Snow is beautiful. It turns the world white and unblemished. Children love it for play, grown-ups love it for what it hides. Soon the snow will melt again, or worse, it will age like the world ages, lose its colour and become pockmarked with dirt. The beauty is temporary, here today, gone tomorrow. What was hidden will be revealed again – a memory come back to life, for better or worse.
Glaciers lack the beauty of snow. They carry the scars of age, the collected debris of the years. The surface can be so dirty that it looks like rubble, and needs a fall of snow to beautify. You need to dig down to find the ice. There may also be black layers embedded in the ice, evidence of nearby volcanic eruptions. Vatnajokul is full of those, caused by the frequent eruptions of Grimsvotn. But like the snow, the glacier hides what lies below. At the bottom is the rock, made barren by the scouring of the ice. Once the ice melts (as it will, one day), the scraped rock surfaces. It can take a long time before new soil builds up. We can still see the aftermath of the ice age in the scoured, soil-less lands of northern Canada, the west coast of Sweden, and many other places.
The rock which lies hidden below the ice contains the memory of a landscape that predates the glacier. There are events here from long ago. And not all that it covers is expected. This is the story of one such event, hidden below the ice of Greenland.
The Nares Strait separates the northwest coast of Greenland from Ellesmere Island. According to political ownership, Europe and Canada are only 30 km apart here. Geology, of course, puts both sides of the divide on the American plate. The straight line of the Nares Strait is typical for a fault line, an old one as there are no earthquakes here now. Indeed, this once was a strike-slip fault with some 70 km offset. The width suggests that there may also have been some extension. The fault was active around the opening of the Labrador Sea to the south, 60 million years ago. The precise date is not known, though.
Along the Nares Strait are some ice-free areas of land. The largest of these is Inglefield Land. It is 10 degrees north of the arctic circle, and the climate is not kind. There are only three months each year where the average temperature is above freezing, and in winter -20C is normal. The US has an airbase nearby, at Thule. It was the cold war at its coldest. Still, in sheltered places there is grass and other tundra vegetation.
There are people for whom this is home. High in the arctic, the area has a small population of Inughuit, polar Inuit who have been here for centuries. About 800 live in the region around Thule.
Archeologists have uncovered some of their ancient dwellings. The people here lived in houses for the nine months of winter, mainly in the area around Marshal Bay. (There is not enough snow for igloos.) The houses were partly dug into the ground, with an entrance tunnel, a stone-paved floor, walls made from cobbles and a roof build from stone and sod. In the summer the people would remove the roof to let the house dry out, and move away to do summerly things (hunting along the coast and in the eastern part of Inglefield Land), before returning in autumn to reassemble the dwelling. The dwellings that were investigated all had the roof removed: the people had left in spring as normal, clearly expecting to return in autumn but this time had not done so. Whether they had not returned or had build a new house nearby is not known. Archaeology cannot answer all questions: it only finds what was left, not what was taken.
In the 1860’s a few Inuit migrated north from Baffin Island and mixed with the local population. They re-introduced some forgotten technologies, such as bow-and-arrow, and kayaks. Westerners too came and they developed the Thule trading station, which later became the airbase.
The Inughuit were thus not the last to live here. Neither were they the first. The Inughuit had arrived in the area around 1300 AD (the people of that era are known as the Thule.) Before them came the Dorset people, seal and walrus hunters who lived here from 2000 BC to around 1100 AD. The longest Late Dorset longhouse known in Greenland was discovered in the Inglefield area, dating from before 1000 AD. Genetic studies have shown that the Inughuit are not their descendants. As an interesting aside, the Dorset people used iron blades on their harpoons, with the iron taken from the Cape York iron meteorite, found on the ice a little to the south. Stone tools were used to hack off iron fragments from the large meteorite. By about 1300 AD, the local Thule obtained some trade good from the Vikings to the south, perhaps traded during the summer walrus hunt. Different worlds mingled here.
Inglefield Mobile Belt
Inglefield Land lies north of Thule. It is an ice-free coastal region, a 20-km wide arctic desert with little precipitation – no more than 150 mm per year. In-land lies the Greenland ice sheet which spills over the mountains into large glaciers. The Hiawatha glacier which lies above Inglefield Land does not reach the coast (unlike the Humboldt glacier to its north), leaving Inglefield Land habitable for non-vegetarians. The area has been ice free since 6000 BC, when the ice barrier in the Nares Strait collapsed and warmer water from the south entered the Strait. After 1500 AD the ice began to advance again as the Little Ice Age took hold, and areas in Greenland further north than Inglefield became abandoned. Nowadays, as almost everywhere on Earth, the ice is retreating again, but the Hiawatha glacier remains larger than it was 2000 years ago.
