White Christmas


From all of Us to all of You: A happy Christmas and a Volcanic New Year!


Victoria Island

Our story begins and ends in the far north of Canada. The sea between northwestern Canada and northwestern Greenland is filled by a series of ancient islands. In winter the sea is deeply frozen, and the islands are a desolate, snow-covered wasteland. In summer the snow melts and the sparse tundra vegetation bursts into colour, but the sea may stay frozen. Many tried to find the fabled northwestern passage, a sea route through the archipelago connecting the Atlantic to the Pacific. Sir John Franklin was one of those. He was familiar with the area, having explored the coast twenty years before, in the 1820’s. But the sea ice enclosed his ships and it didn’t melt in summer, or the summer after. More than 130 men died in this disaster. Sixty years later, Roald Amundsen, the greatest of the polar explorers, finally managed to complete the route.

Victoria Island is the second largest island of the archipelago. It has the shape of a maple leaf – and this counts for something in Canada, far more than it being (reportedly) the 8th largest island in the world. There is a range of low mountains in the northwest of the island, with a plateau of basalt. The Kuujjua river has cut canyons in the basalt. Elsewhere there are patches of tundra, and a few shrub dwarf woods where the birch trees grow a staggering 50 cm tall. In spring and summer, the snow melts and the tundra comes to life with green mosses and lichens, and small plants with purple flowers.

Victoria Island is cold now. It was even colder in the past. During the ice age this area was covered by a thick glacier. When the glacier melted, the ground began to rise. There are now beaches 150 meters high above the current coast. Parallel grooves carved in the bedrock shows the scouring by the old glacier. Glacial debris and dropstones complete the evidence for the long winter of the ice age.

Christmas is the darkest time of mid-winter (well – at least it is in the north). There are months of cold still to come, but there is hope, or rather certainty, that the winter will pass and spring will come. But the ancient ice age seemed to last forever. Century after century, Christmas after Christmas, the cold remained and the ice thickened. It took tens of thousands of years for the melt to finally begin. The memory of that time is still all around on Victoria Island. But so is a far older memory, of a time when not just the north but the entire world was under the rule of the White Witch, beyond any hope of recovery.

Victoria Island is where the world ended. Here is where it began.


The rocks and stones of Victoria Island date to ancient Rodinia. This supercontinent formed around 1.1 billion years ago and existed until 700 million years ago; for much of that time it straddled the equator. Somewhere in the northern half of Rodinia was a basin, which collected sediment. This in-land sea sediment became (amongst others) Victoria Island.

Rodinia. Image from wikipedia

The sediment, now turned to solid, hard wearing rock, contains evidence of a major volcanic episode. The remains of sills, dikes and lavas are seen on the western side of Victoria Island, north of Prince Albert Sound. At least 13 separate sills intruded in the bedrock, with thicknesses between 5 and 100 meters, continuous over as long as 40 km. Many dikes exist, up to 40 meters across, and there are also remnants of thick surface flows. Little is left after so much time. At one time, lava as much as a kilometer thick spilled out, covering the future island in an ancient flood basalt. The sills have been dated to 716 million years ago.

Victoria Island, Canada

A flood basalt is also called a ‘large igneous province’. To avoid any risk of misspelling, it is often abbreviated as ‘LIP’, a term which can be used for either the completed lava sheets or for the eruption itself. A LIP is given a geographical name. This particular LIP covers the entire area of the District of Franklin (abolished in 1999) which covered the Canadian high arctic including Ellesmere, Baffin and Victoria Island. The District was named after Sir John Franklin, the unfortunate explorer who perished within this LIP. His name is now attached, perhaps a bit unfairly, both to the largest disaster of polar exploration, and to one of the biggest catastrophes the world has ever seen.

The Franklin Large Igneous Province has left traces in surprising places. The Coronation sills on the Canadian coast belong to it. The large igneous province around Irkutsk, southern Siberia (think Lake Baikal), which stretches from the Yenesei to Dovyren (‘Y’ and ‘D’ on the figure; ‘B’ is Baikal), has the same age, and is now considered as part of it. It So does the Kikiktat flood basalts (‘Ki’) on the North Slope of Alaska. The continents have been re-arranged, of course: at the time both Siberia and Alaska were rotated and the north slope of Alaska and the southern part of Siberia were attached to Canada. (How times have changed, even more so when you realize that at the time the area was located on the equator!)

