Reassessing the contribution of the Deccan Traps volcanism to the End Cretaceous Extinction.

During the Phanerozoic Eon the Earth has suffered a series of major extinction events almost all of which are considered to have been triggered by continental flood basalt emplacement episodes, which released vast amounts of toxic gasses into the atmosphere. The exception to this is the End Cretaceous Extinction, which is largely considered to have been triggered by the impact of an extra-terrestrial object into what is now the Yucatán Peninsula of Mexico. 

While this is a compelling story, it has faced a number of challenges from rival theories, the most notable of which is that there was a significant outpouring of flood basalts at the End of the Cretaceous, leading to the emplacement of the Deccan Traps Igneous Province in India, This, combined with the fact that we have now identified a number of other large impacts in the Phanerozoic rock record, none of which seem to have been associated with extinction events, presents a serious challenge to the Chicxulub Impact Theory. However, studies of the Deccan Traps Igneous Province have suggested that the majority of the basalt-emplacement happened slightly after the extinction event, implying that it cannot have been the cause. 

In a paper published in the journal GSA Bulletin on 5 November 2025, Vivek Kale of the Advanced Center for Water Resources Development and Management, Devdutt Upasani and Madhu Rajput of the Department of Geology at Fergusson College, Gauri Dole of the Department of Environmental Science at Savitribai Phule Pune University, and Shilpa Patil Pillai of the Department of Earth and Climate Science at the Indian Institute of Science Education and Research Pune, present a re-evaluation of the contribution of the Deccan Traps volcanism to the End Cretaceous Extinction, based upon new geochronological studies of the Deccan Traps Igneous Province.

The Deccan Traps Igneous Province covers about 50 000 km² of Western and Central India, and extends westward beneath the Arabian Sea, where it is thought to cover a further area of about 38 000 km². It was produced by shield-volcano-like eruptions, which produced a series of radially overstepping basalt formations. Studies carried out in the 1980s suggested that the onset of the Deccan Traps volcanism coincided with the End Cretaceous Extinction Event, leading the majority of the geological community to conclude that it could not be responsible for the event, and even some suggestions that the volcanism might have been caused in some way by the Chicxulub Impact.

Geographic sectors of the present-day exposures of Deccan volcanic deposits of central and western India (shaded green) on the backdrop of different cratonic blocks (in shades of pink and named in red) of the Indian Peninsular Shield. The named alignments of deep-crustal tectonic zones from this shield, with Precambrian heritage and late Mesozoic to Cenozoic reactivation are depicted with parallel hatching in their respective strike directions.  Locations of sampled offshore basaltic flows with affinities to the Reunion hotspot are shown with green dots in the offshore areas, along with key ocean floor features of the Arabian Sea off the western coast of India. The locations of Late Cretaceous magmatic activity in the Indian Peninsula (early magmatics) are shown as red stars. The maximum projected extent of the Deccan continental flood basalt is shown with a green dotted line. Kale et al. (2025).

However, much of this earlier work was based upon chemostratigraphic correlation between different parts of the Deccan Traps, something which is now considered unreliable as it has been demonstrated that the Traps are in fact made up of a series of subprovinces, the Western, Satpura, Central, Malwa, Mandla, and Saurashtra, each with their own distinct volcanic history. Thus the work carried out in the 1980s appears to have been valid for parts of the Western Subprovince, but not necessarily for any of the other subprovinces.

Kale et al. combined chemostratigraphic methods with palaeomagnetism and studies of key fossils from sedimentary beds interspaced with the volcanic layers, with the aim of understanding the timing and eruptive history of each subprovince of the Deccan Traps. To achieve a high level of confidence, they carried out extensive fieldwork, preparing more than eighty stratigraphic logs. 

Subprovinces of the Deccan Volcanic Province of India (in different shades of green) and the indicative locations of geochronological sections and fossil-bearing intertrappean sediments used in the age assignments of different stratigraphic units in this study. Abbreviation: KPgB, Cretaceous-Paleogene boundary. Kale et al. (2025).

Kale et al. recognised three phases of volcanic activity within the Deccan Traps deposits, phases which were consistently recovered using chemostratigraphic and palaeomagnetic methods. These were the Early Magmatic Phase, the Main Flood Basalt Eruptions, and the Late Volcanic Phase. Deposits associated with the Early Magmatic Phase outcrops in the Saurashtra Subprovince, along the Narmada Son Lineament Zone between the Bastar and Dharwar cratons, and comprises mixed mantle-derived volcanic sediments more than 67.0 million years old. The Late Volcanic Phase comprises intrusive volcanic material inserted into the main deposits after they had been emplaced. These outcrop in a small coastal strip around Mumbai, and have also been found offshore in the Laxmi Basin of the Arabian Sea. This Late Volcanic Phase material is about 63.0 million years old, making them coeval with the separation of the Seychelles from the Western Margin of the Indian Plate. There may also be some material associated with the Late Volcanic Phase at the top of the Amarkantak Group.

The Main phase therefore accounts for about 90% of the material which makes up the Deccan Traps Igneous Province. Furthermore, most of this material appears to have been produced within a period of less than 1 million years within Chron C29r, a period between two reversals of the Earth's magnetic field, which lasted from about 66.43 million years ago to about 65.8 million years ago (therefore spanning the current accepted boundary between the Cretaceous and the Tertiary, at 66.04 million years ago), and the following Chron C29n, which lasted from about 65.8 million years ago to about 64.745 million years ago.

The Western Subprovince hosts the thickest and (probably the) most continuous deposits of the Deccan Traps, as well as the largest single outcrop, at Kalsubai Peak, where a 1642 m continuous stack of basalts is exposed, and continues some way beneath the ground. These deposits are assigned to the Sahyadri Group, which is divided into the Wai, Lonavala and Kalsubai subgroups, with about 3000 m depth of basalt produced within chrons C29r and C29n. These deposits show few fossiliferous sedimentary layers, with scattered 'interflow horizons' (horizons marking time-gaps between episodes of basalt-deposition) of limited lateral extent, suggesting that gaps between eruptive episodes were localised and brief.

