed to both extreme heat and high levels of air pollution. Urban centers such as Al Rayyan (Qatar), Dubai, Al Farwaniyah (Kuwait), Djibouti, and Manamah (Bahrain) also exhibit high exposure, with between 20 and 55 percent of their populations exposed to combined heat stress and air pollution. Aleppo in Syria has approximately 7 percent of its total population severely exposed to both extreme heat stress and air pollution as of 2020. Notably, Ahvaz in Khuzestan Province, Islamic Republic of Iran, which accounts for around 7 percent of the exposed population, was declared the world’s most polluted city in 2011. 20 3.5 Significant Portions of Arable Land across Various Sub-Regions Are at High Risk Due to Concurrent Drought and Air Pollution, with Potential Severe Impacts on Agriculture and Local Economies In Islamic Republic of Iran, Tehran Province, which constitutes 11 percent of the country’s population, is highly exposed to both drought and air pollution. Tehran is the Islamic Republic of Iran’s most industrialized province, with 86.5 percent of its population residing in urban areas and the remainder in rural areas. Recent reports indicate that rainfall in the province has declined by 67 percent, and the available water volume in reservoirs has fallen to less than 300 million cubic meters. The persistent drought in Tehran, and more broadly in the Islamic Republic of Iran, has led to the loss of livelihoods for more than a quarter of farmers in recent years. Following the Islamic Republic of Iran, the Giza region in Egypt, which comprises 8 percent of the national population, is also highly exposed to both drought and air pollution. The Nile Delta valley, which is critical for agriculture and contributes 15-20 percent to the GDP while employing one-third of the workforce, faces significant threats from these environmental stressors, exacerbated by water scarcity for farming. Similarly, in northwest Syria, the Jebel Saman district in Aleppo, where mixed-method farming is prevalent, small- to medium-scale farmers are particularly vulnerable to these risks. As a share of a country’s total population, Amman in Jordan, which constitutes 22 percent, is highly vulnerable to both drought and air pollution. Figure 17 presents a matrix illustrating the various climatic shocks affecting each country and the share of the population impacted by these shocks. Notably, Iraq is the most affected by the three concurrent shocks of air pollution, drought, and heat. 21 Figure 17: Overlapping Extreme Hazards in MENA: Prevalence of Widespread Air Pollution and Concurrent High Drought Stress or Extreme Heat Stress in Most Countries 3.6 Djibouti, Morocco, and Egypt present the Highest Proportions of Poor Populations Exposed to Heat Stress, Drought, Air Pollution, and Flooding, Respectively, among Countries Where Poverty Data Is Available at the 10 km Resolution Figure 18 illustrates the proportion of poor populations—defined by daily income thresholds of $2.85, $3.65, and $6.85—exposed to these hazards at the country level. In Djibouti, 41% of the total population is classified as extremely poor (below $2.85) and exposed to heat stress. Additionally, 25% of the population is poor at the $3.65 threshold and exposed to heat stress, while 13% falls under the $6.85 threshold with exposure to heat stress. For drought exposure, 18% of Morocco’s population is extremely poor and exposed to severe drought conditions, while 3.7% is poor at the $3.65 level and exposed to drought, and 0.52% is poor at the $6.85 level with drought exposure. High levels of air pollution extremely poor population in Djibouti (52%), Egypt (50%), and Morocco (30%). Figure 19 displays these findings in absolute terms, illustrating the highest number of individuals who are poor at the $6.85 threshold and exposed to various hazards in Egypt. Egypt has the largest population of poor individuals exposed to extreme heat, with many residing in Lower Egypt. 22 Figure 18: Djibouti has the highest share of total population that are exposed to heat stress and poor, Morocco’s poor face severe drought occurrences, and Egypt’s poor to flooding. 