The sills and dikes across the arctic formed about 720 million years ago. The oldest dates are around 723 million years and the youngest 716 millions years. Flood basalt eruptions typically last 1 to 5 million years – Franklin, being a large LIP, may have been at the upper end. Victoria Island has the younger dates, and may only have experienced the final part of the entire event.

The dike system of the Franklin Large Igneous Province. The positions are shown as they were at the time of the eruption. MH: Mount Harper; Co: Coronation sills; Ki: Kikiktat (now Alaska); B: Baikal (Siberia). From Ernst et al. Long-lived connection between southern Siberia and northern Laurentia in the Proterozoic. Nature Geosci 9, 464–469 (2016)

The two basalt areas on Victoria Island (see below) are known as the Natkusiat basalt: they are the only remaining surface flows or the Franklin LIP. Originally the lava sheets must have covered a far larger region, approaching the extent to the dike system. But they only survive on Victoria Island. The rest is either eroded away or still lies buried underneath younger deposits. The area of the Franklin flood basalt is at least 2.5 million km2. It covers 10% or more of the combined Siberia-Canada craton.

Franklin is one of 300 LIPs known in the past 3.5 billion years. It is quite large even for a LIP, at four times the size of the Deccan. For comparison, the Siberian Traps and the CAMP, the largest known, each cover between 4 and 7 million km2. An interesting accident is that Canada hosts another flood basalt of almost the same size: the MacKenzie LIP from 1.27 billion years ago. The Franklin LIP partially overlaps with the older MacKenzie, and the plume centre of the MacKenzie LIP may, in a twist of history suitable for a christmas cracker, have been located at Victoria Island. Some places have all the luck.

Victoria Island was not the centre of the Franklin LIP, however. The dikes fan out from a location further to the north, somewhere in the western arctic. The exact location is difficult to define. Dikes on the Canadian mainland point at a focal point northwest of Banks Island. But other dikes suggests that the centre was further east. Either way, this was in the heart of what remained of Rodinia.

Rodinia broke up (supercontinents always do) in stages between 900 and 700 million years ago. At the time of the Franklin eruption, many parts had already gone their own way. The Franklin eruption was located on the border between Laurentia and Siberia. The fact that the dikes crossed into Siberia shows that the two were still connected, rather than (as normally drawn) with Siberia at some distance. The last surviving part of Rodinia would break apart at this very location. Naturally, the Fanklin LIP is held responsible for this final demise of Rodinia.

This unwanted Christmas present initiated the separation of North America and Siberia. Afterwards, these two continents went their own way and Siberia rotated away, ready to leave. But it didn’t go far. In fact all three arctic continents, long before they became arctic, remained closely knit: Baltica, Siberia and North America were like partners who could neither live together nor apart. Think Shrek, Fiona and donkey. Or, perhaps more in tune with the season, the three wise men: they traveled together, but the story does not say how they got along. Wisdom can grate and wise men can argue too much. We value experts but don’t necessarily like them.


There is evidence that before the Franklin volcanism began, there had been uplift. Beginning between 10 and 20 million years before the eruptions, Victoria Island had been rising as part of a big dome. The doming is attributed to a mantle plume. The size of the dike system also suggests such a plume. A large mantle plume will spread out where it hits the lithosphere, and form a mushroom head with a size that can reach 1000 km in radius. This is indeed approximately the area of the Franklin LIP.

The doming may have been centred several hundred kilometers north of Victoria Island. The remnants that we have in Victoria Island are at the outer edge of a much larger structure. In the dance of the plates, the center has been lost. It may be hiding in an unexpected place – we don’t know the make-up and break-up of Rodinia well enough. It could be on the ancient cratons of China or Australia, both of which were in the general area. Or the old centre has been erased in a later mountain-building plate collision. We only have fragments of our past, and it can be hard to piece together the historical events from those fragments. Too much is gone.