Kale et al.'s revised study places the boundary between chrons C29r and C29n at the base of the Mahabaleshwar Formation from the Wai Subgroup of the Sahyadri Group. This is marked by the widespread presence of giant-phenocryst basalt. The Purandargarh Formation, which underlies the Mahabaleshwar Formation, is calculated to date from the earliest part of the Danian Stage of the Palaeocene, i.e. immediately after the Cretaceous-Tertiary boundary. 

The Poladpur Lavas, which underlie the Purandargarh Formation, have been suggested to be of latest Cretaceous origin, on the basis of dates obtained from zircons, but Kale et al. reject this, on the basis that zircons can survive at very high temperatures and are often reworked within volcanic deposits, and also classify the Poladpur Lavas as earliest Danian. Based upon a sediment layer with key Mammalian fossils exposed at Naskal in Telangana State, Kale et al. estimate that the lowest 50 m of the Poladpur Lavas were erupted within 100 000 years of the Cretaceous-Tertiary boundary. The underlying Lonavala and Kalsubai subgroups, therefore, must have erupted entirely within the Latest Cretaceous. 

The Satpura Subprovince forms the second deepest sequence of the Deccan Traps, and has been divided into six formations, with two of these formations containing giant-phenocryst basalts, something which has led to the correlation of the entire subprovince with the Wei Subgroup of the Sahyadri Group. Kale et al. reject this analysis, assigning the entire sequence to the Maastrichtian (Latest Cretaceous) on the basis of fossil inclusions within sedimentary layers.

The Central Subprovince forms the northeastern part of the Deccan Plateau, with a diffuse boundary with the Western subprovince. The lavas of this the Mahur Formation, which form the base of this sequence, also contain a distinctive giant-phenocryst basalt layer, which had led to them being comparied to the Wei Subgroup, but Kale et al. again reject this, assigning this formation to the Maastrichtian on the basis of fossil inclusions. Instead, they place the Cretaceous-Tertiary boundary at the top of the Ajanta Formation, which overlies the Mahur Formation, on the basis of palynological evidence (fossil pollen). This places the final three formations of the Central Subprovince, the Chikhil, the Buldhana, and the Karanja, within the Early Danian.

The Malwa Subprovince sequence has also previously been assigned to the Wei Subgroup, although in this case because it contains a normal-reverse-normal palaeomagnetic sequence, which was thought to mark the Danian chrons C29r and C29n. However, a revised dating sequence suggests that this sequence contains some of the oldest rocks of the Deccan Traps, with only the final, Singachori Formation being Danian in age, while all the lower formations are Maastrichtian, based upon fossil evidence. 

The Mandla Subprovince has proven much harder to establish a chronological sequence for, but appears to contain both some of the oldest and some of the youngest lavas of the Deccan Traps. However, the lowermost Mandla Formation and unclassified underlying beds contain sedimentary layers which produce fossils of Maastrichtian age, while the higher beds of the Multai, Amarwara, Khamla/Khampla, and Kuleru formations contain magnetic reversals and fossils of Palaeocene origin.

Stratigraphic logs of the subprovinces of Deccan Volcanic Province of India depicting dominant morphological types with the approximate position of the 66.05 million-year-old Cretaceous-Palaeogene boundary (red dashed line). In the Sahyadri Group, the cumulative stratigraphic thickness is used, as there is a well-demonstrated southward and southeastward overstepping of the older formations by younger ones. All other logs are plotted for maximum thickness. The available palaeomagnetic orientations (dark, normal; grey, mixed; and white, reverse) are depicted. Abbreviations: BMBY, Bombay Subgroup; LNVL, Lonavala Subgroup; GPB, giant phenocryst basalt. Kale et al. (2025).

The Saurashtra Subprovince is the only part of the Deccan Traps where an iridium layer has been discovered within the Anjur Section, and used as an identifier for the Cretaceous-Tertiary boundary. However, the deposits which host this layer have been shown to belong to Magnetochron C28r, making them too young to be related to the End of the Cretaceous. Other deposits, from layers beneath those exposed at Anjur, have produced Dinosaur bones and nests, as well as other clearly Cretaceous fossils, indicating a Maastrichtian origin. These deposits are also cut through by a series of dykes which have yielded ages of between 66.06 and 62.4 million years. Given the lack of a clear chronological sequence for this group, Kale et al. do not attempt to calculate its full sequence, but for the sake of modelling assume that it contains 5% of the total volume of the Deccan Traps lavas, and that this can be divided equally between the Danian and the Maastrichtian.

Kale et al. also assume that 75% of the Deccan Traps basalts were erupted on land, with about 25% offshore. This gives a total volume of about 1.8 million km² of erupted lava (higher than any previous estimate) with the terrestrial deposits produced during the Maastrichtian and the Danian, while the offshore deposits are assumed to be entirely Danian in origin.

While Kale et al. produce a higher total volume estimate for the Deccan Traps Volcanism than previous studies, this is not a major increase, with most previous estimates being of a similar order of magnitude. However, by separating the Deccan Traps into a number of subprovinces and studying those individually, they do significantly re-estimate the amount of volcanism that occurred before the Cretaceous-Tertiary boundary.

Kale et al. estimate that early magmatism associated with the Deccan Traps were widely spaced across India in the Late Cretaceous, and probably associated with the Reunion Hotspot passing under part of the Indian Plate. Cainozoic Late Phase Magmatism occurred largely on the spreading western edge of the Indian Plate, and the associated shallow submarine shelf, with much of the material produced probably being better viewed as ocean/island basalt rather than continental flood basalt.

Between these two events, the Main Flood Basalt Eruptions produced about 1.2 million cubic kilometres of continental flood basalt (i.e. about 70% of the total volume of the Deccan Traps) within the last 300 000 years of the Cretaceous. This equates to an eruption rate of about 1000 km³ per year of basalt being produced during the Cretaceous portion of Chron C29r, falling to about 300 km³ per year during the Palaeocene portion of the chron. 

The lethal impacts of flood basalts themselves are rather limited, with most Animals able to out-walk all but the fastest lava flows. Ash clouds associated with such eruptions are more dangerous, potentially smothering plant life far from the source. However, the real threat comes from the gasses such events produce, with large volumes of carbon dioxide, carbon monoxide, gaseous sulphur, chlorine, fluorine, mercury, and other potent toxins into the atmosphere. 

