The Geography of Global Vulnerability

Each of the four hazards affects a large share of humanity. In the recent climate, 1.93 billion people live with frequent hot-humid heat, 1.62 billion within reach of hurricane-force winds, 1.52 billion where five-day rainfall extremes exceed 150 mm and 1.02 billion with long dry seasons. By 2040–2059 under SSP2-4.5, with 2050 populations, the numbers rise to 4.14, 1.92, 2.27 and 1.50 billion. Heat grows fastest, because warming pushes the 35 °C heat-index threshold across the densely populated tropics.

Overlaps matter because compound hazards strain the same household, the same health system and the same budget in quick succession. Today 1.30 billion people face two hazards and 369 million three. Almost nobody faces all four, because cyclones and long dry seasons rarely meet. By 2050, two-hazard exposure rises to 2.00 billion and three-hazard exposure to 907 million, as heat is added to places that already flood or face cyclones.

The exposed populations are concentrated in Asia. India has 569 million people in compound-hazard cells, China 328 million, Bangladesh 157 million (95% of its people), the Philippines 80 million and Vietnam 72 million. Most combine heat, monsoon rainfall and cyclones. These are the places a hazard map would flag first. Section 5 explains why most of them are not hotspots.

The map crosses the two dimensions for every inhabited 1° cell. Columns of the key run from none or one hazard to three or more, rows from adequate to very limited capacity. Pale blue to deep blue marks places where hazards pile up but capacity is adequate. This covers most of South Asia, eastern China, mainland South-East Asia, the Philippines and the US Gulf coast. Amber marks places with limited capacity but at most one hazard, across much of East, Central and Southern Africa.

The red and magenta cells are the hotspots. In the recent climate they form a belt across the Sahel, from Senegal and Mauritania through Mali, Burkina Faso, Niger and Chad to Sudan, where frequent heat meets long dry seasons and very limited capacity. A second cluster runs through the Indus valley in Pakistan, and smaller ones cover Madagascar, Haiti, North Korea and parts of Myanmar, where cyclones meet extreme rainfall.

Switch to 2040–2059 and the red spreads. Frequent hot-humid heat spreads across the rest of the Sahel and its southern margins, deep into northern Nigeria, Niger, Chad, northern Cameroon and Senegal. These are some of the fastest-growing populations on Earth, which is why the hotspot population almost triples. The 2050 view holds capacity at today's level. If development lifts capacity across these regions, much of the new red would turn blue.

Adaptive capacity is the ability of people and institutions to anticipate, absorb and recover from shocks. It depends on income and savings, housing and infrastructure, health and social protection, information and governance. This dashboard uses two measurable proxies. Locally, it counts places where at least 40% of people live below $4.20 a day, which leaves little room to rebuild, relocate or invest in protection. Nationally, it uses the INFORM lack-of-coping-capacity score, which combines institutional and infrastructure indicators, and counts countries scoring 6 or more.

By these measures, 1.37 billion people have limited adaptive capacity: 1.10 billion through local poverty and 1.06 billion through weak national coping, with 790 million meeting both. Limited capacity is concentrated in sub-Saharan Africa, in Pakistan and Afghanistan, in Haiti, and in fragile states such as Yemen and North Korea, with pockets in Myanmar. Most of the populations of South and East Asia sit above both lines.

Crossed with hazards, the overlap is narrower than either layer. Only 158 million of the 1.66 billion people facing compound hazards also have limited capacity. Low-income countries are the exception: all 78 million of their compound-exposed people are in hotspots. In lower-middle-income countries the share is 9%, and in upper-middle- and high-income countries almost zero. Hazard and capacity mostly sort into different places, so the vulnerable minority can be targeted precisely.

Pakistan has the largest hotspot population, 49 million people, mostly in the hot, dry Indus plain, where frequent extreme heat meets long dry seasons and poverty above 40%. The 2022 floods showed how such places can also be swamped by extreme rain. North Korea follows with 17 million, where cyclones and extreme rain meet weak national coping. Then come Madagascar (13 million), Sudan (13 million), Nigeria (12 million, in the far north), Mali (10 million), Niger (9 million) and Burkina Faso (7 million).