23 Figure 19: In absolute terms, Egypt has the highest number of poor ($6.85) exposed to each of the four hazards: heat stress, drought risk, air pollution and flooding 3.7 Shares of Vulnerability Uncover Pronounced Disparities in Shares within Countries Figure 20 provides a detailed breakdown of vulnerability by sub-regions within countries, focusing on those exceeding the regional vulnerability threshold of 0.42. The analysis iden- tifies sub-regions in Syria, Bahrain, Yemen, Morocco, Saudi Arabia, the West Bank and Gaza, and Djibouti as having the highest vulnerability shares in the region. This assessment highlights substantial heterogeneity within countries; for instance, Al Qunaytirah in Syria records the highest 24 vulnerability share in the region, while other Syrian sub-regions such as Lattakia and Homs exhibit considerably lower vulnerability levels. A closer examination in Figure 20 reveals that Al Qunaytirah’s elevated vulnerability is linked to multiple factors: limited access to essential services, including healthcare, education, and markets; low economic activity; a high frequency of conflict events in recent years; and significant food insecurity, as evidenced by aggregate yield gaps. Al Qunaytirah, located in southern Syria bordering Lebanon and Jordan, is divided into the districts of Qunaytirah and Al-Fiq. The region has sustained extensive damage, with the Armed Conflict Location & Event Data Project (ACLED) documenting 25 security incidents in 2019 (an average of 0.5 incidents per week), predominantly involving explosions, remote violence, and battles. In early 2020, the frequency of security incidents increased to an average of one per week. A recent report by the underscores the severe humanitarian crisis and food insecurity in Qunaytirah, where 15.07% of the population resided in damaged buildings as of 2021. The composition of vulnerability in these sub-regions is largely influenced by factors such as access to social protection, financial services, and the proximity to markets, which include urban centers, airports, and seaports as observed in Figure 20. Additionally, the availability of educational and health facilities is a substantial component of the vulnerability dimensions for sub-regions that surpass the regional vulnerability threshold. Sub-regions of Yemen are the highest number within the list despite the pronounced variation between sub-regions. Among them, the most populated are Hajjah and Sana’a with extremely low levels of access to basic services such as water, electricity, education, or social protection, as well as high levels of food insecurity. Agriculture and animal husbandry are key economic activities, where the governorate produces about 4.6% of the country’s agriculture produce but deals with high levels of food insecurity. Poor water availability and a lack of sanitation services, especially for IDPs, has increased the incidence of cholera and other diseases. Following the destruction of several hospitals and health centers, the governorate is suffering from an absence of health services. 3.8 The Analysis Reveals That the Top 10 Administrative Level 1 Subregions with Populations That Are Poor, Exposed to Any Hazard, and Have a High Share of Vulnerability Include Al Daqahliyah in Egypt; Nord Ouest (Now Known as Gharb - Chrarda - B´ e touan 3) in Morocco; e n i Hssen and Tanger - T´ and Sana’a, Dhamar, Hadramaut, Taizz, and Ibb in Yemen The intersection of poverty, exposure to hazards and vulnerability is analyzed in this section of the study. The focus up to this point has been on looking at people, especially those living below the poverty line, who are impacted by hazards disproportionately and lack the resources to recover from climatic shocks. Specifically, exposure to four key hazards have been examined: high levels of air pollution, extreme heat stress, flooding and drought. When examining vulnerability, we identified seven different dimensions of vulnerability, and in Figure 20, we show the sub-regions with the highest shares of vulnerability above 0.42. 3The analysis uses the old classification using the GAUL Administrative Boundaries of 2015. 