From Williamson et al., 2016. The Franklin LIP on Victoria Island. Green: basalt; orange: older sandstone; yellow: sandstone of a flood plain just prior to the eruptions

Studies of the surviving lava flows on Victoria Island (green on the figure) shows that parts rest on the sand from a river flood plain (yellow). The river was flowing to the northwest, and brought sand which had been eroded from higher land to the southeast. This sand contains basaltic and other volcanic fragments, which apparently were still warm when deposited. There had been volcanic activity nearby before the basalt of Victoria Island was erupted. The same sand layer also shows disturbances: these indicate there had been frequent earthquakes.

The rise of the land shows that magma had collected below the crust. Long, approximately vertical dikes formed. Some dikes became deflected and formed large horizontal sills in the crust, tens of kilometers long. The sedimentary rock became riddled with these sills and dikes. The rock was rich in sulphur, dating from the basin in Rodinia where it formed. The magma in some of the sills, picked up this sulphur from the sediment.

The eruption itself consisted of three phases, over a long period of time. The earliest flows at Victoria island were in shallow water, perhaps lakes or ponds. They formed some local halyoclastites and a few pillow lavas. The water quickly disappeared: subsequent flows were rubbly but not wet. Each flow was between 1 and 10 meters in thickness. The top of these flows is weathered, showing there was considerable time between individual flows. The flow rates were not high and the flows are not wide. This suggests that this phase was fed by small, separate vents or fissures, each erupting at its own time.

In the second phase, volcaniclastics (think lahars) filled in a valley in the lava field, perhaps 50 meters deep. The valley may have been tectonic, grabens caused by magma intrusion, which would imply a long hiatus after the first phase. The first debris avalanche came from the edge of the valley and brought only debris from the lava field. The second one came from further away and was carried by water, indicating a river had started flowing over the lava field.

Now the real thing began, the third phase with much larger sheet flows, typical of a flood basalt. Three flows are recognized in the southern lava field of Victoria Island, with a combined thickness of 70 meters. The lower one is recognizable along a length of 25 kilometers. The northern lava field has four further sheets on top of these three, which are absent in the more deeply eroded southern field. The uppermost of these sheets is more than 150 meters tall. Each sheet consists of multiple flows: in the northern field, the five sheets together contain a minimum of 34 flows. Some of the individual flows can be traced over more than 30 kilometers.

From Williamson et al. 2013, https://publications.gc.ca/collections/collection_2013/rncan-nrcan/M44-2013-16-eng.pdf. The southern lava field. Kj is the underlying sand stone, V0 the first phae, C1 and C2 the two volcaniclastics and V1A-C the sheets flows.

The earliest lava was similar in composition to ocean crust with up to 10% continental crust melted into it. Later flows had less continental pollution, and as the melt became more and more voluminous, the lavas became more homogeneous. But this was deep within Rodinia: where did the oceanic crust come from? Isotopic ratios suggest that this oceanic crust was quite old (or ‘mature’). The heat from the plume melted the lithosphere below the continent, and perhaps the upper mantle. The lithosphere had retained material from a long-subducted oceanic plate, most likely a remnant of the ocean that had been lost when Rodinia was assembled. It is sobering to think that this 700-million year old lava field, deep in the Canadian arctic, was made from an ocean that was lost more than a billion years ago and which has seen the underside of Rodinia.

All this was a side show: the real centre of the activity was hundreds of kilometers away. The event left a scar even below the continent. Seismographic imaging has shown that the deep root of the Canadian craton comes to a sudden end underneath the northern shore of the Prince Albert Sound. The Franklin LIP split the craton top to bottom. Rodinia now came to an end as the continents on either side of the developing rift drifted away.


Life goes at its own pace. It can’t be rushed. The Franklin eruption occurred in a world that was poised for the future, but reluctant to enter it. Simple life (single cells) has existed since some 3.5 billion years ago. When photosynthesis developed, it led to the first major environmental crisis when its waste product (oxygen) turned out an indiscriminate killer. It almost ended life altogether. Eventually life developed ways to cope with, and even make use of, this highly poisonous molecule.