The sparse vegetation grows on ancient rock. The bedrock at Inglefield Land is just under 2 billion years ago. It is an interesting region for its geology. In geology too, this was a place where worlds collided. We now recognize an east-west structure extending from here across Ellesmere Island. It is called the Inglefield Mobile Belt. The belt lines up with the edge of the old craton of Northern Canada, all the way to Alaska. Is this all part of the same structure, the old fault where North America and Siberia once came together and separated again? We have written about this in the White Christmas post: the time when the world almost ended. Later, the oldest parts of Alaska migrated 2000 km along this fault, to the west where it rotated into its present location (see the Wrangellia post). We don’t know for certain whether this is all the same structure. Sometimes alignments are real, and sometimes they are just coincidences.
The rocks here are far older than the Franklin event of White Christmas. The rock to the north has an age of 2.5 billion years. There was a later event which intruded magma and granite, dated to 1.9 billion years ago. This was apparently a plate collision which tilted layers vertically. A volcanic island arc may have been involved, coming in from the south. Behind it followed a continental block that is around 2.8 billion years old which now forms the area around Baffin Bay. The two blocks came together in the Inglefield Mobile Belt, with the old volcanic arc contributing to the southern part. The southern craton is now called the Rae craton. The Inglefield Mobile Belt is a young suture, in comparison, a 1.9-billion-year interloper still keeping the two oldies apart.
After the parts had come together, Inglefield had a fairly uneventful life. During the Cambrian it sank below sea level, and sediment covered the old rock. At some time a short fault formed in the northern part, running southwest to northeast. It runs parallel to the Nanes Straits and may have formed at the same time, an ‘also-ran’ fault that eventually failed. It is recognized because hydrothermal activity left a gold/copper mix on the surface.
The Hiawatha glacier lies east of the northern section of Inglefield Land, close to this gold-copper belt, and rises above it. It forms a semi-circular shape, attached to the Greenland ice sheet further in-land; a thin lobe of ice extends from Hiawatha to the northwest, flowing through an opening in the rock wall into a valley. The photo below shows the snow-covered ancient bed rock, with the thin lobe of ice and in the background the arc of the large glacier rising above it all. In the summer, pools of melt water develop on the far side of the Hiawatha glacier.
The Hiawatha glacier never attracted much attention. Ice is not scarce in Greenland, and the area was of no particular importance. Science looked elsewhere.
The first direct evidence that the glacier was well worth a closer look came during a visit in July 2016. A few hundred meters beyond the end point of the Hiawatha glacier lies a sandy flood plain, where a river deposits sediment from below the glacier. This floodplain had started to form after 2010, in response to the rapid local warming. During the one-day visit, sand was collected from the floodplain for analysis. The team found crystals of shocked quartz among the sand. Such crystals form in highly energetic events. A large impact seemed the most likely explanation, and the implication was that somewhere under the Hiawatha glacier there was an impact crater. It was time to remove the ice.
Of course, removing the ice is not trivial (though it is not beyond our technology – pumping CO2 into the atmosphere will eventually do the trick). To look below the ice, there are two faster options. One way is to drill through it. That requires luck, as the drilling only samples one location and that may miss the expected crater. The other is to use radar. Perhaps surprising, a radar signal at a frequency of tens of MHz (wavelengths of 10-20 meters) can penetrate ice. Not perfectly: imperfections in the ice will scatter some of the radar signal. Especially layers of volcanic ash can be problematic. This makes the method perhaps less suitable for Iceland, where glaciers are full of these layers, but Greenland has had no volcanic activity for the last tens of millions of years. The little Icelandic or Alaskan ash is too little to interfere with the radar. Internal boundaries in the ice also reflect the radar signal. This may for instance be at a boundary between stationary and flowing ice. A shear zone with broken ice will give lots of reflections.
The strongest boundary is at the bottom of the ice sheet, where the ice lies on what remains of the bedrock, after 3 million years of scraping. If enough of the radar signal makes it through to the bottom, a nice reflection signal tells you the depth of the ice. Tests have shown that such a radar can see through 3 kilometers of ice. Put the radar on an airplane and fly it across the entire ice sheet, and the radar will produce a map of the real, deeply buried surface of Greenland. The job is perfect for NASA which owns several research planes, used for various purposes. The mapping has been done over the past decades. The instrumentation has improved over that time. It now uses an ultra-wide band system developed in Kansas; it transmits 6 kW of power using 8 antennas, covers a frequency range of 150-600 MHz, and emits pulses that repeat at a frequency of 10 kHz. The transmitter is on the body of the plane; two of the three receivers are mounted on the wings. (The third receiver is the transmitter.) The aircraft flies 350 meters above the ice. The radar can measure the depth with an accuracy of better than 20 m.