25 Figure 20: Within sub-region (ADM1) vulnerability with share of vulnerability >0.42 26 Figure 21 illustrates these populations, identifying those who are poor (earning below $3.65 per day), exposed to any of the aforementioned hazards, and residing within 10km grid cells with a vulnerability share above 0.42. The analysis reveals that the top 10 administrative level 1 sub-regions with populations that are poor, exposed to any hazard, and have a high share of vulnerability include Al Daqahliyah in Egypt; Nord Ouest in Morocco; and Sana’a, Dhamar, Hadramaut, Taizz and Ibb in Yemen. Al Daqahliyah is notably the most affected by flooding and is particularly vulnerable to lack of forest cover, compounded by limited access to markets and a low level of non-oil economic activity. The number of poor individuals in this region who are exposed and highly vulnerable is 2,468,880. Figure 21: Number of Poor ($3.65/day), Exposed to Any of the Four Hazards, and Vulner- able[share of vulnerability >0.42] Sana’a in Yemen and Nord Ouest in Morocco are among the most affected sub-regions with 1,491,688 and 1,781,674 poor individuals, respectively, facing exposure and high levels of vulnerability. Figure 22 illustrates the breakdown at the 10km grid cell level the number of poor exposed and vulnerable. In Sana’a in Yemen, the confluence of poverty, exposure to hazards and vulnerable is most acute within the country. The lack of access to social protection and essential services like electricity, water, and education is the primary driver of vulnerability (shown in Figure 20). The scarcity of financial services, including access to a bank account, aggravates the condition of the poor in the region. The sub-region’s limited access and low yield gap ratios suggest a heightened risk of food insecurity. In Morocco, the Nord Ouest region emerges as the most affected. The area suffers from significant forest cover loss and food insecurity, which are the main contributors to its vulnerability. Additionally, the region is characterized by high flood risk, which represents the most considerable hazard exposure to its poor populations. 27 Figure 22: Poor ($3.65), Exposed to any Four Hazards, and Vulnerable[share >0.42] Note: We are not displaying the Western Sahara area here 4 Discussion This study offers a high-resolution analysis of the intersection between poverty, exposure to climate hazards, and vulnerability in the Middle East and North Africa (MENA) region. By integrating geospatial data with poverty estimates and hazard indices, the study highlights significant disparities in vulnerability across and within countries, revealing critical insights for targeted policy interventions. The results indicate that the MENA region faces a ”triple environmental challenge” of air pollution, extreme heat stress, and severe drought, each of which disproportionately affects poor populations. Particularly in Yemen, Morocco, and parts of Syria, the compounding effect of multiple hazards on vulnerable populations un- derscores the urgency for comprehensive resilience-building efforts. In Yemen, for example, limited access to essential services, such as clean water, healthcare, and electricity, exacer- bates the vulnerability of impoverished communities already exposed to frequent droughts and high levels of food insecurity. Similarly, Morocco’s Nord Ouest region, heavily impacted by forest cover loss and flood risks, experiences heightened vulnerability among its poor populations, reinforcing the need for both environmental and economic support systems. A notable contribution of this study is the use of a multidimensional approach to measure vulnerability, incorporating economic, social, and environmental factors. This approach surpasses traditional poverty metrics by accounting for complex vulnera- bilities, such as limited access to social protection, inadequate infrastructure, and high rates of food insecurity, that amplify the effects of climate hazards. For instance, in Syria’s Al Qunaytirah region, overlapping vulnerabilities from conflict, restricted market access, and limited healthcare availability illustrate how socio-political factors intersect with environmen- tal hazards to intensify risk. These findings align with prior research on the compounding effects of climate and conflict in fragile states, underscoring the need for integrated strategies that address both environmental and social drivers of vulnerability. Our methodology, which utilizes a 10km grid-level resolution, represents an advancement 28 in spatially disaggregated poverty estimation for the MENA region. By combining national household survey data with the Relative Wealth Index, this approach provides more granular poverty estimates that enhance our understanding of exposed populations at a subnational level. While this high-resolution poverty data offers new insights, further efforts are needed to validate these estimates across different time points and under varying socio-economic conditions. Future research