The Geography of Growing Human Exposure

Each engine is well known on its own. The distinctive feature of the next quarter-century is that all three are concentrated in the same places. Population growth is concentrated in sub-Saharan Africa, South Asia and parts of the Middle East. Urbanisation is fastest in the same regions. And the tropical and subtropical lowlands where they meet are where heat-index days rise fastest and where large rivers and deltas flood.

This dashboard therefore asks a geographic question: not only how much exposure grows, but where, and through which engine. It decomposes the change in heat exposure from 2020 to 2050 into three parts. The growth part is what would happen if each country's population grew evenly with today's climate. The redistribution part reflects people moving within countries, for example from countryside to city. The climate part is the effect of warming on the 2050 population.

The answer differs by hazard. For heat, warming is the larger engine almost everywhere, but growth adds 0.44 billion exposed people, more than a fifth of the rise. For river floods in cities, there is no climate term in the data used here. The rise is entirely people settling in flood zones, which makes it the clearest case where planning, not only adaptation, determines exposure.

World population grows from 8.23 billion in 2025 to 9.66 billion in 2050 in WUP 2025, an increase of 1.43 billion. The growth is extraordinarily concentrated. Sub-Saharan Africa adds 820 million (57%), Central and Southern Asia 451 million and Northern Africa and Western Asia 202 million. Latin America adds 62 million. Eastern and South-Eastern Asia lose 111 million, and Europe and Northern America are roughly flat.

Five countries add 650 million between them: India 216 million, Nigeria 122 million, Pakistan 117 million, the DR Congo 105 million and Ethiopia 90 million, with Tanzania next at 59 million. China loses 156 million. Forty-two countries with 2.3 billion people today shrink, while the 119 that grow add 1.68 billion, so the net figure understates how much new housing, infrastructure and services the growing countries will need.

Those growing countries are also the more hazardous ones. Of the people added in countries that grow, 83% are in countries rated high or very high on the INFORM Risk Index 2026, and 43% are in countries with high natural-hazard scores. Growth adds people precisely where coping capacity is thinnest.

Under the Degree of Urbanisation, the world was 31% city, 38% town and 31% rural in 1975, and 45% city, 36% town and 20% rural in 2025. By 2050 cities reach 48%, towns 35% and rural areas 17%. The rural population still grows slowly, to a peak of about 1.65 billion in the mid-2040s. Cities add 986 million people by 2050, towns 406 million and rural areas 40 million.

Towns are the overlooked middle of this transition. They hold 2.9 billion people, more than a third of humanity. In sub-Saharan Africa towns add 318 million people by 2050 against 410 million in cities, and rural areas still add 93 million. Towns typically have weaker planning, drainage and health systems than cities, and far less data. Much of the new exposure will be in places that are growing too fast to be planned and are too small to be studied.

The countries adding the most city dwellers are India (+150 million), Nigeria (+82 million), Pakistan (+79 million), the DR Congo and Egypt (+54 million each), Bangladesh and Ethiopia (+49 million each). Egypt and Bangladesh add city dwellers while their towns shrink, as cities expand over the land of former towns. China's cities shrink slightly (−26 million) while its towns and countryside lose 130 million people.

In 2020, 1.93 billion people lived where the heat index exceeded 35 °C on at least 30 days a year in the 1995–2014 climate. By 2050, under SSP2-4.5 warming and SSP2 population, the figure is 4.14 billion. Holding each country's population distribution fixed and scaling it to 2050 adds 0.44 billion. Within-country shifts of where people live add only 0.02 billion in these grids. Warming adds 1.74 billion. Measured in person-days, the total rises from 162 billion to 439 billion a year, and the average person's days above the threshold go from 21 to 48.

The mix differs sharply by region. In South Asia, where 1.3 billion people already face such heat, growth adds 342 million exposed people and warming 347 million. In sub-Saharan Africa, warming reaches large populations for the first time: the exposed population rises from 90 million to 709 million, with 69 million from growth and 549 million from climate. East Asia's exposure rises by 653 million through warming alone as its population shrinks. North America goes from about 1 million to 61 million.

The scenarios bracket the outcome. Under SSP1-2.6, with lower population and less warming, 3.6 billion people are exposed by 2050. Under SSP3-7.0, with faster growth and more warming, 4.7 billion are. The countries adding the most exposed people are India (+543 million), China (+375 million), Nigeria (+289 million), the Philippines, Pakistan and Indonesia.

Grouping 0.25° grid cells by population density shows how urbanisation and heat combine. The densest cells (average density of 1,000 people per km² or more across the cell), home to 2.0 billion people, averaged 33 days a year above a 35 °C heat index in the recent climate, rising to 65 by 2050. Intermediate cells averaged 22, rising to 49. Sparse areas averaged 7, rising to 27. By 2050, 52% of people in the densest cells will face 30 or more such days, against 28% in sparse areas.

This is partly geography: the world's great dense populations are in the hot lowlands of South Asia, the North China Plain and West Africa. Cities also add their own heat island, which these grids do not capture, so the urban figures are conservative. Either way, concentrating growth in dense places concentrates exposure.

Growth also leans toward heat. Of the 1.75 billion people added in grid cells that grow between 2020 and 2050, 49% are in cells that will face 30 or more hot days by 2050, against 44% of the 2020 population living in such cells. In Pakistan, India, Nigeria and Bangladesh, more than four-fifths of growth lands in such places. Drying adds a further dimension: 630 million people in 2050 will live where the 12-month drought index falls by a standard deviation or more, up from 480 million on today's population.

The GHSL Urban Centre Database maps the 11,422 urban centres that match WUP's definition of cities, with the population living in JRC 100-year river flood zones (assuming no defences). That population grew from 338 million in 2000 to 507 million in 2025, faster than cities overall: the flood-zone share of city dwellers rose from 12.5% to 13.8%. Another 349 million city dwellers live less than 10 metres above sea level.