Kale et al. estimate that the Main Flood Basalt Eruptions of the Deccan Traps would have produced over 6000 gigatonnes of carbon emissions, less than that produced by the Siberian Traps Basalts (associated with the End Permian Extinction) or the Central Atlantic Magmatic Province (associated with the End Triassic Extinction). However, around 4200 gigatonnes of this would have been produced within the last 300 000 years of the Maastrichtian (i.e. immediately before the Cretaceous-Tertiary boundary) with the remaining 1800 gigatonnes released over a longer period of time.

Fuethermore, Kale et al. conservatively estimate that 1300 gigatonnes of sulphur was released into the atmosphere during the Maastrichtian portion of Chron C29r, with about 200 gigatonnes being released during the Danian portion. 

High mercury levels have been observed around the world towards the end of the Maastrichtian, something which has previously been linked to the onset of the Deccan Traps Volcanism, although why mercury should peak at this point was unclear. Kale et al.'s results suggest that this spike did in fact coincide with the main phase of Deccan Traps volcanism, although they do not attempt to calculate the volumes of mercury produced. Estimates of the volumes of chlorine and fluorine produced by the eruptions were also beyond the scope of the study, although these are also likely to have been substantial.

Emissions of carbon dioxide from volcanic sources are typically enriched in the isotope carbon¹³ compared to other sources, something which is often used to track volcanic activity in the sediment record. Kale et al. examined the ratios of carbon¹³ in terminal Cretaceous sequences from the South Atlantic, North Atlantic, and Spain (areas which would have been far from India at the time), and found a significant spike in carbon¹³ levels immediately before the Cretaceous-Tertiary boundary, something they do not believe is coincidental. They also not fluctuations in other isotopic proxies which begin 600 000-800 000 years before the boundary, approximately the time frame for the onset of the Satpura, Malwa, and Mandla subprovince eruptions.

Based upon this, Kale et al. conclude that the majority of the Deccan Traps Flood Basalts were produced in a short period of time in the terminal Cretaceous, causing an environmental collapse which was the main driver of the End Cretaceous Extinction. The Chicxulub Impact could potentially have caused a single large mortality event against this backdrop, but is unlikely to have been the main cause of the extinction. The delayed recovery of the biosphere seen in the Early Danian is unlikely to have been caused by a 'nuclear winter' triggered by the impact, but instead probably relates to the ongoing, albeit reduced, volcanic emissions coming from the Deccan Traps at this time.

See also...

A possible crown-group Bird from the Late Cretaceous Lance Formation of Wyoming.

The origins and evolution of Mesozoic Birds are now well understood, but the emergence and development of the crown group Birds (a crown group contains all living members of a group, their most recent common ancestor, and everything descended from that ancestor) remains largely clouded in mystery. Most living Bird groups have a very poor fossil record, if they have a fossil record at all, despite Birds being the most diverse group of flying Vertebrates alive today, with more than 10 000 species. Fossils, where known, tend to be extremely fragmentary in nature, with most phylogenies of the group based entirely upon genetic data. This is particularly frustrating as the living Birds are the only group of Dinosaurs to have survived the End Cretaceous Extinction, something which has been taken to imply they had some quality missing in all other Avian and non-Avian Dinosaur groups. However, while crown group Birds are known to arisen before the End of the Cretaceous, they appear to have been at best a minor component of the Cretaceous Fauna, with few-or-no specimens found even in deposits which have produced numerous fossils of extinct Mesozoic Bird groups.

In a paper published in the journal BMC Ecology and Evolution on 9 February 2024, Chase Doran Brownstein of the Department of Ecology and Evolutionary Biology at Yale University and the Stamford Museum and Nature Center, describes a possible crown group Bird from the End Cretaceous Lance Formation of Wyoming.

The specimen, YPM VP 59473, comprises partial skeleton consisting of the complete left quadrate, portions of the skull roof a partially articulated, though very poorly preserved, cervical series, a fragment of the synsacrum, the left humerus, the articulated left radius and ulna, partial left tibiotarsus, and a partial pes. The material is largely disarticulated, but all of the bones are from a young juvenile and no duplicate bones are present, supporting the idea that they came from a single Animal.

Preservation of YPM VP 59473. The blocks containing all bones of the holotype (except for the humerus, tibiotarsus, synsacrum fragment, and large distal pedal phalanx) are shown under light microscopy (a), (e), (h) and with multiple x-ray views of the largest (b), (c), (d), second largest (f), (g), and smallest (i) blocks as rendered in VGStudio, showing the relative placement of bones in the matrix blocks. Brownstein (2024).

Despite the extremely fragmentary nature of the material, Brownstein feels confident in assigning the specimen to the Galloanserae, the group which includes the living Land and Water Fowl, and one of the three groups of Neornithine Birds thought to have diverged before the End of the Cretaceous, with the Palaeognaths and the Neoaves. This diagnosis is on the basis of the clear separation of the otic and squamosal capitula on the quadrate, the presence of a subcapitular tuberculum below the squamosal capitulum on the quadrate, the expansion of the ventral condyles and pterygoid condyle on the quadrate, the humeral head being dorsally offset from the rest of the proximal margin of the humerus, tricipital fossa being deeply excavated, and the dorsal tubercle of the humerus being large and offset from the rest of the proximal margin, all of which traits are typical of Galloanserine Birds, but absent in the various Mesozoic Avian stem groups.

Forelimb of YPM VP 59473. Humerus in (a) posterior, (b) anterior, (c) lateral, and (d) medial views. In (a) and (b), both CT scans and colour images are shown. Radius in (e) anterior, (f) posterior, (g) lateral, (h) medial, and (i) distal views. Ulna in (j) posterior, (k) anterior, (l) lateral, and (m) medial views. Brownstein (2024).

While the presence of a Galloanserine Bird in an End Cretaceous deposit is not unexpected, the presence of the specimen in the Lance Formation is significant in two ways. 

Firstly, because the deposit is from the Northern Hemisphere; phylogenetic studies of Birds based upon genetic data have found that the earliest diverging members of many groups have Southern Hemisphere distributions, which has led to speculation that the Neornithine Birds might have had a Southern Hemisphere origin, and the establishment of YPM VP 59473 adds to a growing body of data which contradicts that, suggesting that Neornithine Birds already had a global distribution in the Late Mesozoic. 