could focus on creating poverty proxies through geospatial and micro data in areas lacking recent household surveys, which remains a challenge in conflict- affected regions like Libya and Syria. The study also presents innovative geospatial indicators of vulnerability, such as yield gaps, conflict incidence, and nighttime light intensity as a proxy for economic activity, to capture localized risks. These indicators reveal distinct spatial patterns of vulnerability, which can inform region-specific adaptation strategies. For example, the high yield gap ratios in Yemen’s agricultural areas, coupled with frequent droughts, signal a critical need for agricultural resilience programs to address food insecurity. Similarly, regions with significant nighttime light declines, such as parts of Islamic Republic of Iran and Saudi Arabia, may benefit from targeted economic development efforts to reduce vulnerability related to economic stagnation. Despite the study’s contributions, limitations exist. The reliance on thresholds to define hazard severity may not fully capture the nuanced impacts of moderate shocks, which can still have significant cumulative effects on vulnerable populations. Moreover, the exclusion of certain climate-related hazards, such as cyclones, limits the study’s applicability to regions where such hazards are prevalent. Additionally, while this analysis highlights the present- day distribution of vulnerability, there remains a need for longitudinal studies that project future hazard impacts under various climate change scenarios. Incorporating predictive models would allow policy makers to anticipate shifts in exposure and vulnerability, enabling proactive adaptation measures. In conclusion, this study underscores the urgent need for targeted resilience-building interventions across the MENA region, especially in high-risk areas where poverty, hazard exposure, and vulnerability intersect. Policy makers should prioritize investments in social protection, infrastructure, and environmental restoration, particularly in sub-regions where vulnerability is driven by multiple overlapping factors. The methodology and findings pre- sented here serve as a foundation for further research and operationalization in climate- vulnerable regions, both within MENA and beyond, contributing to a growing body of evidence aimed at mitigating the disproportionate impacts of climate change on the world’s poorest communities. 29 References Acharya, V. V., Johnson, T., Sundaresan, S., & Tomunen, T. (2022, September). Is phys- ical climate risk priced? evidence from regional variation in exposure to heat stress (Working Paper No. 30445). National Bureau of Economic Research. https://doi. org/10.3386/w30445 Adler, C., Wester, P., Bhatt, I., Huggel, C., Insarov, G., Morecroft, M., Muccione, V., & Prakash, A. (2022). Cross-chapter paper 5: Mountains. In H. O. Portner, D. C. Roberts, M. Tignor, E. S. Poloczanska, K. Mintenbeck, A. Alegr´ıa, M. Craig, S. Langsdorf, S. L¨ o schke, V. Moller, A. Okem, & B. Rama (Eds.), Climate change 2022: Impacts, adaptation and vulnerability. contribution of working group ii to the sixth assessment report of the intergovernmental panel on climate change (pp. 2273–2318). Cambridge University Press. https://doi.org/10.1017/9781009325844.022.2273 Auffhammer, M., & Schlenker, W. (2014). Empirical studies on agricultural impacts and adaptation. Energy Economics, 46, 555–561. https://doi.org/https://doi.org/10. 1016/j.eneco.2014.09.010 Baylis, P. (n.d.). Temperature and temperament: Evidence from a billion tweets [Job Market Paper]. Include%20URL%20if%20available Burgess, R., Deschenes, O., Donaldson, D., & Greenstone, M. (2017). Weather, climate change and death in india [Unpublished manuscript]. Burke, M., Hsiang, S. M., & Miguel, E. (2015). Climate and conflict. Annual Review of Economics, 7 (1), 577–617. https://doi.org/10.1146/ANNUREV- ECONOMICS- 080614-115430 Chi, G., Fang, H., Chatterjee, S., & Blumenstock, J. E. (2022). Microestimates of wealth for all low- and middle-income countries. Proceedings of the National Academy of Sciences, 119 (3), e2113658119. https://doi.org/10.1073/pnas.2113658119 Deryugina, T., & Hsiang, S. (2014, December). Does the environment still matter? daily temperature and income in the united states (Working Paper No. w20750) (NBER Working Paper No. w20750). National Bureau of Economic Research. https://ssrn. com/abstract=2538336 