Naples sits between Vesuvius and Campi Flegrei. Yogyakarta looks up at Merapi, Mexico City at Popocatépetl, Taipei at the Tatun volcanoes. This is every volcano with signs of activity in the last 12,000 years, the number of people who live within 10, 30 and 100 km of each one, and when each last erupted.
The Smithsonian Institution's Global Volcanism Program keeps the world's reference list of volcanoes. Its Holocene list, the volcanoes with evidence of activity in roughly the last 12,000 years, holds 1,215 of them, from Antarctica to the Arctic and from the middle of the Pacific to the middle of Africa. For each one it also gives the number of people living within 5, 10, 30 and 100 km. Put those numbers on a map and the Ring of Fire around the Pacific becomes a ring of towns and cities: Java, Luzon, Honshu, the highlands of Mexico and Central America, and, on the far side of the world, the Bay of Naples.
Each circle on the map below is a volcano, sized by the number of people within 30 km (you can switch to 10 or 100 km) and coloured by when it last erupted. Click one to draw its 10, 30 and 100 km rings and see who lives inside them. Click a country for its totals, or search for a volcano or a country.
Volcanoes, last eruptions and eruption counts: Global Volcanism Program, Volcanoes of the World 5.3.5 (Smithsonian Institution). Populations within 5, 10, 30 and 100 km: the GVP's own figures, from its 5.1.7 list (a fixed snapshot, not current counts). Plate boundaries: Bird (2003), via Hugo Ahlenius, Nordpil (ODC-By). Countries and cities: Natural Earth. Base map tiles © Esri.
Adding up the GVP's figures for every volcano gives 56.4 million people within 10 km, 273.5 million within 30 km and 2.5 billion within 100 km. The last number is far too big, because many people live near more than one volcano and are counted again for each. Naples is the classic case: Vesuvius and Campi Flegrei are 24 km apart, and most of the city lies within 30 km of both. West Java, with a dozen volcanoes in a cluster, is counted many times over.
To count each person once, I kept only volcanoes whose circles do not overlap, choosing the most crowded first. Those circles hold different people, so their total is a firm lower bound. The true number lies between the two: at least 52.9 million people within 10 km of a Holocene volcano (and at most 56.4 million), at least 147.5 million within 30 km (at most 273.5 million), and at least 532 million within 100 km. Within 10 km the two bounds almost meet; at 100 km the lower bound is the useful one, and studies that count people on a population grid, such as Brown and others (2015), put the number within 100 km at roughly 800 million.
People living near a Holocene volcano
Solid bar: people within each distance of at least one volcano, counting each person once (a lower bound, from volcanoes whose circles do not overlap). Pale bar: the simple sum over all 1,209 volcanoes with population figures, which counts some people several times (an upper bound). Hover for numbers.
Most volcanoes are not crowded. 357 of the 1,209 volcanoes with figures (30%) have nobody at all within 10 km, and the median volcano has 4,732 people within 30 km. The crowds are concentrated: the 20 most crowded volcanoes hold 27% of the 30 km total, and 76 volcanoes have a million or more people within 30 km.
Indonesia has 101 Holocene volcanoes and at least 41.8 million people within 30 km of one, 28% of the world's lower-bound total; within 100 km, at least 120 million, much of the population of Java. Mexico follows with at least 18.7 million within 30 km, then Japan (10.1 million), the Philippines (9.7 million) and Italy (8.2 million), where Naples and Rome both sit beside volcanoes. Taiwan has only 2 Holocene volcanoes, but one of them, the Tatun Volcanic Group, rises on the northern edge of Taipei.
Countries with most people near a volcano
People within the chosen distance of the country's Holocene volcanoes: solid, each person counted once (lower bound); pale, the simple sum (upper bound). A volcano on a border counts for both countries, and circles can cross borders. Click a bar to see the country on the map.
Counting volcanoes gives a different list. The United States has the most, 165, mostly in Alaska and the Cascades, followed by Japan (118) and Indonesia (101), but few Americans live close to one: at least 549,000 within 30 km. Russia, Chile and Iceland have many volcanoes and few neighbours. Central America is the opposite: El Salvador and Guatemala are small countries built around their volcanoes, with at least 5.0 million and 5.3 million people within 30 km.
By people within 30 km, the most crowded volcano in the list is the Tatun Volcanic Group above Taipei (6.7 million), then Michoacán-Guanajuato in Mexico (5.8 million), Tangkuban Parahu above Bandung (5.7 million), Penanggungan near Surabaya (4.6 million) and Merapi above Yogyakarta (4.3 million). Vesuvius has 3.9 million people within 30 km and Campi Flegrei 3.0 million; within 10 km, Campi Flegrei has 2.2 million people living on and around the caldera itself.
The 20 most crowded volcanoes
People within the chosen distance, coloured by last eruption. Click a bar to see the volcano and its rings on the map.
The ranking depends on the distance. Within 10 km, the top places go to volcanic fields that underlie whole cities: Michoacán-Guanajuato (5.8 million), the Tatun group (5.1 million), Campi Flegrei (2.2 million) and Ilopango beside San Salvador (2.0 million). Within 100 km, Java takes over: Gede-Pangrango and Salak, both within sight of Jakarta, each have more than 38 million people within 100 km, and Hakone and Mount Fuji bring in Tokyo and Yokohama. A note of care: for 16 volcanoes, most of them large volcanic fields, the GVP gives the same figure at 10 and 30 km (Michoacán-Guanajuato has 5.8 million at every distance), which suggests the count covers the whole field rather than circles round one point.