Here is an example of such a radar profile, taken somewhere in Greenland. It shows the different layers where the ice has internal reflections, the chaos of a shear zone, and very clearly the grounding layer. The empty regions in the left part are cause by strong reflections higher up, leaving too little signal to get through. On the right there is another empty region; it is caused by a steep cliff on the bottom which reflects the radar signal away from the receiving aircraft.
The search for the shocked crystals had of course not been done on a hunch. It was still a closely guarded secret, but the radar mapping done over decades had shown signs of something extraordinary. A few months earlier, a more detailed radar survey of the Hiawatha glacier had confirmed the signs. The team had chartered the helicopter for a day to confirm what they already suspected. There were only 18 hours to find the confirmation, but that was enough.
There was indeed a meteor crater underneath the Hiawatha glacier, and it was not a small one. Neither was it a typical large, 10-km crater. It wasn’t large: it was huge. It turned out, the Hiawatha glacier was the crater. The semicircular shape of the glacier was caused by the crater rim. The glacier had grown inside a huge impact crater, filled it to the brim and grown up further. The glacier extended only about 1 kilometer outside of the crater – that single kilometer had hidden the secret. This was among the 25 largest impact craters known on Earth, measuring a tremendous 31 kilometers across. And no one knew.
The crater rim turned out to be 320 meter high, with a further 600 meter of ice thickness above this at the highest point of the Hiawatha glacier.
The bottom of the crater was rather flat, with a 50-meter tall rise in the central 8 kilometer. This shape is typical for such a large crater. Small craters (such as Meteor Crater in the US) have a bowl shape, with the deepest point at the centre. This what you get from an explosion in solid rock. Volcanic craters, if not filled with lava, have the same shape. But explosions in soft rock create a flat surface inside the crater. This is seen for instance in maars, where the explosion happens in wet rock. The explosion originally makes the same bowl shape, but the soft rock flows back and reshapes the bowl into a plate. Big impacts do this because they have so much energy that they partially melt the rock below. A very large impact can even cut through the crust, and expose the ductile layers below.
Why the central rise? It is known from large craters on the moon: they often have a central peak. It comes from rebound. The explosion pushes down the rock below, and after the explosion the rock rebounds and overshoots, finally freezing into position as a central mountain. (In an even larger impact, the rebounded mountain collapses again before it freezes, and in a second rebound forms a ring around the centre.) In Hiawatha, the central peak has been largely removed by erosion from the ice, and remains as a low plateau.
The crater rim is breached in two places. One is on the southeastern side, where two channels have merged. The second is on the northwest, is much smaller and is the source of the small lobe of ice flowing into the valley. Apart from this, the structure is remarkably well preserved. This may be because the bedrock in Greenland is so hard and so resistant to erosion (all erodable material has long gone). It might also be because the structure is young.
The impact happened within the Inglefield Mobile Belt. We know that the belt extends to Hiawatha: the sediment carried by the river from the northwestern glacier has the same composition as the material from the belt. The crater is symmetric and shows no deformation that would have happened during the formation of the belt. Therefore, it formed after the Mobile Belt had become immobile. This however is not a severe constraint. Very little deformation has happened here since. Even the opening of the Nares Strait might not have affected the area of the crater.
However, the crater is not as deep as it must have been originally: when it formed, it would have been 800 meters deep. That suggests that the rim and central peak have been significantly eroded. Erosion rates for Greenland are not well known: it could have taken anywhere from 50,000 to 50 million years.
Dating the crater is hard enough. Can we date the ice? Ideally, that would be done by extracting an ice core and counting the annual layers. The desert-like climate ensures that the annual layers are thin. At 10 cm precipitation per year, a kilometer of ice can build up in 10,000 years. In reality the layers will be much thinner than the annual precipitation rate, because of evaporation and compression. Thus, the ice could be as old as the ice age.
About 100 meters above the bed rock, the radar mapping shows a layer which reflects a lot of the radar signal. This has been seen elsewhere in Greenland. Where it reaches the surface, it has been shown to date to the Younger Dryas, the sudden cold and dry period when the ice age briefly came back, 13,000 years ago. The layer contains a lot of debris. If the layer inside Hiawatha is the same it means that the ice here is at least as old as the start of the holocene. Below this layer, the structure of the ice is complex. Ice from the late glacial period, (more than 20,000 years ago) which is present in the Greenland ice sheet, is not seen or is strongly deformed inside Hiawatha.