Doan, M. K., Hill, R., Hallegatte, S., Corral, P., Brunckhorst, B., Nguyen, M., Freije- Rodriguez, S., & Naikal, E. (2023). Counting people exposed to, vulnerable to, or at high risk from climate shocks: A methodology [World Bank publication]. http: //www.worldbank.org/prwp Economic, U. N., for Western Asia (ESCWA), S. C., et al. (2023). Second arab multidimen- sional poverty report. United Nations. FAO. (2023). Agricultural stress index system (asis) [[Online; accessed July 2024]]. Food Agriculture Organization of the United Nations. https://www.fao.org/giews/earthobservation/ Graff Zivin, J., & Neidell, M. (2014). Temperature and the allocation of time: Implications for climate change. Journal of Labor Economics, 32 (1), 1–26. https://doi.org/10. 1086/671766 Guo, Y., Gasparrini, A., Armstrong, B., Li, S., Tawatsupa, B., Tobias, A., Lavigne, E., De Sousa Zanotti Stagliorio Coelho, M., Leone, M., Pan, X., Tong, S., Tian, L., Kim, H., Hashizume, M., Honda, Y., Guo, Y.-L. L., Wu, C.-F., Punnasiri, K., Yi, S.-M., & Williams, G. (2014). Global variation in the effects of ambient temperature on 30 mortality: A systematic evaluation. Epidemiology, 25 (6), 781–789. https://doi.org/ 10.1097/EDE.0000000000000165 Hallegatte, S., Vogt-Schilb, A., Bangalore, M., & Rozenberg, J. (2017). Unbreakable: Building the resilience of the poor in the face of natural disasters. World Bank. https://doi. org/10.1596/978-1-4648-1003-9 Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S. J., Loveland, T. R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O., & Townshend, J. R. G. (2013). High-resolution global maps of 21st-century forest cover change. Science, 342 (6160), 850–853. https://doi. org/10.1126/science.1244693 Heal, G., & Park, J. (2013). Feeling the heat: Temperature, physiology & the wealth of nations (tech. rep. No. w19725). National Bureau of Economic Research. https://doi.org/10. 3386/w19725 Heger, M., Vashold, L., Palacios, A., Alahmadi, M., Bromhead, M.-A., & Acerbi, M. (2022). Blue skies, blue seas: Air pollution, marine plastics, and coastal erosion in the middle east and north africa. The World Bank. https://doi.org/10.1596/978-1-4648-1812-7 Hidalgo, F. D., Naidu, S., Nichter, S., & Richardson, N. (2010). Economic determinants of land invasions. Review of Economics and Statistics, 92 (3), 505–523. https://doi.org/ 10.1162/REST a 00007 Hsiang, S., Kopp, R., Jina, A., Rising, J., Delgado, M., Mohan, S., Rasmussen, D. J., Muir- Wood, R., Wilson, P., Oppenheimer, M., Larsen, K., & Houser, T. (2017). Estimating economic damage from climate change in the united states. Science, 356 (6345), 1362– 1369. https://doi.org/10.1126/science.aal4369 Hsiang, S. M., Burke, M., & Miguel, E. (2013). Quantifying the influence of climate on human conflict. Science, 341 (6151), 1235367. https://doi.org/10.1126/science.1235367 Hsiang, S. M., & Meng, K. C. (2015). Tropical economics. American Economic Review, 105 (5), 257–61. https://doi.org/10.1257/aer.p20151030 HSIANG, S. M., & NARITA, D. (2012). Adaptation to cyclone risk: Evidence from the global cross-section. Climate Change Economics, 03 (02), 1250011. https://doi.org/ 10.1142/S201000781250011X Jones, M. W., Abatzoglou, J. T., Veraverbeke, S., Andela, N., Lasslop, G., Forkel, M., Smith, A. J. P., Burton, C., Betts, R. A., van der Werf, G. R., Sitch, S., Canadell, J. G., Sant´ın, C., Kolden, C., Doerr, S. H., & Le Q u´ e , C. (2022). Global and regional trends e r´ and drivers of fire under climate change [e2020RG000726 2020RG000726]. Reviews of Geophysics, 60 (3), e2020RG000726. https : / / doi . org / https : / / doi . org / 10 . 1029 / 2020RG000726 Omidvar, N., Ghazi-Tabatabie, M., Sadeghi, R., Mohammadi, F., & Abbasi-Shavazi, M. J. (2013). Food insecurity and its sociodemographic correlates among afghan immigrants in Iran. Journal of Health, Population, and Nutrition, 31 (3), 356–366. https://doi. org/10.3329/jhpn.v31i3.16828 Osipov, S., Chowdhury, S., Crowley, J. N., et al. (2022). Severe atmospheric pollution in the middle east is attributable to anthropogenic sources. Communications Earth & Environment, 3 (1), 203. https://doi.org/10.1038/s43247-022-00514-6 31 Raleigh, C., Kishi, R., & Linke, A. (2023). Political instability patterns are obscured by conflict dataset scope conditions, sources, and coding choices. Humanities and Social Sciences Communications, 10, 74. https://doi.org/10.1057/s41599-023-01559-4 Rentschler, J., Salhab, M., & Jafino, B. A. (2022). Flood exposure and poverty in 188 countries. Nature Communications, 13, 3527. https : / / doi . org / 10 . 