The great oxygenation happened 2 billion years ago. But after the initial spike, oxygen levels had dropped back again and a billion years followed of low oxygen (perhaps 1-2%). Life was hanging on but nutrient levels in the ocean were very low. Much of the sea was anoxic, and mainly supported sulphur-eating bacteria. This sulphur is still seen in the sills of Victoria Island. The climate was pleasant, with temperatures a few degrees higher than nowadays. The atmosphere was not, with very little oxygen, five to ten times higher CO2 than nowadays, and significant methane and nitrous oxide. But change was coming. Stromatolites, build by cyanobacteria, disappeared 1 billion years ago, perhaps because of grazing, from a new life form which had developed. Microfossils became more diverse. Their origin is not known but they may come from algae. Bacteria ruled the waves! 800 million years ago the first animals evolved. Don’t hold your breath (however tempting in that atmosphere): these were only the forerunners of the sponges. But it was sign of what was to come.

If you wonder how such a high level of greenhouse gasses could keep the Earth no warmer than 4C above current levels, the Sun was fainter in those days. And as Rodinia was breaking up, those CO2 levels were falling. Winter was coming. And what a winter it was. It even was given its own geological name: this was the time of the Cryogenian.

In the years before the Franklin eruption, the climate had become variable. Models suggest that global temperatures may have been a little cooler than today. There had already been some excursions to even cooler climates. One study finds evidence for a glacier in southwest Virginia, about 750 million years ago, at a time Virginia was tropical. This was likely a mountain glacier, and it has been suggested to be associated with rifting, perhaps akin to (but lower than) the Mountains of the Moon along the Africa rift valley which nowadays carries glaciers. This brief cold phase is now known as the Kaigas glaciation. A bit earlier, around 810 million years ago there was a longer phase of colder weather known as the ‘Bitter Springs’.

From MacDonald et al. 2010, Calibrating the Cryogenian, Science Vol 327, pp. 1241-1243. Mount Harper glacial indicators. Top: glacial dropstone; Middle: Striations caused by a moving glacier; Bottom: Soft sediment which has been folded underneath a grounded glacier.

Why had the climate been cooling, after the warmer days of early Rodinia? There are two reasons. The continents had all moved to tropical and subtropical latitudes. Land reflects more radiation than the sea, and with the continents on the equator where most of the sunlight falls, the Earth now reflected more and retained less solar heat. And as Rodinia began to split, the climate may have become wetter (supercontinents have dry interiors) and this would have increased weathering. Weathering of rocks removes CO2 from the atmosphere, so the CO2 content was going down a bit: there was less greenhouse warming. This is common when supercontinents break up.

However, this was no ice age. With the continents clustered around the equator, large glaciers could not form. The sea near the poles might freeze, but snow had nowhere to settle. There would be no white Christmas – and certainly not at tropical Victoria Island.

Not far from Victoria Island lies Canada’s Yukon territory – at least, it is close enough to be in the same country. Rocks of the time of the Franklin LIP are exposed here in various locations stretching to the Alaskan border. And these show unexpected features. Around Mount Harper, there are scratches on the rock, similar to those seen on the modern-day Victoria Island, left by recent ice. A layer of diactimite is seen: a layer of sediment containing unsorted fragments up to boulders in size. This is the kind of sediment that can come from a glacier. Debris and dropstones show evidence for a floating glacier. The conclusion was unavoidable: Mount Harper had been near the grounding line of a marine glacier.

This glacial deposit has been dated to 716.5 million years ago. Mount Harper lies within the area of the Franklin dikes, and therefore already was close to Victoria Island. The lava at Victoria Island has kept the direction of the magnetic field of the time. From that, we know this region was no more than 10 degrees from the equator. That Mount Harper glacier was floating on a tropical sea.

Evidence for this glaciation has been found around the world, in Africa, Australia and America. They all tell us the same story of sea-level glaciers in the tropics. The ice had reached everywhere and the sea had frozen from pole to pole and from shore to shore. No open sea remained anywhere on Earth, not even the most remote ocean. The long winter had begun. Victoria Island finally had its white Christmas.