Applying each country's 2025 flood-zone share to WUP's projected 2050 city population gives 626 million city dwellers in flood zones by 2050. If the 2000–2025 trend in each country's share continues, the figure is 689 million. The largest additions are in India (+44 million), Egypt (+29 million, along the Nile valley and delta), Bangladesh (+22 million), China (+18 million), Pakistan (+14 million), Iraq and Nigeria.

Unlike heat, this rise owes nothing to climate change in the data used here, which hold flood zones fixed. It is the result of where cities grow. Flood defences protect part of these populations, especially in China, so the numbers measure exposure, not risk. But they show that land-use planning is itself a lever on exposure, and that climate-driven increases in heavy rainfall, projected for 180 million people's homes by 2050 on these grids, will add to it.

The growth layer shows population change from 2025 to 2050. It is above 50% across much of sub-Saharan Africa and parts of the Middle East, and negative across East Asia, Europe and parts of Latin America. The city-share layer shows the share of people living in cities in 2025, and the city-additions layer shows how many city dwellers each country adds by 2050.

The heat layers show days a year above a 35 °C heat index for the average person in 2050, and the number of people newly exposed to 30 or more such days between 2020 and 2050. They are highest across South Asia, the Sahel and West Africa, the Gulf and South-East Asia. The flood layer shows the share of city dwellers in 100-year flood zones, highest in Thailand, Iraq, Bangladesh, Viet Nam, Egypt and Cambodia, where cities sit on great river floodplains and deltas.

The hot-growth layer shows the share of each country's growth landing in places that will face frequent extreme heat. Hover over or tap a country for its card, which breaks down its population by settlement type and its exposure by engine. Clicking a country loads it into the calculator in section 8. The sortable table below lists all 161 countries.

Pick a country and a shared socioeconomic pathway. Each pathway pairs a population projection with a warming level: SSP1-2.6 (lower growth, less warming), SSP2-4.5 (middle), SSP3-7.0 (fast growth, more warming) and SSP5-8.5 (lower growth, most warming). The calculator shows the people living with 30 or more days a year above a 35 °C heat index in 2050, and splits the change since 2020 into the part from population growth and the part from warming.

For floods, choose whether each country's share of city dwellers in flood zones stays at its 2025 level or continues its 2000–2025 trend. Then set how much of the additional flood-zone growth to 2050 is avoided by steering new urban development to safer land. The output shows city dwellers in 100-year flood zones in 2050 and the people kept out of them.

These are exposure counts, not impacts. Exposure to frequent heat matters less where homes are cooled, work is adapted and health systems are ready, and flood zones matter less where defences are maintained. But exposure is the base on which adaptation must be built, and in fast-growing countries most of the 2050 exposure has not yet been built.

Plan expansion before it happens. Cities in sub-Saharan Africa and South Asia will add hundreds of millions of people, mostly on land that is not yet urban. Mapping flood zones, reserving safe land and infrastructure corridors ahead of settlement, and enforcing simple rules on where to build cost far less than relocating people or defending them later. River floods are the clearest case: their exposure growth in this analysis comes entirely from where cities grow.

Build for heat. Warming will expose about 1.7 billion more people to frequent extreme heat by 2050 regardless of where they live, and the densest places will see the most hot days. Shade, trees, reflective roofs, ventilated building design, cooling centres, heat early warning and adapted working hours are the urban fabric of heat adaptation, and they are cheapest to build in from the start. Towns, which add 400 million people, need the same attention as cities.

Target the overlap, and cut the hazard. The 25 countries where fast growth and frequent heat coincide, from Nigeria and Pakistan to the Sahel and the Gulf, will add 483 million people by 2050. They need finance for urban infrastructure as much as for disaster response. Warming is the largest engine of heat exposure, so every tenth of a degree avoided reduces the number of people exposed: SSP1-2.6 instead of SSP3-7.0 means 1.1 billion fewer exposed in 2050.

Settlements: UN DESA World Urbanization Prospects 2025 (file F01), population in cities, towns and rural areas by Degree of Urbanisation, 1950–2050, for 161 countries with at least one million people. Heat: World Bank CCKP 0.25° gridded population for SSP1–SSP5 (Jones and O'Neill, GPWv4-based), and CMIP6 multi-model median climatologies of days with heat index above 35 °C (hi35) for 1995–2014 and 2040–2059. Exposed = population in cells with at least 30 such days. Drying = SPEI-12 at or below −1 and heavy rain = Rx5day at least 20% above the historical value, both 2040–59 under SSP2-4.5. Countries are assigned with a rasterised world-atlas boundary map.

Decomposition (SSP2-4.5): base = 2020 population × historical mask; growth = Σ countries (2050 ÷ 2020 population) × base; redistribution = 2050 population × historical mask − growth; climate = 2050 population × 2040–59 mask − 2050 population × historical mask. Density classes use 2020 cell density (≥1,000, 150–1,000, <150 people per km² of cell area). Floods: GHSL Urban Centre Database R2024A population within JRC 100-year undefended river-flood zones, 2000 and 2025, aggregated by country; 2050 = WUP city population × 2025 share (constant) or × 2025 share × (2025 ÷ 2000 share), capped (trend).

Caveats: SSP2 grids total 9.2 billion in 2050 against WUP's 9.7 billion; 0.25° cells miss urban heat islands and intra-urban patterns; heat index above 35 °C marks frequent hot-humid days, not lethal heat; multi-model medians hide model spread; flood zones ignore defences and future flood change; WUP and GHSL use matching definitions but different vintages; exposure is not impact.