As shares of national populations, the Sahel stands out. Nearly half the people of Mali and Madagascar live in hotspots, 43% in Niger, a third in Burkina Faso, and 29% in Chad and Sudan, as do a quarter of Haitians and of Pakistanis. In these countries, compound climate risk on fragile foundations is the national condition, not a regional exception.

Heat combined with long dry spells accounts for 71% of hotspot people, from the Sahel to Pakistan. Extreme rain combined with cyclones accounts for 24%, in Madagascar, Haiti, North Korea, Myanmar and the Philippines. Few hotspot people face three hazards today, but by 2050 the number rises from 2 million to 12 million as heat extends into rain-and-cyclone coasts.

The largest compound-hazard populations are not hotspots on these thresholds. India has 569 million people where two or more hazards overlap, China 328 million, Bangladesh 157 million, the Philippines 80 million and Vietnam 72 million. India's latest national poverty rate at $4.20 a day is 16%, Bangladesh's 21% and Vietnam's lower still, so few places cross the 40% local threshold. All have national coping scores below INFORM's high range. Only the Philippines has a small hotspot, 2 million people.

This is a real achievement. Bangladesh's cyclone preparedness, India's heat action plans and early-warning systems, and rising incomes across Asia have reduced disaster deaths and kept many hazards from becoming disasters. But it is a race. Heat exposure in South Asia is rising fast, the poverty threshold used here is not far above many districts' rates, and the poorest households within these countries face the same hazards with fewer resources than national averages suggest.

The threshold choice matters most here. With a 30% local-poverty threshold instead of 40%, 14 million people in Bangladesh and 7 million in the Philippines would join the hotspots, and the global total would rise from 158 million to 182 million; India would still have none. Read the result as follows: these countries are exposed but, on average, coping. Keeping capacity rising faster than hazard is the central adaptation task for the world's most exposed region, and the hotspots show what happens where it does not.

Projecting to 2050 with SSP2 population and SSP2-4.5 hazards, and holding capacity at today's level, the hotspot population rises to 455 million. Separating the drivers: with 2050 population but today's climate it would be 241 million, so population growth adds about 83 million and warming about 214 million. The increase is overwhelmingly in sub-Saharan Africa, where hotspot populations rise from 79 million to 333 million: Nigeria from 12 to 107 million, Niger from 9 to 50 million and Sudan from 13 to 46 million.

Climate pathways shift the result. Under SSP1-2.6, with less warming and slower population growth, the 2050 figure is 356 million. Under SSP3-7.0, with faster growth and more warming, it is 556 million. SSP5-8.5 gives 422 million, because its lower population offsets its higher warming by mid-century. Emissions reductions matter most after 2050, when the pathways diverge most.

Capacity is the stronger lever by 2050. If local poverty were halved and national coping improved by a point, the 2050 hotspot population under SSP2-4.5 would be 263 million, lower than under SSP1-2.6 alone. Combining stronger capacity with SSP1-2.6 gives 204 million. The gain is lumpy, though. 72% of the 191 million reduction comes from Pakistan, Mali, North Korea, Senegal and Burkina Faso, where halved poverty or a one-point coping gain crosses the line. In Nigeria, Niger, Sudan, Madagascar and Chad, which hold 242 million of the 455 million in 2050, national coping is so weak (INFORM 7 or worse) that a one-point gain changes nothing. In those countries capacity has to improve much further, through peace, state capability and services, before it brakes vulnerability at all.

The country atlas summarises the grid. The hotspot layers show the number of people where compound hazards meet limited capacity, now and in 2050, and the share of each country's population that represents. The Sahel, Sudan, Pakistan, Madagascar, Haiti and North Korea dominate the present. By 2050 Nigeria, Niger and Sudan add the most people, and Niger, Sierra Leone, Sudan and Senegal see the largest rises in the share of their populations in hotspots.

The compound-hazard layer shows the share of people in places with two or more hazards, highest across South and South-East Asia, the Gulf, the Sahel and the cyclone coasts. The limited-capacity layer shows the share of people below the capacity line, highest across sub-Saharan Africa and fragile states. The coping layer shows the national INFORM score that feeds the capacity measure.