We have long known that the Earth is living precariously. Modeling shows that there are three stable types of climate that the Earth can have. One is what we have, where temperatures are moderate, the seas are open, and snow and ice are found only at higher latitudes or not at all. There are excursions to more extreme temperatures, as there were during the ice age or from our attempt at global warming, but these are within the range of this type of temperate, liveable climate which we depend on.

The second climate type is that of the White Witch where the entire Earth is frozen. Open oceans absorb a lot of sunlight and turn it to warmth. Ice, on the other hand, reflects most of the sunlight. A frozen Earth would have a stable temperature far below freezing, all over the globe, kept that way by the reflecting ice.

There is a third stable climate but we don’t want to go there: evaporate the entire ocean and we get a steam atmosphere where the blanket of moisture keeps the temperature high enough (several hundred C) to keep it that way. Somehow the Earth has avoided both extremes and kept temperatures at the moderate level which life can live with, over a staggering 4 billion years. For much of that time, the Earth was an ocean planet with only small pockets of land. Oceans are excellent at providing stability. But over time, the land areas had grown and had reached continent size. Land creates climate instability: it has low heat capacity, and the temperature can fluctuate rapidly. And now all this land was in the tropics, and much of the sunlight falling on Earth fell here. Franklin erupted into this world of danger.

The temperate climate had been lost. In an ice age, the whiteness of the ice and snow keeps the temperatures low, and allows the glaciers to creep towards temperature latitudes. Models show that if sea ice were to reach 30 degrees latitude, the expansion becomes unstoppable. Temperatures plummet, and within 200 years the sea ice will have reached the equator. A few thousand years later, in those models the ice is several hundred meters thick across the entire world ocean. This had come to pass.

The global freeze had begun immediately after the Fanklin eruption of Victoria Island. Somehow, this flood basalt, among the largest continental flood basalts known, had tipped the Earth into a snowball. We still don’t quite know how the LIP did this – other, even larger LIPs never managed. There are some suggestions. The Earth had already been relatively cool, which helped. The Franklin eruption was quite sulphur rich because of the sedimentary rock through which it erupted. The sulphate in the atmosphere may have cooled the planet. The problem with this is that sulphate drops out of the atmosphere quite fast and if the eruptions lasts very long (centuries), it has to compete with the CO2 which the eruption also produces.

More likely is CO2 scrubbing. Volcanoes put CO2 into the atmosphere – flood basalts therefore often go together with a strong bout of global warming. But later, they can remove CO2 again. This is because basalt, when it weathers, takes up CO2. The suggestion is that this scrubbing had been much stronger than usual after the Franklin eruption, firstly because the basalt covered such a large area, and second because it erupted in the tropics with high rainfall and therefore high weathering. The result was that after an initial CO2 spike, levels began to drop – and reached values well below those of before the eruption. As CO2 fell, temperatures did too, sea ice expanded and snow began to settle on the fringes of Rodinia. When half the ocean was covered by ice, the planet’s fate was sealed.

Let it snow, let it snow

The ice catastrophe may have been an accident. The Franklin eruption had been going on, intermittently, for one million years or more. But now it was in its declining phase. Basalt was weathering and CO2 was going down. At that moment, in a final flash, Victoria Island erupted its sulphuric magma. This caused a sharp, temporary drop in temperatures. This drop was enough to push things over the edge, perhaps allowing glaciers to form in parts of Rodinia. It never recovered, and the volcanic winter became a permanent one.

And so the ice came, the snow fell and the world turned into a frozen, lifeless desert. The bleak winter would last for 50 million years.

The bleak mid-winter

With the sea frozen, the continents covered in snow and CO2 at record low levels, temperatures plummeted. The average temperature over the year quickly dropped to -40C, even at the equator. These temperatures are predicted by climate models, but have been confirmed by the study of oxygen isotopes in rocks of the time. The poles were far colder, with an average annual temperature of -80C and midwinter temperatures of -110C. It was like living on Mars. Summer temperatures were a balmy -20C at all latitudes. There are no seasons in the tropics, and the cold here was year round. The sea ice grew to a thickness of over 1 kilometer.