1038 / s41467 - 022-30727-4 Roberts, M. J., & Schlenker, W. (2013). Identifying supply and demand elasticities of agri- cultural commodities: Implications for the us ethanol mandate. American Economic Review, 103 (6), 2265–95. https://doi.org/10.1257/aer.103.6.2265 Rom´ a n, M. O., Wang, Z., Sun, Q., Kalb, V., Miller, S. D., Molthan, A., Schultz, L., Bell, J., Stokes, E. C., Pandey, B., & Seto, K. C. (2018). Nasa’s black marble nighttime lights product suite. Remote Sensing of Environment, 210, 113–143. https://doi.org/10. 1016/j.rse.2018.03.01 Stokes, E. C., & Rom´ a n, M. O. (2022). Tracking covid-19 urban activity changes in the middle east from nighttime lights. Scientific Reports, 12, 8096. https://doi.org/10. 1038/s41598-022-12211-7 van Donkelaar, A., Hammer, M. S., Bindle, L., Brauer, M., Brook, J. R., Garay, M. J., Hsu, C. N., Kalashnikova, O. V., Kahn, R. A., Lee, C., Levy, R. C., Lyapustin, A., Sayer, A. M., & Martin, R. V. (2021). Monthly global estimates of fine particulate matter and their uncertainty [Erratum published in Environmental Science Technology, 2024, 58(9): 4463-4464. doi: 10.1021/acs.est.4c01477]. Environmental Science Technology, 55 (22), 15287–15300. https://doi.org/10.1021/acs.est.1c05309 World Bank. (2023). Geospatial poverty portal [Accessed July 2024]. https://www.pipmaps. worldbank.org World Bank. (2024a). Lebanon poverty and equity assessment 2024: Weathering a pro- tracted crisis (tech. rep.). World Bank Group. Washington, D.C. http://documents. worldbank.org/curated/en/099052224104516741/P1766511325da10a71ab6b1ae97816dd20c World Bank. (2024b). Yemen poverty and equity assessment: Living in dire conditions (tech. rep.). World Bank Group. Washington, D.C. http : / / documents . worldbank . org / curated/en/099030424081530447/P17919414df5860011a33814c87a62dce86 WorldPop & Center for International Earth Science Information Network. (2018). Global high resolution population denominators project [School of Geography and Envi- ronmental Science, University of Southampton; Department of Geography and Geo- sciences, University of Louisville; Departement de Geographie, Universite de Namur]. https://doi.org/10.5258/SOTON/WP00671 Zittis, G., Almazroui, M., Alpert, P., Ciais, P., Cramer, W., Dahdal, Y., Fnais, M., Fran- cis, D., Hadjinicolaou, P., Howari, F., Jrrar, A., Kaskaoutis, D. G., Kulmala, M., Lazoglou, G., Mihalopoulos, N., Lin, X., Rudich, Y., Sciare, J., Stenchikov, G., . . . Lelieveld, J. (2022). Climate change and weather extremes in the eastern mediter- ranean and middle east [e2021RG000762 2021RG000762]. Reviews of Geophysics, 60 (3), e2021RG000762. https://doi.org/https://doi.org/10.1029/2021RG000762 32 5 Appendix 5.1 Hazards 5.1.1 Heat Stress Index The WBGT data is sourced from the global daily maximum air temperature readings from the ERA-Interim reanalysis fields for the period 1981-2010. This three-decade span pro- vides a solid foundation for climate analysis. Using a high-resolution global digital elevation model, the data is adjusted for altitude on a 0.75-degree latitude/longitude scale. This pro- cess yields global daily maximum WBGT fields at an approximate 10km spatial resolution. The data is further refined by applying a 3-day smoothing filter, aiding in the evaluation of prolonged heat effects. The final output is daily maximum WBGT values, smoothed over 3 days and spanning 30 years, at a 10km resolution. These values create a Generalized Extreme Value (GEV) distribution for each grid cell. The distribution uses hazard intensity maps at a 10-km resolution for 5, 20, and 100-year return periods. In our analysis, we create the upper bound of the probability distributions of the three return periods by bootstrapping the data. We compute the upper bound by taking the expected value of the three GEV distributions. 