Secondly, unlike other deposits which have yielded Mesozoic Neornithine Birds, the fossils of the Lance Formation are thought to have been buried in situ, rather than being an accumulation deposit. This is important because the deposit has also produced toothed stem-Birds from at least four major clades, as well as Eudromaeosaurian, Alvarezsaurid, Troodontid, and potentially ‘four-winged’ Microraptorine Dinosaurs, all of which are thought to have been ecologically close to Birds. This is significant, as it suggests that the Neornithine Birds were not occupying some ecological niche which protected them from the impacts of the End Cretaceous Extinction, but instead were part of a community of ecologically similar Animals living in similar environments. This undermines the idea that Neornithine Birds were able to survive the End Cretaceous Extinction because they were in some way special, supporting the alternative hypothesis that they survived due to simple luck an important but sometimes overlooked factor in evolutionary biology.

The ecological and temporal origins of living Birds. Left side of the diagram shows the temporal and spatial range extensions and records of key small-bodied non-Avian Theropod clades found in the Lance Formation assemblage, and cladogram at right shows the major clades of stem and crown Birds that survive to or past the End  Cretaceous extinction, with ecologically relevant features that have been considered important to differential Avian survival through that event noted along branches. All clades shown on tree are unambiguously represented in the Lance Formation assemblage, except Neoaves and Paleognathae. Bird illustrations by John Gould. Brownstein (2024).

See also...

The Nadir Crater: A possible second End Cretaceous Impact Structure off the coast of West Africa.

High velocity impacts by large asteroids and comets are a known, but poorly understood, threat to life on Earth. Impactors with diameters of about 50 m (the estimated size of the object which caused the Tunguska explosion) are thought to hit the Earth roughly once every 900 years, and objects with diameters of 1 km or greater, likely to cause devastation on a regional or global scale, to impact the Earth roughly every million years. The largest impacts cause devastation on a global scale, with the best known of these being the Chicxulub impact 66 million years ago, an event thought to have caused the End Cretaceous Extinction Event. Despite the importance of such events, very little evidence of their having occurred is known, with only about 200 large impact craters known on Earth, of which only 15-20 are interpreted as having happened in a marine environment, despite 71% of the Earth's surface being covered by water. This lack of impact craters, and in particular well preserved impact craters, severely hampers our ability to understand the effects of the events which cause them.

In addition to this general lack of impact craters in the rock record, there is a proportional lack of double impact craters. About 15% of the asteroids in near-Earth orbits are thought to be binaries (i.e. pairs of gravitationally locked asteroids, whereas only 2-4% of known impact craters on Earth have ever been suggested as potentially being parts of pairs. Furthermore, most of these suggestions have either been disproven or are disputed. On Venus, where meteor impacts are generally recorded as 'dark splotches' rather than true craters, due to the much more destructive nature of the Venusian atmosphere, about 14% of impacts appear to be binary. Even allowing for a considerable degree of inaccuracy in these numbers, binary impacts seem to be significantly under-represented on Earth.

Another phenomenon seen elsewhere in the Solar System but largely absent on Earth is impact clustering, i.e. the formation of a series of impact craters over a relatively short period of time. This is typically caused by the breakup of a large object due to a gravitational encounter with a planet, followed by the debris from this breakup remaining on an orbit which intersects that of the planet, resulting in a series of impacts over a drawn out period of time. The only observed instance of such an event was the breakup of Comet Shoemaker-Levy 9 in 1992, which was followed by a series of impacts in 1994 as the debris from this breakup collided with Jupiter, but it is thought that the breakup of larger bodies can result in strings of impacts which could continue for several million years. Only a single such event has been recorded on Earth; a series of linked craters in the Ordovician, which is combined in a shift in the ratio of argon isotopes in the Earth's atmosphere at that time, and evidence for an increase in the amount of extraterrestrial dust raining down on the Earth at that time. Other proposals have been made for clusters of impacts in the Cretaceous and Eocene, but the evidence in support of such events is sparse. This is despite the Moon showing clear evidence of a number of periods of impact clustering, events which must surely also have affected the Earth.

In a paper published in the journal Science Advances on 17 August 2022, Uisdean Nicholson of the School of Energy, Geoscience, Infrastructure and Society at Heriot-Watt University, Veronica Bray of the Lunar and Planetary Laboratory at the University of Arizona, Sean Gulick of the Institute for Geophysics, Department of Geological Sciences and Center for Planetary Systems Habitability at the University of Texas at Austin, and Benedict Aduomahor, also of the School of Energy, Geoscience, Infrastructure and Society at Heriot-Watt University, announce the possible discovery of a second End-Cretaceous impact structure on the Guinea Plateau of the coast of West Africa, based upon evidence from two-dimensional seismic surveys.

The Guinea Plateau is an area of submerged continental crust extending southwest from the coasts of Guinea and Guinea Bissau for about 400 km, where it meets with the Guinea Fracture Zone, beyond which lies the deep ocean. The plateau can be divided roughly into two parts, an inner zone which is about 400 m deep across most of its extent, and an outer zone, known as the Guinea Terrace, which deepens gradually from about 400 m to about 1200 m. The Guinea Fracture Zone is home to a series of seamounts (inactive submarine volcanoes), including the Nadir Seamount.

Map and regional seismic sections showing location of Nadir Crater. (A) Regional bathymetry map of the Guinea Plateau and Guinea Terrace showing location of 2D seismic reflection and well data used in the study. JS, Jane Seamount; NS, Nadir Seamount; PS, Porter Seamount. The white dashed line shows the NE extent of high-amplitude discontinuous seismic facies at the top Maastrichtian interpreted as ejecta deposits and associated tsunami deposits. The north-east limit of this facies closely corresponds with the Maastrichtian shelf-slope break at the landward margin of the Guinea Terrace. Inset map shows a palaeogeographic reconstruction of the Atlantic near the end of the Cretaceous, about 66 Ma ago. Ch, Chicxulub Crater; Nd, Nadir Crater; Bo, Boltysh Crater. (B). Regional composite 2D seismic reflection profile extending from the GU-2B-1 well in the east to the deep Atlantic basin in the west, showing the structural and stratigraphic character of the Guinea Plateau and Guinea Terrace. (C) North-South seismic profile from the salt basin in the north to the Nadir Seamount, south of the Guinea Fracture Zone. Nicholson et al. (2022).

