Life on Earth 66 million years ago may have faced two disasters back-to-back: new modelling suggests impact dust trapped a heat pulse that gave exposed land animals up to 17 times the thermal dose lethal to humans within hours, then helped block sunlight for years—giving burrowers and aquatic animals an early advantage, though shelter alone could not guarantee survival through the famine that followed.

The asteroid that struck near Chicxulub 66 million years ago did not present life with one uniform catastrophe. Around the Gulf of Mexico, the first destruction came from the impact itself: shock, heat, earthquakes, ejecta and enormous water movement. Far beyond the crater, material launched into the sky turned a regional collision into a planetary extinction.

A new study in the Journal of Geophysical Research: Biogeosciences argues that life then faced two atmospheric disasters in close succession. Fine silicate dust intensified an hours-long pulse of thermal radiation. When that pulse faded, the same dust helped block sunlight and sustain the colder, darker conditions that damaged food webs for years.

The sequence can explain why a burrow or enough water might have been useful in the first hour without being a ticket through the extinction. It also needs a firm boundary: this is a model-led reconstruction, not a direct temperature record. The geological evidence for global dust is strong, but direct evidence for widespread K-Pg fire is still concentrated in North America.

The heat did not have to come from the crater

Chicxulub excavated a crater roughly 200 kilometres wide and vaporised part of the asteroid and target rock. A hot plume expanded above the atmosphere. As it cooled, some of the rock condensed into molten droplets, or spherules, about 250 micrometres across. Those droplets followed ballistic paths around Earth and fell back at several kilometres per second.

The returning spherules slowed through the upper atmosphere at altitudes around 70 to 80 kilometres. Their kinetic energy became heat. Thermal radiation from that hot layer could reach the ground thousands of kilometres from Mexico, giving the impact a global pathway that did not depend on proximity to the crater.

How much heat reached the surface has been debated for decades. A 2009 Geology model showed that the first falling spherules would partly screen radiation from those behind them. Its lower estimate looked closer to an oven set to broil than to a global blowtorch: damaging to exposed animals and capable of lighting tinder, but probably insufficient to ignite thick wood everywhere.

The missing ingredient may be vapour that became dust

The new work, led by Purdue University planetary scientist Brandon C. Johnson, revisits material that did not condense into the first generation of spherules. Earlier impact-plume simulations suggested that about 44 per cent of the vaporised rock mass remained as vapour during the plume’s initial expansion.

Johnson and colleagues calculated how that vapour and the larger droplets would behave when they encountered the atmosphere. The components would slow differently and substantially separate. The vapour could then cool into far smaller silicate dust rather than remain attached to the spherules.

A geological clue comes from Tanis in North Dakota. The site preserves freshwater fish with impact spherules lodged in their gills, described in a 2019 PNAS paper. Above that spherule-bearing deposit is a fine layer carrying the impact’s iridium signature but no spherules. Its median grain size is about 2.88 micrometres, consistent with later fallout from a much finer cloud.

A similar separation appears in the Raton Basin across Colorado and New Mexico. Together with the plume model, these sites support the idea that abundant fine dust formed after the larger droplets and was distributed globally. More boundary sites are needed to test how uniform that dust really was.

An infrared blanket changes the surface dose

The team estimated that the plume ultimately produced about 1.6 quadrillion kilograms of fine dust. In the calculation, it had an optical depth of at least 620, making the upper layer effectively opaque to the thermal radiation trying to escape into space.

This did not create the original energy. The returning spherules supplied most of the heat considered in the model. Dust changed where that energy went. It acted like an infrared blanket, redirecting upward radiation and increasing the flux sent back toward the ground.

With the dust included, the surface pulse was about 3.5 times more intense than in the spherule-only calculation, according to the American Geophysical Union’s account of the paper. The authors note that this may still be conservative because their radiation calculation did not add heat carried by the fine dust itself, which initially held kinetic energy of its own.

There are countervailing uncertainties. A cloud or haze beneath the hot ejecta could reduce the radiation reaching the surface. The end-Cretaceous atmosphere was not identical to the modern atmosphere used in some earlier calculations. Ejecta would also have varied with direction, impact angle and distance from Chicxulub.

What 17 times the lethal dose actually means

The paper needed a biological scale for the modelled energy. It used a human reference from industrial-fire research: exposure to 10 kilowatts per square metre for 150 seconds is treated as 100 per cent lethal. When fine dust was included, the integrated Cretaceous exposure reached about 17 times that reference dose.