5.1.2 Flood Risk The Fathom dataset is highly detailed, with a 3 arcsecond resolution, and covers areas be- tween 56°S and 60°N globally. It has demonstrated accuracy in predicting flood events, such as the recent floods in Libya. However, due to the global nature of the model’s assump- tions, the data is not suitable for detailed engineering-level analysis as it does not account for constructed flood protection systems. The maps used in this study are the undefended flood raster files, publicly available through the Global Facility for Disaster Reduction and Recovery (GFDRR). Fathom has released updated estimates for 2023, featuring enhanced spatial accuracy, which can be used for future analyses. In the flood risk analysis as well, we take the expected value of all the return periods and generate a flood risk expected value for water depth to which we apply the threshold of greater than 50cm. 5.2 Data Sources 33 Table 1: Detailed List of Data Sources HAZARDS Measure Threshold Spatial Source Resolution EXTREME HEAT Expected value Temperature > 10km Ridder et al. STRESS of the annual 32°C, considered (2017). maximum of an acute health Indicator average daily risk (OCHA) developed by maximum the World Bank wet-bulb and used by temperature GFDRR over 3-day running average FLOOD RISK* Expected value Water depth > 90m FATHOM of water level 0.5m for a given year (fluvial) SEVERE DROUGHT 30% of cropland Severe drought 1km FAO RISK affected by more than 20% Agriculture severe drought of the time Stress Index more than 20% of the time AIR POLLUTION Average annual 15µg/m3 (WHO 1.1km Van Donkelaar (PM 2.5) PM 2.5 daily threshold) et al., 2021 concentration over the period (2016-2021) *Fluvial data for Bahrain is missing. EXPOSURE Measure Spatial Source Resolution ADMIN SHAPES Administrative - Humanitarian Boundary level Data Exchange 2 Portal Continued on next page 34 Asset-wealth 1km Blumenstock et RELATIVE index measured al. (2021) WEALTH INDEX** within the country, values range from -1 to 1 Poverty rates - Global POVERTY for thresholds Subnational RATES*** $2.15/day, Atlas of Poverty $3.65/day, and (GSAP), World $6.85/day at Bank administrative or national levels POPULATION Number of 1km WorldPop people per km2 for 2020 (unconstrained) **Data for Yemen, the Islamic Republic of Iran, Iraq, the West Bank and Gaza, Syria, and Saudi Arabia are not available. ***Data for Libya, Syria, Saudi Arabia are not available. VULNERABILITY Measure Threshold Spatial Source Resolution NIGHTTIME Average annual Luminosity > 0, 450m or 15-arc Black Marble, LIGHTS luminosity for proxy for second NASA 2020 economic activity AGGREGATE Sum of yield Yield 6.7km or 5 FAO YIELD- achievement Achievement arc-minutes ACHIEVEMENT ratios over 22 Ratio > 5 RATIO major crops/crop groups MARKETS Number of big Travel time to 10km OpenStreetMap cities, airports, these markets < and seaports 1 hour within a 10km grid 35 Continued on next page 36 EDUCATIONAL Number of Travel time to 10km OpenStreetMap INSTITUTIONS institutions these within a 10km institutions < 1 grid hour HEALTH Number of Travel time to 10km OpenStreetMap FACILITIES facilities within these facilities < a 10km grid 1 hour CONFLICT Number of - Armed Conflict EVENTS conflict events Location and within a 10km Event Data grid (ACLED) FOREST COVER Binary variable Any grid cell 30m/pixel at Global Forest LOSS indicating any with value 1 the equator Change Data, loss within the UoMaryland 10km grid (2000-2022) 37 5.3 Downscaling Poverty Estimates with RWI A significant obstacle in the Middle East and North Africa, as well as globally, is the scarcity of detailed poverty data. Typically, poverty estimates are only available at broader adminis- trative (admin 1 or 2) or national levels. Our study addresses this by employing the Relative Wealth Index to refine these general estimates to more precise 10-kilometer grid levels. This analysis stands out for its innovative approach: using the Relative Wealth Index to ad- just poverty estimates from larger administrative or national levels to a detailed 10-kilometer grid scale. We incorporate two key datasets: the Global Subnational Atlas of Poverty from the Poverty and Equity group and the Relative Wealth Index developed by Meta, Facebook. A limitation of the Global Subnational Atlas of Poverty is its lack of completeness in many countries and, where available, its confinement to admin 1 or 2 levels. A limitation of the Relative Wealth Index data is that it is time-invariant, in that it used a combination of different sources including household surveys, spatial data and mobility information to estimate these levels of wealth. Therefore, it would be a good assumption that this estimate will not hold in the future unless the underlying sources of data are updated. However, since we use publicly available sources of data, this analysis is replicable within and outside the World Bank network. A key contribution