Classification: Documented

Darvaza Crater: the burning gate and the history satellites rewrote

In the Karakum Desert, a gas-fed crater became the “Door to Hell.” Satellite imagery shows it existed by 1972, yet persistent fire probably began only in 1987-1988. This dossier follows its uncertain origin, a scientific descent in 2013, and the methane that can still escape as the flames fade.

Darvaza, Deserto de Karakum, Turcomenistão · See stories from this place and nearby

A light in the desert

At night in Turkmenistan’s flat Karakum Desert, a glow seems to rise from beneath the sand. Up close, it is a depression roughly 60-70 metres across and about 30 metres deep. Natural gas feeds flames from many fissures on its floor. The tourist name “Door to Hell” describes an appearance, not a supernatural cause. This is neither a volcano nor a meteorite impact crater.

The map point marks the publicly known site near Darvaza/Derweze. The sight tends to eclipse two more revealing questions: when did the ground collapse, and when did it start to burn? For decades, precise answers circulated without contemporary records. Satellite archives now give parts of that story firmer dates.

The holes in the 1971 story

The familiar version places Soviet drilling in 1971: a gas-filled cavity was struck, the rig and ground collapsed, and workers lit the escaping gas to remove a hazard, expecting a short fire. A 2011 NASA article illustrates how widely this account travelled; it is evidence of the story’s circulation, not an incident report. Other accounts place the collapse in 1963. The sources examined here do not include a contemporary technical record that settles the dispute.

In a 2014 National Geographic interview, explorer George Kourounis said local geologists had suggested a collapse in the 1960s and ignition only in the 1980s. He cautioned that he could not find old incident reports. His account alone cannot prove which drilling operation made the crater. It does show that the single-day 1971 tale was being challenged before the satellite study.

What the 1972 image proves

A 2026 Geophysical Research Letters study by Adriana Valverde and colleagues revisited the Landsat archive. The crater is already visible in an image from 25 October 1972. That establishes a documentary limit: the cavity existed by then. It cannot tell us whether it formed in 1963, 1971, or another earlier year. The date of a photograph is not necessarily the date of the event in it.

Landsat 1 lacked the thermal bands needed to establish the start of burning from that scene. Seeing the hole in 1972 does not establish that it was alight. Nor did the study recover a drilling log or identify the person who ignited it. The history becomes more useful because those gaps stay visible.

When fire appeared in the archive

To date the fire, the team compared Landsat 4 and 5 imagery from 1984-1988 using infrared and thermal bands. The last image without sustained flaring is dated 20 September 1987; the first thermal anomaly, 27 February 1988. Persistent ignition therefore fits between those observations. The precise night, the person responsible, and any short-lived earlier attempts remain unknown.

This window changes the legend: the cavity could have stood open for years or decades before burning. The earlier interval matters for climate because escaping gas without a flame would enter the atmosphere directly. Fire did not create the methane; it changed the fate of some of it. Figure 2 of the 2026 paper, linked under Explore, sets the two scenes side by side.

The mechanism below the rim

Darvaza is a collapse feature connected to a gas field. The study describes sedimentary layers, aquifers and routes that may let gas rise; the flow need not come from one visible pipe. Instead of a single uniform bonfire, many small flames move between vents on the floor. The position of the original borehole and the precise mechanical sequence of the collapse remain uncertain.

When methane reacts completely with oxygen, it produces mainly carbon dioxide and water. Real combustion may be incomplete, and some gas can evade the burning area through other fractures. Consequently, less light does not automatically mean less methane. The difference between measured heat and unburned gas is central to the recent evidence.

Seventeen minutes on the floor

In November 2013, George Kourounis descended on ropes with a heat-reflective suit, breathing protection and instruments. He described a jet-engine-like roar and numerous separate flames to National Geographic. His words are testimony about a visit, not proof of the Soviet accident. The descent required permission and a support team; the familiar lone-adventurer image obscures that work.

Filming was only one aim. He gathered soil samples to look for microorganisms adapted to the hot, nutrient-poor setting. The team reported a small number of bacterial types in crater samples that were absent from comparison samples outside. This is a local result reported by expedition participants; it does not establish extraterrestrial life or show that the entire crater is habitable. Their work turned Darvaza into a biological question as well as a geological one.

Methane the eye cannot see

The 2026 paper used hyperspectral sensors that can detect methane’s signature in reflected light. Between April 2020 and May 2025, researchers identified 44 plumes over the crater area, with estimated rates on detected occasions of roughly 0.6-3 tonnes per hour. They estimated 71 ± 21 thousand tonnes of methane over those five years. These are satellite-based estimates, not continuous measurement of every minute.

Wind, dust, ground reflectance and satellite timing affect what each instrument sees. An image without a plume does not prove that emissions stopped. Across the crater’s lifetime the authors extrapolate about 800 ± 250 thousand tonnes if it formed in 1971, or 900 ± 300 thousand if it formed in 1963. Both totals depend on an uncertain starting date and assumptions about past flow; they are not six decades of directly measured emissions. They also exclude methane that burned.

Trying to dim the fire

Turkmenistan’s government says it has reopened and drilled nearby wells to draw gas away from the crater. In 2025, its oil and gas agency said the glow could only be seen close up and burning intensity had fallen more than threefold. Capterio’s 2024 satellite analysis also recorded a sharp fall in flaring from 2023. Both accounts concern flame and heat; the official claim of reduced methane must be checked against measurements of that gas.

Valverde’s team found declining fire intensity, especially after summer 2023, but no clear correlation between this decline and the strength of measured methane plumes. Gas escaping through fractures without burning is one possible explanation. Dimming the spectacle without controlling the underground flow could leave an invisible source. The fire had not completely disappeared in the observations published in 2026; its later condition needs continued monitoring.

What each image leaves out

Tormod Sandtorv’s 2011 panorama records how the site looked then. A satellite view shared by NASA shows that the feature and its heat could be seen from above, although the accompanying 2011 text repeats a chronology written before the 2026 investigation. A 2022 ERT video interview preserves Kourounis’s account of the 2013 mission; it is not footage of the crater forming. Every item has its own date and evidentiary reach.

The research joins a long multispectral archive for the onset of fire, recent hyperspectral sensors for methane, and fire radiative power data for burning intensity. None automatically substitutes for the others. Open figure 2 to compare 1987 with 1988, then figure 3 for later plumes, and finally the 2011 photograph. The mystery changes shape when the evidence is read in the right order.

The boundary of knowledge

Established by the published analysis: the crater was visible by 25 October 1972; sustained flaring became detectable between September 1987 and February 1988; methane plumes were measured in 2020-2025. Well documented by testimony and imagery: the 2013 descent and the reduction in visible fire after 2023. Still open: the exact date and circumstances of collapse, who began the fire, the total flow in each past decade, and the eventual effectiveness of nearby wells.

Darvaza shows how a tourist icon can conceal an incomplete archive and a measurable climate problem. To move beyond the “gate,” compare the legend preserved in NASA’s older account with Valverde’s data and the state’s description of interventions. The larger question is not how long a flame lasts, but what continues to escape as its light fades.

Open in Google Maps