Surprisingly, the radar reflections show evidence for ground water underneath the glacier. This may be the source of the sediment in the river. Could the ground water come from the residual heat of the impact? That heat could linger for up to 100,000 years. However, this is speculative.
Based on the evidence from the crater and the ice, the original discoverers argued for a date within the past few million years (the pleistocene). This implies a highish erosion rate to bring the rim down by 500 meters.
The Younger Dryas
Could the crater be much younger than this? The crater lacks ice from the ice age. This has led to the suggestion that the crater did not exist during the ice age: it formed in the warmer period between the end of the ice age and the onset of the Younger Dryas. This was an extraordinary claim. Impacts of this size (requiring an impactor of 1.5-2 km across) may occur on Earth perhaps once every few million years. To have one only barely 10,000 years old would be a major event.
There had been other, earlier suggestions that a major impact had happened somewhere on Earth around this time. The original claim came from spherules and melt-glass found on sites in Pennsylvania and several other places. The sites had been dated to the start of the Younger Dryas. These locations also show enhanced iridium and platinum, as might come from a major impact. It quickly developed into a catastrophe theory, with the impact being the cause of the extinction of the large mammals in North America, the disappearance of the Clovis culture, and even the Younger Dryas itself. Enhanced iridium and platinum was also reported from a few sites in the southern hemisphere. Evidence for widespread forest burning in North America added to the evidence for the disruption. But a culprit had not been found: there was no impact crater of the right age and size.
The theory ran into trouble when later analysis showed that several of the original sites did not date to the Younger Dryas. A statistical analysis of the carbon dates from the sites confirmed that the various sites did not trace the same time. To get around the lack of a crater, the suggestion had been made that instead of a single event, it had been a meteor swarm with airbursts across the globe. To explain the variety of dates, a cometary swarm was proposed with impacts over a period of time. When a theory needs such complexities to explain missing evidence, it begins to look rather weak. What it needed was an actual crater. Could Hiawatha be that crater? It was in the right location for North America, and such an age would neatly explain the missing late glacial ice inside Hiawatha.
A clue was found in the crystals collected from Inglefield Land. Unshocked fragments showed compositions consistent with the local bedrock, not unexpectedly. But the shocked crystals and melt fragments show additions to this: they were enhanced in elements such as copper, chromium and gold, and with high rhodium. No local rocks show this and it likely came from the original impactor. Most impactors are made of silicate rocks, but about 10% are metallic, with high iron content. The impactor that made the Hiawatha impact crater appeared to have been such a rare metallic object. Platinum is often associated with iron meteorites, and the spike in platinum abundances at the start of the Younger Dryas had already led to the suggestion of an iron impactor. However, the fragments analyzed from Inglefield Land showed low platinum content, so the evidence did not quite come together.
There were some problems with the idea of such a young age for the Hiawatha crater. The impact would have ejected some 20 km3 of rock, forming a debris layer 200 meters thick around the crater, and 20 meters thick 30 kilometers away. The debris should have shown up in the Greenland ice cores, but nothing has been seen. Could the debris all have been carried away by the ice? Or was the impact under a shallow angle, directing the ejecta away from Greenland?
An association of the Hiawatha impact crater with the start of the Younger Dryas has run into too many problems: the low probability of an impact this size this recent, the lack of evidence in the ice cores and the evidence for 500 meters of erosion of the crater walls point at a much larger age. It may be possible to circumvent the problem by having the impactor hit the ice sheet. Models show that for a 2-km deep ice sheet, a much shallower crater wall and central peak is formed, because the ice absorbs the impact. There is also much less rock ejected from the site. However, this is double the current ice thickness (already amplified by the presence of the crater) and that would put us deep within the glacial maximum, long before the Younger Dryas. A search was made for impact grains in marine deposits associated with the Younger Dryas, but none were found. An association with the Younger Dryas seemed unlikely.
We learned more about the impact from the crystals. The structures of the crystals indicated that the crater cooled rapidly after the impact. Soon after the impact the crater seems to have filled with water. This is consistent with an impact in ice, but it could also be a water-rich, warmer era.
Many of the crystals that were analyzed contain organic material. In one study, organics were seen in 5 out of 6 grains. Carbon dating failed to give an age: this indicates that the organics is older (perhaps much older) than 40,000 years. The material has been identified as coming from burned pine trees. But pine trees have not grown here since the early pleistocene, 2.4 million years ago. Could the crater be this old, predating the ice age altogether, and having formed in a much warmer period? Could it even be the eocene, a warmer period before the world began to cool towards the ice ages?