of our analysis is its ability to estimate the number of poor affected by hazards at a highly detailed spatial resolution. For this, we use the Relative Wealth Index and available poverty estimates, re-weighting the poverty share through specific steps for countries where these data are accessible. Step 1: Aggregation of the RWI to a 10km Grid. Initially, we address the granularity mismatch as the RWI is available at a 2.4km resolution while our analysis operates on a 10km-by-10km grid. To reconcile this, we aggregate the RWI data to match our 10km grid, ensuring that the wealth index is appropriately scaled to the unit of analysis, the 10-km grid. Step 2: Alignment of Aggregated RWI with GSAP Data. The Global Subnational Atlas of Poverty (GSAP) provides poverty estimates at administrative levels 1 or 2, or at national scales, based on household surveys. By overlaying the aggregated 10km RWI grid onto the GSAP data, we can assign poverty estimates to specific 10km grid areas, even within larger admin level polygons. Step 3: Re-weighting the RWI values within GSAP Data. The objective of Step 3 is to redistribute poverty estimates within each administrative region (admin 1 or 2), by applying Relative Wealth Index (RWI) as weights. Our goal is to maintain the overall GSAP poverty level reported for each region but to reallocate the distribution of that poverty within the region according to RWI values. This adds a layer of spatial detail and enhances the under- standing of poverty’s distribution. For areas without RWI data, we apply the default GSAP poverty rate to ensure uniformity. 38 Figure 23 illustrates the process of re-weighting to transform ’relative wealth index’ (RWI) values into a measure of ’relative poverty.’ This process involves several steps: 1. The left side of the figure shows six boxes, labeled A to F, each representing an inverse RWI value (calculated as -True RWI + 1.5). This modification adjusts the RWI scale, which originally ranges from -1.5 to +1.5, to represent ’relative poverty’ on a scale from 0 to 1. Surrounding these boxes are larger ones, colored blue (representing a 40% poverty rate) and yellow (representing a 60% poverty rate), as per the Global Subnational Atlas of Poverty (GSAP) rates. 2. The next step involves calculating the population weighted average RWI for each col- ored box (like a polygon at admin level 2). For instance, in the blue box, this average is computed as (0.5*100 + 1*200)/300, resulting in 0.83. In the yellow box, the calcu- lation is (0.8*800 + 0.3*200 + 1*0 + NA*200), yielding an average of 0.7. 3. Subsequently, we determine the RWI weight for each box, using the formula (-True RWI +1.5)/ Population-weighted average. This results in RWI weights of 0.6 for box A, 1.2 for box D, 1.14 for box B, 0.42 for box C, 1.42 for box E, and NA for box F. 4. In the final step, we apply these RWI weights to the GSAP poverty rates. In cases where RWI data is missing, we directly use the GSAP poverty rate. For boxes within the blue area, for example, box A’s adjusted poverty rate is 24% (0.6 multiplied by the GSAP rate of 0.4), and box D’s is 48% (1.2 multiplied by 0.4). The same method is applied to calculate the adjusted poverty rates for the remaining boxes. 39 Figure 23: Explaining the weighting methodology (B) Depiction of the re-weighting of poverty rates using RWI when we overlay the GSAP map Figure 24: Map showcasing the disaggregated estimates using RWI In Figure 23, (A) provides a clear explanation of the weighting methodology. It details the steps and formulas used to convert RWI values into a measure of relative poverty. (B) visually demonstrates the re-weighting of poverty rates using RWI. This is depicted through an overlay of two maps: the Global Socioeconomic Allocation of Poverty (GSAP) map on 40 the top and the RWI map at the bottom. The map showcases how poverty rates are adjusted using the RWI data. Notably, darker brown pixels on the map on the right highlight the variations in poverty rates that this method can achieve. 5.4 Robustness of RWI Estimates Figure 25: Correlation between Figure 26: Correlation between GSAP and ”GSAP+RWI” methods GSAP and ”GSAP+RWI” methods for $2.15/day for $6.85/day Figure 27: Correlation between GSAP Figure 28: Correlation between GSAP and ”GSAP+RWI” methods for poor ex- and ”GSAP+RWI” methods for poor ex- posed to heat stress across countries posed to air pollution across countries 41