The Guinea Plateau formed by extensional rifting in the Triassic-Jurassic, as the opening of the Atlantic Ocean pulled the North American landmass away from the west coast of Africa. The plateau remained part of the passive (i.e. seismically inactive) margin of the Central Atlantic until about 100 million years ago, when the its southern margin underwent further rifting, as South America separated from Africa.

The crater which Nicholson et al. propose is located on the southwestern part of the Guinea Terrace, roughly 60 km to the north of the Guinea Fracture Zone and 100 km northwest of the Nadir Seamount. The crater is covered by about 400 m of sediment, and a water column about 900 m deep.

The stratigraphy of the area has been determined from boreholes and seismic surveys. Cretaceous deposits here can be divided into a Lower Cretaceous sequence, roughly 145-100 million years old, and an Upper Cretaceous sequence, 100-66 million years old, separated by an unconformity known as the Top Albian. The deposits making up the Upper Cretaceous sequence are largely undeformed, and extend laterally across the Guinea Terrace. These deposits can be further subdivided by a series of (seismically) reflective surfaces, numbered KU1-KU4, which are interpreted as representing major flooding events. The Upper Cretaceous sequence can be divided into a Lower Unit, which is about 300 m thick and contains a series of high amplitude reflective surfaces, and an Upper Unit, which is about 500 m thick and contains moderate-to-low amplitude reflective surfaces.

Seismic characteristics of the Nadir Crater. (A) Seabed depth map of crater showing seismic line locations and the mapped extent of the crater rim and damage zone.  (B)  West-east  seismic  section  (pre-stack  depth  migration  –  depth  domain)  across  the  crater,  highlighting  the  crater  morphology  and  damage  zone,  and  the  extent  of  subsurface deformation. (C) Detailed seismic stratigraphic and structural elements of the crater. KP, Cretaceous-Paleogene sequence (KP1 equivalent to Top Maastrichtian); KU, Upper Cretaceous seismic horizons. KU1 and KP1 'regionals' are schematic reconstructions of these seismic horizons before formation of the crater at the end of the Cretaceous and have been used to reconstruct a conceptual model of crater formation. (D) Southwest-northeast seismic section (pre-stack time migration – time domain) across the crater, showing crater morphology and seismic facies outside the crater, including high-amplitude seismic facies sitting above a roughly 100-ms-thick unit of chaotic reflections, interpreted to have formed as a result of seismic shaking following the impact event. Nicholson et al. (2022).

The end of the Cretaceous sequence here, identified as the Top Maastrichtian, is clearly identifiable, and well defined, but the age of other reflective surfaces in the sequence is less certain. The KU-1 reflective surface may mark the Top Turonian, making it about 90 million years old, and the high amplitude surfaces between this and the underlying Top Albian surface may represent black shale deposits, with elevated levels of organic carbon. The lower amplitude reflective surfaces above KU-1 probably represent marine shales and marls.

The Top Maastrichtian Surface lies at the top of another set of distinctive, high amplitude reflective surfaces, and is marked by an erosional unconformity on the southwest part on the Guinea Terrace, which is not present to the north and east. The reflective surfaces at this level extend laterally across the Guinea Plateau, although they are eroded away on both the continental slope and the inshore shallows. On the Guinea Terrace these surfaces overlie a chaotic seismic surface about 100 m deep.

Overlying the Top Maastrichtian, the Cainozoic deposits begin with a thick sequence of carbonates and clastic sediments, although these are considerably thinner on the Guinea Terrace than elsewhere on main Guinea Plateau. Precise dating of these sediments is again difficult, though sediments younger than about 23 million years old are apparently absent on the Guinea Terrace; strat here comprise a Palaeocene-Eocene sequence about 300 m thick, overlain by a series of Oligocene mass transport deposits (debris from submarine landslides).

The feature which Nicholson et al. identify as the Nadir Crater is a depression at least 8.5 km wide in the Top Maastrichtian. This can be seen where two separate seismic profiles cross one-another, suggesting a rounded or elliptical shape. Since there is no particular reason to believe that the intersection of the two seismic profiles lies at the centre of the depression, 8.5 km wide must be taken as a minimum measurement. The crater bed is 200 m below the 'seafloor' represented by the Top Maastrichtian, while the crater rim is raised 20-40 m above this level. There appears to be a slight uplift in the centre of the crater at the Top Maastrichtian level; this is more pronounced at the Top Albian level, where it rises 350 m above the surrounding terrain. On either side of this central uplift the strata below the crater floor and rime are intensely deformed, with a column of extensively folded and fractured material extending downwards 700 m from the Maastrichtian surface. A wider area of faulting extends 10-12 km from the crater in all directions, with all faults dipping towards the crater.

All of these phenomena are consistent with large impact craters elsewhere on Earth and other Solar System bodies. The rim is raised  above the surrounding terrain approximately one tenth as much as the crater floor is depressed beneath it. The crater is approximately one twentieth as deep as it is wide. The damage zone around the crater is approximately twice as wide as the crater itself. The crater has a terraces structure and a raised central peak.

The deformation beneath the crater is also consistent with the predictions for large craters of this sort. The central uplift seen in such craters is caused by the elastic rebound of shock waves passing through the rock, causing material beneath the crater to flow upwards, initially flowing freely the brecciating as the rock regains its brittleness. The deformations seen in the zone 10-12 km from the crater are consistent with these areas being subjected to a significant shockwave, then flowing back towards the crater to fill a partial void.

Based upon examination of other impact craters on Earth, the zone of uplift beneath a crater 8.5 km wide would be expected to extend downwards to between 630 and 780 m beneath the impact surface. That of the Nadir Crater extends down about 800 m, with some deformation extending downwards into underlying Jurassic Sediments. However, this does not necessarily imply a much larger crater; rather it may be a consequence of a bolide impacting soft, unconsoidated marine sediments. The Mjølnir Crater, in the Barents Sea off the coast of Norway, shows a similar pattern of deformation beneath the crater, although here there is less deformation outside the crater rim, as is the case at Chicxulub. This deeper pattern of damage at sites where large, high velocity objects impact shallow marine sediments may be caused by the sudden decrease in the porosity of the strata impacted, resulting greater brecciation, and the formation of hydrothermal systems which could potentially last for millions of years.

