Globally, natural disasters are estimated to cause an average of over $300 billion in direct asset losses every year; this estimate increases to $520 billion when considering well-being (or consumption) losses. While each country faces its individual set of natural hazards, including cyclones, earthquakes, or wildfires, floods are among the leading threats to people’s livelihoods and affect development prospects worldwide. Especially in lower-income countries—where infrastructure systems, including drainage and flood protection, tend to be less developed—floods often cause unmitigated damage and suffering. Recent disastrous floods in countries as diverse as Nigeria, Bangladesh, Vietnam, the United States, and the United Kingdom illustrate that the threat is a global reality. Rare, major floods and smaller, frequent events alike can revert years of progress in development and poverty reduction. Understanding the scale and distribution of risks is crucial for devising targeted mitigation measures and allocating adequate resources.
While the threat is already substantial, several ongoing trends could result in significant increases in flood risks in coming years. For a high-concentration climate change scenario, estimates from 11 climate models converge to the conclusion that flood frequencies in Southeast Asia, East and Central Africa, and large parts of Latin America could increase substantially by 2100. Even in an optimistic climate change scenario (RCP 2.6), sea levels are estimated to rise up to 0.55 m by 2100, putting especially large coastal cities at risk. Land subsidence, often caused by unsustainable ground water extraction and drainage, has been shown to increase coastal flood risks at a rate four times faster than sea level rise.
Flood risks are also driven by socioeconomic change, as the number of people, assets, and value of economic activities increase over time3. By one estimate, in the absence of risk-mitigating measures, socioeconomic growth could result in the absolute damages from flooding to increase by a factor 20 by 2100. Considering the compounded effect of these drivers in the world’s 136 largest coastal cities, one study has shown that population and asset growth, climate change, and subsidence are likely to contribute to a drastic increase in global average flood losses, from $6 billion per year in 2005 to over $60 billion in 2050.
Recognizing the severe impacts of disasters on socioeconomic development, many flood exposure assessments have been conducted at local and national scales, often leveraging the recent availability of high-resolution flood, asset, and population maps, enabling increasingly accurate risk assessments. Yet, local studies have focused predominately on high-income countries like the European Union, United States, and Japan, not least due to data availability and the large economic values at risk,. While studies exist for developing countries, attention is focused on large economic centers like Jakarta, Dhaka, Dar es Salaam, Accra, and Ho Chi Minh City ,,,,,; few systematic assessments exist for the poorest countries and subregions, where floods are likely to have the most devastating impacts on livelihoods.
Overall, there is limited evidence on the global scale of flood exposure and how it relates to the incidence of poverty. Previous global flood risk assessments suffer from multiple limitations. By using global historical inventories of recorded flood events (e.g., from EM-DAT), studies have estimated exposure indicators at the country level . Yet, the lack of data on the spatial distribution and coincidence of flood risk and populations means that this approach does not allow a robust estimation of exposure headcounts17,. A more recent study documents the worrying trend of increasing flood exposure using satellite data for 2000 to 2018, though omits at-risk populations who remained unaffected during the study period.
Studies that use relatively coarse spatial resolution by current standards cannot account for the highly localized nature of flood risks, and thus drastically underestimate exposure3,5,,. One study projects that the global number of flood-exposed people will reach 1.3 billion by 205019, but our study shows that this threshold has already been exceeded by at least 15%. Other global studies have only focused on certain types of flood, rather than assessing the combined risks from fluvial floods (rivers exceeding their capacity due to excessive precipitation), pluvial floods (surface water build-up due to extended precipitation and insufficient drainage), and coastal floods (due to tidal or storm surges, or sea level rise)2,,,,,. For instance, a recent study conducted a detailed global assessment of the risk of sea level rise to the world’s coastal population, estimating that over 190 million people live in areas that could be inundated by sea level rise by 2100; but it does not consider inland flood risks. Other studies have only assessed risks for a subset of countries, falling short of full global coverage. Most importantly, none of the existing global studies consider the intersection between flood exposure and poverty incidence, which is a crucial indicator for people’s vulnerability, resilience, and ability to cope with and recover from floods1.
In this study, we close these gaps by conducting a high-resolution global exposure assessment for 189 countries, reaching within rounding errors of the entire world population. We assess people’s exposure to all current flood risks—that is, pluvial, fluvial, and coastal flooding. Flood data from Fathom-Global 2.0 are based on latest generation terrain and hydrographic models, while population density uses WorldPop 2020 maps calibrated on census and satellite data. The global coverage of these datasets enables an overlay analysis with 3 arc-seconds resolution (equivalent to about 90 x 90 meters at the equator), providing a more granular assessment than previous studies and eliminating the need for analytical assumptions besides the ones employed for producing the datasets. In addition, we leverage the latest edition of the World Bank’s Global Subnational Atlas of Poverty, which harmonizes household survey data and offers poverty estimates with global coverage and statistical representativeness at the subnational level. Full technical details on data and computational process are provided in the “Methods” section.
We find that 3 billion people, or 38% of the world population live in areas that would experience some level of inundation during a 1-in-100 year flood event. About 1.81 billion people, or 23% of the world population, are directly exposed to inundation depths of over 0.15 meters, which would pose significant risk to lives, especially of vulnerable population groups. The majority (1.24 billion) are located in South and East Asia, where China (395 million) and India (390 million) account for over one-third of global exposure. Low- and middle-income countries are home to 89% of the world’s flood-exposed people. Of the 170 million who face high flood risk and extreme poverty (living under $1.90 per day), 44% are in Sub-Saharan Africa. At least 780 million people face high flood risk, while living on less than $5.5 per day.
We conclude that the number of poor people living under severe flood risk is substantially higher than previously thought and concentrated in vulnerable regions that face compounding risks from climate change, sociopolitical instability, and resource constraints that hamper effective risk management. By offering global, yet disaggregated, insights on flood risk exposure and poverty incidence, this study highlights the scale of the needs and priority regions for flood risk mitigation measures that can safeguard livelihoods and prevent prolonged adverse impacts on development.