That comparison does not establish an exact lethal dose for a Tyrannosaurus, a bird, a mammal or an insect. The authors explicitly say the threshold for Cretaceous organisms cannot be known with certainty. Feathers, scales, body size, skin thickness, posture and access to shade or moisture would change the outcome.

The number communicates the scale of exposure predicted for an unprotected animal. Even relatively thick-skinned animals could have suffered severe burns or fatal heat stress. The model concerns thermal radiation arriving over minutes, not a claim that the surrounding air everywhere held one fixed temperature.

Fire thresholds require the same care. Grass, lichen and pine needles can ignite at fluxes around eight kilowatts per square metre for less than a minute, and the new pulse exceeded that level. Burning fine fuels could then spread into larger vegetation. Direct ignition of dry wood is harder; the paper says radiation alone may not clear the conservative threshold of 20 kilowatts per square metre for 20 minutes.

A burrow could beat the pulse but not the extinction

Infrared radiation does not penetrate deeply into soil or water. Animals underground, in cavities or beneath more than the shallowest water therefore had a plausible refuge. Roots and buried seeds had similar protection. The selectivity fits an older observation that small, sheltering organisms appear disproportionately among terrestrial survivors.

That pattern is suggestive, not a controlled experiment. Many burrowing and aquatic lineages still vanished, and many traits changed survival odds at once. Location, diet, reproductive speed, dormancy and the ability to use detritus would matter after the first physical insult.

A 2013 analysis of marine and freshwater extinction makes the distinction especially clear. Water could shield organisms from the heat pulse, yet marine ecosystems still suffered severe losses when darkness suppressed phytoplankton. Freshwater communities may have benefited from dormancy, groundwater refuges, variable habitats and food chains with more stored organic matter.

Shelter, in other words, solved an exposure problem. It did not manufacture food for the world that followed.

The same dust then switched from heater to sunshade

The apparent contradiction is a matter of timing and wavelength. During spherule re-entry, the fine layer stopped infrared energy from escaping upward and helped return it to the surface. Once the short-lived energy source was gone, dust remaining overhead intercepted incoming sunlight. The atmospheric blanket became a planetary sunshade.

A 2023 Nature Geoscience model, based on the measured size distribution of North Dakota boundary dust, estimated that micrometre-scale silicate particles could remain in the atmosphere for about 15 years. In that simulation, dust contributed to global-average cooling of as much as 15 degrees Celsius and nearly shut down photosynthesis for almost two years.

A separate 2026 Nature study of marine plankton modelled photosynthetically active radiation at roughly zero in the first year, 20 per cent of its former level in the second and 80 per cent in the third. Its global marine primary production fell by 99.5 per cent, while mean sea-surface temperature dropped from 26.4 to 12.3 degrees Celsius within three years.

Those numbers come from different models and should not be combined as if they were one measured timeline. They agree on the central ecological problem: surviving the initial heat left organisms inside food webs whose primary energy source had almost disappeared.

The two disasters overlapped

“Back-to-back” is useful shorthand, but the boundary between the crises was not clean. The strongest thermal pulse faded within roughly an hour. Fires could continue to add smoke and soot. Fine particles could injure lungs and gills. Cooling, darkness, acid rain and food loss then unfolded over different timescales.

Conditions were not globally identical either. Ejecta arrived differently by direction and distance. Oceans buffered temperature while their sunlit food webs collapsed. Land refuges reduced direct heating but varied in depth, moisture and access to stored food.

A previous SpaceDaily overview followed the older heat-pulse debate. The 2026 paper adds a specific reason why spherule-only models may have underestimated the surface flux: they did not include a geologically supported, optically thick dust layer above the returning droplets.

This sharpens the sequence without closing the case

Johnson and colleagues argue that heat and fire were primary killers of exposed terrestrial life. Other work gives more weight to the impact winter, and the contribution of Deccan Traps volcanism remains an active question. The new calculation is a strong case for restoring heat to the sequence, not permission to discard every slower mechanism.

The evidentiary gap concerns fire. The model supports global ignition of fine fuels, but the best geological wildfire record remains North American. Expanded K-Pg boundary deposits elsewhere could show whether the dust size distribution and charcoal record match the global prediction.

Laboratory tests using the model’s changing radiation pulse could also determine which fuels ignite under realistic durations. Better impact-plume calculations could track dust heat, atmospheric composition and geographical variation together.

For now, the two-stage picture explains why survival traits had to work on different clocks. A burrow, pond or river could turn away the first hour of heat. It could not guarantee passage through years of darkness, cold and broken food chains. Chicxulub may have selected its survivors twice: first by where they were when the sky became an oven, then by what they could eat after the light went out.