From space, the Earth would have appeared white. This is called a ‘Snowball Earth’ – for obvious reasons. Snowfall was sparse, as cold air contains very little moisture. The whole Earth was a desert. But much of modern Antarctica is such a cold desert, and still it has been able to grow enormously thick glaciers. Even a little snow can build up if it never melts.

You may wonder how there could have been any snow at all when the main source of moisture – sea water – had been locked away below the thick sea ice and evaporation of water had stopped. But ice could also sublimate into vapour, bypassing the state of liquid water. Sublimation happened mainly at the tropics, with the Sun directly overhead and no clouds to shield it. A reversed Hadley circulation channeled the moisture to the subtropics where it became snow. Glaciers could thicken at a rate of a kilometer in a million years. That rate was high enough for the glaciers to start moving, but the ice growth was 10 times less than that in our modern ice ages. Models show that the glaciers covered the entire supercontinent within 200,000 years, apart from a few coastal strips. Sometimes the glaciers would surge (very slowly) seaward; at other times there was a sudden retreat, leaving sand exposed. Meltwater would flow underneath the glaciers, just as it does in modern glaciers. The air was dusty, dry and cold. And so it continued, year after year after year, in this everlasting winter.

But during this winter, there was a slow change, imperceptibly at first but becoming larger as the season grew older and the ten million years of December turned into the ten million years of January, and then February. For this was a broken Earth. In our modern world, CO2 is kept at bay through weathering of rocks and through the carbonate cycle of the oceans. There is a balance between how much CO2 is produced by volcanoes , and how much is removed from the atmosphere. But this cycle had been broken. Volcanoes remained active, breaking through the ice in immense eruptions. But the CO2 they breathed out was no longer removed from the atmosphere. CO2 can dissolve into rain, but not into snow. There was no weathering of rocks and there was no open water. What was added, stayed added because of a lack of subtraction. And so CO2 began to multiply, from 1000 ppm (0.1 mbar) to 10,000 and then 100,000 ppm. It still wasn’t enough to fully counter the enormous reflectivity of the sea ice, and Winter’s rule remained undivided. But temperatures slowly went up from their early extremes. Eventually, March came and during the afternoon, for the first time in 50 million years, the temperature rose above freezing. There was hope of a Spring.

During those millions of years, the rising temperature had already changed the land. Glaciers were retreating and more of the land became ice free. Even at 10,000 ppm, the open land remained limited to the equator, but at 100,000 ppm, most of Rodinia had become bare. But the ocean was still a wide expanse of ice, thinner than before but still frozen solid. At 100,000ppm, the sea glacier was still 500 meters thick on average.

The ides of March

But change was in the air. Once CO2 had increased to 200,000ppm, tropical average annual temperatures were above freezing, sea ice at the equator began to melt from the top, and open water appeared. This water did something that had been missing for a very long time: it absorbed rather than reflected sunlight. When this open water lasted year-round, the winter reached its end. The sea warmed, and more open water appeared. The sea ice began to retreat towards the poles. The change was now unstoppable; the retreat accelerated and within only 2,000 years, even the poles had become ice free. Spring had arrived.

It was a catastrophe. Rain was now falling and CO2 was being scrubbed from the atmosphere. But half of all air was CO2, and the scrubbing took too long. Temperatures were rising, at first welcome but soon the rise became too much. A torrid greenhouse developed. The oceans melted, warmed, and kept warming. Water temperatures reached between 40C and 60C. The warming water expanded, and sea levels rose by 50 meters. Much of the coast became deeply flooded by the unbearably hot water. The Earth was out of the frying pan into the fire.

The flooded coast came to the rescue. The hot water was highly oversaturated in carbonates, caused in part by tens of millions of years of volcanic activity along the mid-ocean ridges and in part by run-off from the sudden weathering in the continents. The carbonates came out of the water, and were deposited on the continental shelves. It is seen as a distinct thick cap on top of the sediments left by the Snowball Earth glaciation. Eventually CO2 regained its equilibrium. The Earth finally became habitable again, ready for a very long summer.