The Top Maastrichtian layer within the crater is overlain by a series of high amplitude reflective layers identified as KP-1 to KP-3, with KP-1 marking the boundary between the Maastrichtian and the Palaeocene. Between the Top Maastrichtian and KP-1 is a layer about 100 m thick and relatively transparent to seismic waves, which Nicholson et al. interpret as a possible suevite layer (i.e melt materia), similar to that seen as the Chicxulub Crater. Above this, between KP-1 and KP-2 is a unit made up of a series of low-amplitude reflective surfaces onlapping onto the inner crater walls, which Nicholson et al. interpret as reworked ejecta material. 

The crater is surrounded by a blanket of material made up of a series of high amplitude reflective surfaces, which Nicholson et al. also interpret as reworked ejecta, combined with material deposited by tsunamis triggered by the impact. Beneath this is a layer of chaotic material, which may represent older strata re-organised by the shock wave from the impact.

The Nadir Crater a deep inner crater with in inner peak and a flat outer zone, surrounded by a raised rim, features also seen in confirmed marine craters such as the Chesapeake Crater (off the east coast of North America), Lockne Crater (in northern Sweden), and Flynn Creek Crater (in Tennessee), strongly supporting the idea that this is an impact crater, although this hypothesis will only be confirmed by drilling into the structure.

There are other processes which can form structures resembling impact craters, such as the dissolution of salt deposits, escape of gas or fluids, volcanic activity, deformation due to tectonic stresses, or any combination of these. However, none of these possible alternatives appears consistent with the geology of the Guinea Terrace,

Salt diapers (domes formed when salt intrudes into overlying rocks) can collapse as the salt within them dissolves into the water column, leaving a circular depression surrounded by radial faults. However, such structures do not have uplifted central zones, nor would they be surrounded by a ring of high amplitude material of the type Nicholson et al. interpret as ejecta at the Nadir Crater. Furthermore, while there are salt diapers on the Guinea Terrace, these are in a zone far to the northwest of the Nadir Crater, with the closest diaper being over 250 km from the crater. 

Craters left by fluids escaping from mud layers as they are buried and then subjected to pressure, are common features on the seafloor, being particularly common on the continental margins, where large amounts of methane are generated. Most of these are a few metres to a few hundred metres across, but the largest can exceed 10 km, making ruling our such an origin for the Nadir Crater impossible on the basis of size alone. However, such structures are almost always found in clusters, and no structures similar to the Nadir Crater can be seen anywhere else on the Guinea Plateau, nor are there any other signs of gas or fluid escape in this region. Furthermore, craters formed by fluid escape do not typically produce crater rings, central uplift, or ejecta, all of which can be seen at the Nadir Crater. 

A variety of depressions, typically bounded by fault zones, can be formed when slip-strike deformation leads to areas of seafloor being pulled apart. However, such depressions are rarely circular, typically being more than twice as long as they are wide. Nor do such structures have rims or centrally uplifted zones. More importantly, the Nadir Crater is located on an area of the African continental margin where slip strike activity is thought to have ceased about 44 million years before it formed, making such an origin highly unlikely.

The most plausible alternative hypothesis for the origin of the Nadir Crater is volcanism. Phreatic explosions, which occur when hot magma comes into contact with water, can form maars (broad, low relief volcanic craters) similar in form to the Nadir Crater. Furthermore, there clearly was volcanic activity nearly contemporary to the formation of the crater in the area; the Nadir Seamont is only 100 km away from the Nadir Crater, and 7.4 million years younger, i.e. close enough to be part of the same volcanic complex, along with the nearby Grimaldi Seamounts. However, maar structures seldom exceed 2-3 km in diameter, and their three dimensional structure has a distinctive funnel-shape unlike anything revealed by the seismic surveys at the Nadir Crater. Furthermore, maars do not have a central uplift structure, making it highly unlikely that the Nadir Crater is a volcanic maar.

Next Nicholson et al. used a hydrocode model to simulate the impact of a 400 m wide object (the estimated size of the Nadir impactor) into soft sediments covered by a range of water depths. All of the models produced a crater between seven and nine kilometres in diameter, with elevated rims and significant central uplift. The amount of subsequent rim collapse varied with water depth, with impacts into deeper water showing more collapse. This lead to more infill into the crater, resulting in a flatter profile and a less prominent central uplift zone. At depths of greater than 1 km, the rim completely collapsed into the crater, creating a very flat profile. In all cases, regardless of depth, the impact resulted in a crater surrounded by a network of faults.

Numerical model results of iSALE hydrocode simulations of final crater architecture for different water depths (200, 500, 800, 1100, and 1500 m, as indicated on the figure) Models assume a 400 m diameter impactor with an impact angle of 90°. The water layer has been removed from each image to better highlight the final crater morphology. Model outcomes show total plastic strain on the left and deformed lithologies on the right - the gray unit represents the assumed Cenomanian-Turonian black shale deposits as a marker horizon. Nicholson et al. (2022).

All models show the development of a central uplift, although the extent of this varies with depth, with larger central uplift zones in impacts at greater depth, possibly representing the rocks rebounding more when the weight of the overlying water is removed by the flash-vaporization cause by the impact. These simulations showed material from as deep as 2 km below the surface being moves upwards, with material from up to 1 km deep reaching the subsurface within the central peak; based upon the available data Nicholson et al. estimate that material from as deep as 750 m beneath the Nadir Crater is present within its central peak.

Snapshots of numerical model results of an iSALE hydrocode simulation for 800 m target water depth. Considered this to be the most likely water depth for the impact, based on crater morphology. Model snapshots show transient crater formation and generation of a rim-wave tsunami after 3 seconds; rebound of central uplift and propagation of a rim-wave tsunami away from crater at 38 seconds; crater collapse, resurge, and central jet formation at 85 seconds; and further resurge at 245 seconds. Nicholson et al. (2022).

All of the simulated impacts produced zones of brecciation beneath the crater, with the depth to which this extended ranging from about 1 km up to about 2.5 km. Areas of damage were also present around all the simulated craters, with faults and fractures extending at least 5 km from the craters and reaching depths of at least 1 km.

The models are consistent with the observed features at the Nadir Crater, insofar as all of the craters forming at depths shallower than 1 km produced raised rims, uplifted central areas, terraces made up of debris, and areas of damage outside the crater. The models which produced craters most similar to the Nadir Crater had water depths of about 800 m.