At least 52.9 million people live within 10 km of a volcano that has erupted in the last 12,000 years.
A volcano that last erupted 5,000 years ago is not the same threat as one that erupted last year. 231 volcanoes have erupted since 2000, and 21 of them have a million or more people within 30 km. Top of that list is Tangkuban Parahu, which last erupted in 2019 with 5.7 million people within 30 km, then Merapi (4.3 million), Slamet (2.9 million), Pacaya near Guatemala City (2.5 million) and Taal, south of Manila (2.4 million). In all, at least 84.4 million people live within 30 km of a volcano that has erupted since 1900.
Crowded volcanoes that have erupted since 2000
The 20 volcanoes with an eruption in 2000 or later with most people within 30 km. Click a bar to see the volcano on the map.
People within 30 km, by when the volcano last erupted
Solid: each person counted once within each group (lower bound); pale: simple sum. Volcanoes with no dated eruption have Holocene evidence such as fresh lava flows but no eruption the GVP can date.
The eruption list runs to 2026. Eruptions that began in 2025 or 2026 came from 30 volcanoes with 5.1 million people within 30 km between them, led by Mayon in the Philippines (1.2 million) and Poás in Costa Rica (951,000). And 366 volcanoes have no dated eruption at all: they are on the list because of young lava flows or other Holocene evidence, not because anyone saw them erupt.
The GVP's eruption list holds 9,918 confirmed eruptions since the start of the Holocene. Counted by century, they climb steeply: 337 in the 1600s, 499 in the 1700s, 1,502 in the 1800s and 2,957 in the 1900s. In the 27 years from 2000 to 2026 there have already been 947, about 35 a year against 30 a year in the twentieth century.
Confirmed eruptions per century, 1 CE to 2026
Eruptions by century of start. Top: all confirmed eruptions, including those of unknown size. Bottom: eruptions of VEI 4 or more (large explosive eruptions, the size of Eyjafjallajökull in 2010 or bigger), on their own scale. The last bar covers only 2000 to 2026. Hover for numbers.
This is a record of watching, not of the Earth speeding up. Small eruptions on remote volcanoes go unrecorded unless someone is there to see them or a satellite to detect them, and 64% of all eruptions in the list are dated by reported observation. Large eruptions leave deposits that geologists can find and date long afterwards, so their record is more complete. The comparison shows it: from the 1600s to the 1900s the number of recorded eruptions rose 8.8 times, but the number of eruptions of VEI 4 or more only 2.0 times, and even that rise is likely to be mostly better reporting. 193 volcanoes have no eruption in the list before 1800.
Volcano deaths are rare but concentrated. NOAA's Significant Volcanic Eruptions Database lists 442 eruptions with deaths recorded, from 141 BCE to 2024, with 286,065 deaths between them. The deadliest are Tambora in 1815 (60,000 deaths in this database, mostly from the famine that followed), Krakatau in 1883 (36,417, mostly from tsunamis), Ilopango in El Salvador in about 450 CE (30,000, the date and toll NOAA gives), Mont Pelée in 1902, which destroyed the town of Saint-Pierre on Martinique (28,000), and Nevado del Ruiz in 1985, whose mudflows buried the town of Armero (23,080).
The deadliest eruptions in the NOAA database
Deaths as recorded by NOAA NCEI; figures for older eruptions are rough estimates.
Volcanoes and eruptions. The volcano list is the GVP's Holocene list from Volcanoes of the World 5.3.5 (1,215 volcanoes), and the eruption list is its list of confirmed Holocene eruptions from the same version (9,918 eruptions). Both were downloaded from the GVP in 2026; I used copies posted to public GitHub repositories, whose headers give the version and download date. A volcano's last eruption is the GVP's "last known eruption": for a long eruption that is when it ended, or the present year if it is still going on.
Populations. Current downloads of the GVP list no longer include population, so the populations within 5, 10, 30 and 100 km come from a copy of the list from version 5.1.7 that still has them, matched by volcano number. 6 volcanoes added since then have no figures. The same population figures appear unchanged in a GVP export from 2020, so they are a fixed snapshot of a gridded population estimate, not a census, and they predate a decade of growth in many of these places. The export does not say which population grid or year they come from. For 16 volcanoes, most of them large volcanic fields, the figure is the same at 10 and 30 km, and I flag those.
Counting each person once. Summing the figures over volcanoes counts some people several times. For a lower bound I sort volcanoes by population and keep each one whose circle does not overlap a circle already kept; the people inside non-overlapping circles are all different people, so their sum cannot be too high. It can be too low, because it leaves out everyone who lives only near a volcano that was skipped. The truth is between the bounds. Countries use the same method on their own volcanoes; a volcano on a border, such as those between Chile and Argentina, counts for both, and its circles take in people on both sides.
What this is not. Living within 30 km of a volcano is not the same as being at risk: hazards depend on the volcano, the type of eruption, wind, valleys, rivers and warning systems, and the GVP's list includes volcanoes that may never erupt again. These are distances, not hazard zones.