Hover over or tap a country for its card: population, people exposed to each hazard now and in 2050, compound-hazard and hotspot populations, poverty in compound-hazard places against the national rate, and the scenarios. Clicking a country loads it into the calculator in section 8. The sortable table below lists all 158 countries.

The calculator uses the grid results aggregated to each country. Choose a pathway: SSP1-2.6 (lower population, less warming), SSP2-4.5 (middle) or SSP3-7.0 (faster growth, more warming), or "today's climate" to isolate population growth. Then choose whether adaptive capacity stays as today or improves, meaning local poverty is halved and the national coping score improves by a point.

The outputs show the 2050 hotspot population against today's, the share of the projected population that represents, the change from climate and population, and the reduction from stronger capacity. The chart compares all six combinations of pathway and capacity for the chosen country, so the two levers can be compared directly.

Cyclone hazard projections are available only for SSP2-4.5, so cyclone exposure uses that pathway in all scenarios. Capacity improvement is stylised, and the poverty threshold is a judgement. Use the results to compare levers and countries, not as forecasts.

Target the overlap. Adaptation finance, social protection and humanitarian preparedness are scarce and are usually allocated by country. A small set of places holds most of the people where compound hazards meet limited capacity: the Sahel belt, Sudan, Pakistan's Indus plain, Madagascar, Haiti and North Korea today, with northern Nigeria and Niger growing fastest. Geographic targeting within countries, using gridded hazard and poverty data like those here, can reach them at far lower cost than blanket programmes.

Match the response to the combination. In heat-and-drought hotspots, the priorities are water security, drought-tolerant agriculture and livestock systems, heat-health action plans, cooling and shade, and shock-responsive cash transfers that scale up in bad years. In rain-and-cyclone hotspots, they are early warning that reaches the last mile, cyclone shelters and safer housing, drainage and flood-plain management, and pre-arranged finance for rapid recovery. Both need functioning local government and health systems.

Keep capacity ahead of hazard, and reduce the hazard. Strengthening capacity cuts the 2050 hotspot population by about 40%, more than the lower-emissions pathway does by mid-century, but the most fragile states need far more than a marginal gain to cross the line. In Asia's exposed but coping regions, keeping development and preparedness ahead of rising heat stops new hotspots from forming. And because hazard growth drives most of the increase to 2050 and diverges sharply after it, cutting emissions is what keeps the hazard side of the map within reach of adaptation later this century.

Grid: 0.25° (721 × 1,440 cells). Population: SSP grids from the World Bank Climate Change Knowledge Portal (Jones and O'Neill, GPWv4-based), 2020 and 2050 (SSP1, SSP2, SSP3 and SSP5). Hazards: CMIP6 multi-model median climatologies, 1995–2014 and 2040–2059: heat index above 35 °C on ≥ 30 days (hi35), maximum five-day precipitation ≥ 150 mm (rx5day) and maximum consecutive dry days ≥ 90 (cdd). Tropical cyclones: CHAZ ensemble median annual exceedance probability of category-1 winds ≥ 2%, 0.5°, 1951–2014 and 2035–2064 (SSP2-4.5 only; used for all 2050 pathways). Compound = 2+ hazards.

Capacity: local share below $4.20 a day (2021 PPP), taken from the GSAP gridded $3.20 (2011 PPP) pattern, or the $1.90 pattern where missing, and rescaled so each country's population-weighted mean equals its latest WDI SI.POV.LMIC value (2015 or later). Limited = local share ≥ 40% or national INFORM 2026 lack of coping capacity ≥ 6. Very limited = both. Improved = local share halved and coping score − 1. Hotspot = compound and limited. Countries are assigned with a rasterised world-atlas map. The 1° map shows population-weighted mean hazard count and capacity tier.

Caveats: thresholds are judgements; a 30% local-poverty threshold gives 182 m hotspot people now (493 m in 2050) and adds Bangladesh, while a 50% threshold gives 151 m (438 m). Floods from rivers and storm surge, sea-level rise, wildfire and wet-bulb extremes are not mapped directly. Climatologies are model medians. Poverty maps are old patterns rescaled to new totals. Countries without poverty data rely on the national coping score (for example, North Korea). The improved-capacity scenario is a step: it removes people only where halved poverty falls below 40% and the coping score below 6. Future capacity is not projected. Exposure is not impact.