Based upon the data gained by running these simulations, Nicholson et al. developed a conceptual model for the chain of events around the impact. It is assumed that the initial stratigraphy of the area comprised a series of flat, horizontal bedding plains, overlain by about 800 m of seawater. The initial impact would have formed a transient crater about 1 km deep, followed rapidly by the formation of the central uplifted zone, and the collapse of the initial crater rims. This would produce the current observed structure; a crater with a marked central uplift and observable rims surrounded by terraces of collapsed material, and a wider area of fallen ejecta, a hypothesis which they propose could by tested by drilling in the area, in order to confirm the actual nature of the deposits observed in the seismic profiles.

Conceptual model of the impact sequence at the Nadir impact site, based on seismic observations and analog models.  (A) At time zero, the impactor hits the water surface at a velocity of about 20 km per second, initiating a rim-wave tsunami in its wake. (B) Several seconds later, the transient crater forms, as the impactor and a substantial body of water are vaporised. Impactites (melt rock and breccias) line the transient cavity. Tsunami waves propagate away from the evacuated crater. Shock waves cause substantial damage below and around the impact site, and seismic waves propagate across the plateau; (C) major uplift (roughly 400 m relative to pre-impact regional) occurs in the 'rebound' crater modification phase, resulting in the formation of a raised crater peak; (D) radial collapse of the subsurface damage zone results in further modification of the crater, including the formation of terraces at the surface. Resurge of water transports substantial volume of ejecta and other sediment into the crater, deposited above the impactites. Nicholson et al. (2022).

Taking the data from the simulations, Nicholson et al. used the Earth Impacts Effect Program to assess the environmental damage caused by the event which formed the Nadir Crater. The initial impact would vaporise vast amounts of seawater and seafloor sediment and release an amount of energy roughly equivalent to 5000 megatons of TNT. This in turn would lead to the formation of a fireball over 5 km in diameter, an an air blast which would be travelling at about 470 km per hour at a distance of 50 km from the impact site. 

The event would release seismic waves equivalent to a Magnitude 7.0 Earthquake, fluidising seafloor sediments for hundreds of kilometres in all directions. Following fluidisation of the sediments they would likely be reworked extensively by the tsunami waves caused by the impact. The seismic waves would also be likely to trigger massive landslides along the Guinea Terrace.

The initial impact would cause an 'ejecta curtain', made up of water and sediment, over 2 km high, which would then collapse back onto the sea surface, triggering the formation of a tsunami wave over 500 m high, which would spread at about 400 m per second. As well as spreading out from the rim of the crater, this tsunami would spread inwards, forming a central water jet rising to over 2 km. This would then collapse back into the crater, forming a new tsunami, a process which might repeat itself several times. These repeated tsunamis would cause extensive reworking of sediments along the Guinea Terrace, at depths of up to 800 m. They would also cause extensive scouring of the West African coastline, and to a lesser extent that of South America, which, despite being about 1000 km away, would still be hit by waves about 5 m high. 

The impact would lead to large volumes of water entering the atmosphere, as well as black carbon dust derived from the black shale deposits laid down on the Guinea Plateau during the Cenomanian  and  Turonian.  The high temperatures and pressures unleashed by the impact would turn much of the organic material present in the area directly into methane, with global climatic implications, albeit short lived ones. The seismic shock wave could potentially release a lot more methane from gas hydrate deposits on the Guinea Plateau, and adjacent areas of the deep sea floor.

The impact which caused the Nadir Crater appears to have happened at, or very close to, the Cretaceous-Palaeocene boundary. This places the Nadir Impact very close to the Chicxulub Impact, chronologically speaking, raising the question as to whether the two events were connected. Potentially the Nadir and Chicxulub impactors might have been two parts of a binary object, two fragments of an object which broke up as it came within the Earth's gravitational field, two objects from a longer impact cluster, or two unrelated objects which happened to fall at about the same time. 

Impacts by binary asteroids have previously been suggested as the cause of other crater pairs, including the Lockne and Målingen craters in  Sweden, the East and West Clearwater craters in Quebec, and the Suvasvesi craters in Finland, although more refined dating has shown that neither the Clearwater and Suvasvesi craters are in fact binaries, and no there is no evidence in favour of the Lockne and Målingen craters being related, and therefore no evidence of a binary impact anywhere on Earth. 

Reconstructions show that the Nadir Crater site would have been about 5500 km from the Chicxulub Crater site at the end of the Cretaceous; significantly less than the 8000 km that separates them today, but still to far apart to have been caused by the impacts of two parts of a single object which broke up within the Earth's Roche limit (the minimum distance to which a large satellite can approach a larger body without tidal forces overcoming the internal gravity holding the satellite together), or within the Earth's atmosphere. 

This does not, however, rule out two fragments of a single object which was broken into fragments by an earlier encounter with the Earth's tidal field, then impacted the Earth on a subsequent encounter, in the way that Comet Shoemaker-Levy 9 was torn apart by Jupiter's gravitational field, then impacted Jupiter in several fragments on a subsequent pass several years later. There is some evidence of other large impacts around the End of the Cretaceous, most notably at Boltysh in Ukraine (currently dated to 65.4 million years ago, or 650 000 years younger than the Chixulub impact), as well as preserved fossil meteorites from Poland and the North Pacific. This could imply that the Earth was not just hit by a single large impact, but a protracted asteroid shower, lasting from months to one or two million years, with each impact triggering earthquakes and tsunamis, and contributing to rising atmospheric levels of greenhouse gasses and aerosols.

Estimates as to how often objects the size of the Nadir impactor hit the Earth vary from about once every 100 000 years to about once every 700 000 years. Thus even if it can be demonstrated that the Nadir Impactor fell within a million years of the Chicxulub impactor, it would be impossible to demonstrate that the two events were part of a single asteroid shower (Near Earth Asteroid (101995) 1999 RQ36 is roughly equivalent in size to the Nadir impactor, and is estimated to have a 1 in 1750 chance of hitting the Earth within the next few hundred years). However, if it can be demonstrated that the Chicxulub, Nadir, and Boltysh events all happened within a relatively short period of time, then it would present a much stronger case for the events being related, with important implications for our understanding of events at the End of the Cretaceous.

See also...

Follow Sciency Thoughts on Facebook.