The controversy around the age was finally resolved this year. Carbon dating is suitable for young dates but not old ones: C-14 decays too fast, leaving nothing after some 40,000 years. But there are other radio isotopes that decay much slower, and these can be used even for geological time scales. Argon is itself not radioactive. When a rock solidifies it has no argon: as a noble gas it escapes the melt. But potassium remains in the rock, and the isotope potassium-40 (40-K) decays (very slowly) into an isotope of argon. Because now it is solid rock, this argon cannot escape. By measuring the amount of argon and the amount of potassium-40, the ratio can be give the age of the sample. The half time is more than a billion years, so that the method works for very old rocks (but not for very young ones!). The second method uses uranium, which over very long times decays into lead. Again, by measuring what fraction of uranium has decayed, the age of the rock can be determined. Because of these methods, we can measure surprisingly accurate ages even for very old rock.
The U-Pb method showed that some of the grains had ages of 1.9 billion year. This dates them to the time when the Inglefield Mobile Belt formed. The map below shows the various ages of local rocks, which agree well with these U-Pb dates for the grains from Hiawatha. (No other rocks in North Greenland have this age – it is unique to the Inglefield Mobile Belt.) Clearly though, this is when the rock formed – not the crater. These grains had not been changed by the impact.
The other grains showed a component with uranium-lead ages around 58 million years, at which time there was a ‘reset’ event which had partially melted the grains.
The argon method indicated ages in excess of 25 million years. The two grains with the best result showed ages of 58 and 60 million years. This agrees very well with the U-Pb method, which shows that ‘something happened’ at this time. This ‘something’ was the impact.
The Hiawatha impact is therefore dated to 58.0+-0.5 million years ago. To put it in context, this was 8 million years after the demise of the dinosaurs, caused by another, much larger impact.
This new age shows that the impact crater has nothing to do with the Younger Dryas, the ice age, or even the pleistocene. It existed long before the ice came. Did the organic remains date from the time of the impact? That is very much plausible: at this time the area was abundantly vegetated at the time and may have had conifer forests. The organic remains became part of the sediment washed out on the river floodplain.
This new age solves the problems that the young ages had: the lack of an ejecta blanket or ice core trace, the significant erosion, and the organic remains. It also leaves the Younger Dryas impact hypothesis without a smoking gun and without its strongest evidence. Unless new evidence emerges, this hypothesis may have to be abandoned.
The asteroid caused a dramatic crater. It is perhaps the best surviving example on Earth of a large crater. However, the impact was too small to change the world’s climate or cause a worldwide extinction. That would have required a much larger impact as had happened 8 million years earlier. The effects of the Hiawatha impact would have been major across the local continent, but survivable further afield.
Were there such local effects? It is interesting that the crater is so close to the copper-gold belt which runs parallel to the rim. Did this fault open by the nearby impact? Copper and gold can be deposited by hydrothermal activity: the impact happened in a water-rich region, and the fault could have filled with hot water after the impact? This is of course speculative. The impact was also approximately when the Nares Strait opened. Did the impact break the final connection between Greenland and Ellesmere Island and triggered the activity of the Nares Strait fault? Nearby is an island, which was only recently revealed by the melting ice. It is now claimed by both Denmark and Canada. Perhaps, if this impact had not happened in this place, the two would not have separated and all of the region would now be Danish. It is just a thought.
The era when the crater formed was an exciting time in this part of the world. The volcanism in western Greenland related to the opening of the Labrador Sea occurred 60 million years ago. The paleocene/eocene boundary with its global warming event happened 56 million years ago, related to a flood basalt in eastern Greenland and the opening of the north Atlantic. The Hiawatha impact neatly occurred in between these events. What a time to be alive.
Dramatic results in science should always be looked at with suspicion, and with an eye to finding confirmation. This is how science works – checking, and checking again. In the end, the data will win. Sometimes, the extraordinary is true. Sometimes, there are changes to the interpretation. In this case, the Younger Dryas impact hypothesis lost out. But we are still left with a city-sized crater no one knew about, hidden in Greenland. The world can still surprise us.
Albert, March 2022
K. H. Kjær et al. A large impact crater beneath Hiawatha Glacier in northwest Greenland. Sci. Adv. 4, eaar8173 (2018) https://www.science.org/doi/10.1126/sciadv.aar8173
G. G. Kenny et al. A Late Paleocene age for Greenland’s Hiawatha impact structure. Sci. Adv. 8 (2022) https://www.science.org/doi/10.1126/sciadv.abm2434