Follow Sciency Thoughts on Twitter.

Dortoka vremiri: A new species of Dortokid Turtle from the Late Cretaceous of the Hațeg Basin, Romania.

The Cretaceous deposits of the Hațeg Basin of Romania are noted for the production of a rich diversity of endemic fossil Vertebrates, interpreted as an island biota, with members of many taxa exhibiting dwarfism compared to relatives elsewhere in the world. The fossils recovered from the Hațeg Basin include Fish, Amphibians, Squamates, Turtles, Crocodyliforms, Pterosaurs, non-Avian Dinosaurs, Birds and Mammals, with Turtles being among the most common. Despite this abundance, only two types of Turtle are known from the Hațeg, Kallokibotion bajazidi, a stem-group Turtle (i.e. a member of a group of Turtles that are not descended from the last common ancestor of all living Turtles), and an unnamed Dortokid.

 

In a paper published in the Journal of Systematic Palaeontology on 8 February 2022, Felix Augustin of the Institut für Geowissenschaften at Eberhard Karls Universität Tübingen, Zolt
án Csiki-Sava of the Faculty of Geology and Geophysics at the University of Bucharest, Andreas Matzke, also of the Institut für Geowissenschaften at Eberhard Karls Universität Tübingen, G
ábor Botfalvai of the Department of Palaeontology and Geology at the Hungarian Natural History Museum and the Department of Paleontology at Eötvös Lor
and University, and M
árton Rabi, again of the Institut für Geowissenschaften at Eberhard Karls Universität Tübingen, and of the Natural Sciences Collections at Martin-Luther University Halle-Wittenberg, formally describe the Hațeg Dortokid as a new species.

The Dortokidae are interpreted as an extinct, early-branching group of Pleurodire (side-necked) Turtles, likely to have been semi-aquatic and to have lived in freshwater environments. Sadly, no known example of the group has a preserved group, preventing the exact relationship of these Turtles to other Pleurodires being established with any confidence, but members of the group can be confidently identified by their micro-reticulate shell ornamentation. The earliest known members of the group date from the Early Cretaceous, while the latest date from the Early Eocene.

The new species is placed within the established genus Dortoka, which gives its name to the whole group, and given the specific name vremiri, in honour of the late M
aty
as Vremir, who prepared the specimen from which the species is named, and who previously worked upon Dortokid Turtles from the Hațeg Basin. The specimen is a partial skeleton with most of the carapace and the complete plastron preserved in contact, together with the exposed in situ right scapula and right pubis; the left-side counterparts of these girdle elements are potentially also preserved inside the shell.

 

Holotype carapace of Dortoka vremiri, LPB (FGGUB) R.2297, from the Upper Cretaceous Sînpetru Formation of the south-central Hațeg Basin, near Sânpetru. (A) Photograph and (B) drawing of the carapace in dorsal view, both to the same scale. Abbreviations: c, costal; M, marginal scale; n, neural; nu, nuchal; p, peripheral; PL, pleural scale; VE, vertebral scale. Augustin et al. (2022).

Dortoka vremiri is a medium-sized Dortokid Turtle which the first pair of costal plates meet behind the first neural plate, and the last pair of costal plates meet behind the eighth neural plate. This last pair of costal plates has a well defined suture, and prevents the eighth neural plate from making contact with the suprapygal bone. The specimen does not have a cervical plate. The first pair of pleural plates connect only to the first pair of costal plates, not the second. The second pair of pleural plates stops before making contact with the fifth pair of costal plates. The sulcus between the fourth and fifth vertebra plates is located towards the rear of the last neural plate, with the fifth vertebral having no contact with the costal plates. The entoplastron is at its greatest width towards its front. The Pectoral scales contact the entoplastron, and the extragulars are less than a third the size of the gulars.

Dortokid Turtles are known from the Hațeg Basin both before and after the End Cretaceous Extinction event, whereas Kallokibotion bajazidi, a species interpreted as having been more terrestrial in nature, disappears at the end of the Cretaceous, leaving no known close relatives. Thus, it is possible that the extinction favoured more aquatic forms living in this area, but also that all life here was wiped out, and that semi-aquatic species such as Dortokid Turtles were able to recolonise the Hațeg island from elsewhere.

A phylogenetic analysis of Dortokid Turtles carried out by Augustin et al. found that Dortoka vremiri was the closest known relative of Dortoka botanica, the species found in the post-Cretaceous Hațeg. This supports the idea that occupation of this area by Dortokid Turtles was continuous, and that therefore the End Cretaceous Extinction event was more lethal to terrestrial forms than semi-aquatic ones in the Hațeg Basin.

 

Phylogenetic relationships, as well as temporal and palaeogeographical position of Dortoka vremiri within the Dortokidae. (A) Strict consensus tree of the in-group phylogenetic analysis of the Dortokidae. There are two distinct lineages of derived Dortokids, a western lineage comprising Dortoka vasconica from the Late Cretaceous of Spain and France as well as an eastern European lineage comprising Dortoka vremiri from the Late Cretaceous of Romania and Dortoka botanica from the uppermost Paleocene of Romania. (B) Palaeogeographical map of the Late Cretaceous European Archipelago depicting the distribution of the different Dortokids. Abbreviations: Do, Dortoka vasconica from the Late Cretaceous of Laño and Armuña, Spain; Ds, Dortoka sp. (Dortoka vasconica?) from the Late Cretaceous of southern and south-eastern France; El, indeterminate Dortokid from the Early Cretaceous of El Castellar, Spain; Eo, Eodortoka morellana from the Early Cretaceous of Morella, Spain; Ha, Dortoka vremiri from the Late Cretaceous of the Hat¸eg Basin, Romania; Ih, indeterminate Dortokid from the Late Cretaceous of Ihark
ut, Hungary; Mu, indeterminate Dortokid from the Late Cretaceous of Muthmannsdorf, Austria; Ro, Dortoka botanica from the uppermost Paleocene of Rona, Romania; Sb, indeterminate Dortokid from the lower Eocene of the Şimleu Basin, Romania; Va, indeterminate Dortokid from the Early Cretaceous of Vallip
on, Spain. Augustin et al. (2022).

See also...

Online courses in Palaeontology. 

Follow Sciency Thoughts on Facebook.

Follow Sciency Thoughts on Twitter.