Source: http://www.snowballearth.org Cap carbonate in Namibia, following the second Snowball event. TST (CD): dolostone; MFZ: limestone of the maximum flooding stage; HST, flaggy limestone and dolostone grainstone

But the long summer did not last. Perhaps this was all a false spring, the pleasant week in March before winter returns with a vengeance. After 10 million years, for reasons not well understood, the temperatures dropped again. Glaciers reformed, and sea ice again advanced. The continents were still close to the equator. History repeated itself: eventually, the ice again enveloped the globe in a snowball Earth. It stayed that way for another ten million years, before spring finally took hold. The Earth’s long summer had begun.

Life in the freezer

How ever did life survive the snowball? Where could it have survived, with the continents covered in glaciers, the sea frozen to a kilometer depth and the waters below pitch dark, hypersaline and anoxic? How did photosynthesis survive without photons for 50 million years? We still don’t know. Deep water creatures will survive near black smokers on the sea floor, but these do not photosynthesize.

There are suggestions. Some have suggested that areas of open water existed throughout the snowball in the warmest regions near the equator. These are called ‘loophole’ or ‘waterbelt’ models. But physics is not encouraging. There is no obvious way to stop the ice spreading out and filling in any gaps. The most promising waterbelt model is one where the ice is deemed to be much less reflective in the tropical region. This can give a ten-degree wide belt of water which migrates back and forth with the seasons. However, there is no clear reason why the ice here would be less reflective. Cracks in the sea ice are a perhaps more plausible alternative. They are seen on Europa, the ice-covered moon of Jupiter. They would require that life can quickly jump from one random crack to another, half a planet away.

Other ideas ignore the ocean and focus on the surface. Perhaps liquid water could exist on top of the ice. A possibility is dust. A dry world is a dusty world, even if only from volcanic ash. Over the entire globe, the dust layer would increase by one to ten meter per million years. Left alone, the dust would be 300 meters deep by the end of snowball Earth; wind would bring it mainly to the tropics. Clearly, something recycled all that dust. Perhaps every now and then the ice would collapse and turn over, as it does in sea glaciers around Antarctica. Dust on the ice will absorb sunlight, melt the underlying ice, and sink in. Quickly it develops pockets, and later ponds, of typically half a meter deep, which can be filled with liquid water and covered with thin ice. In this way, an equatorial dust band could provide the conditions for life to continue, and for photosynthesis to take place, without the need to artificially keep the equatorial ocean ice-free. It would also explain why the atmosphere never became fully anoxic, even during the depth of snowball Earth. Life in the cold ponds would be slow, but possible.

The final models propose that life continued on land, in ice-free locations perhaps similar to the dry valleys of Antarctica. Could cyanobacteria perhaps survive millions of years of complete dessiccation?

Later in the snowball, the slowly rising temperatures would make the presence of liquid surface water more likely. The coldest period during the first ten million years would be the most difficult to survive.


In the end, life survived everything life threw at it. Fossils show that the diversity of life suffered tremendously: there is a poverty of species, and low numbers, in the fossil record of the snowball Earth and of its immediate aftermath. But the recovery was spectacular. When the Cryogenian ended, the next geological epoch was ready to go: the Ediacarian. This was the time of the most amazing fossils, which look like nothing we know today. And after that came the Cambrian, with the explosion of life which Darwin already wondered about. After the Christmas flood basalt of Victoria Island, and the long winter with its false spring, came the real Spring where life sprang to new life. And after that came Summer, the never-ending days of warmth and of dinosaurs and of birds and mammals, the playtime of the Earth.

The Sun is brighter now than it was in those days, and there may never be another snowball Earth. But it is good to remember. Once, long ago and far away, a volcano brought a winter which almost ended life on Earth. Other volcanoes brought back the warmth, but it took a long long time and the recovery almost brought its own disaster. Life survived hanging by a thread. But survive it did, against all odds, and how it flowered in the next season of the Earth.

Victoria Island, in winter a place of desolation and human suffering, in summer springs to life when it becomes a place of wild beauty. It is a place of hope.

Albert, December 2021

A Christmas puzzle

To provide hours of entertaining: trace the curved line

(addendum, June 2022) For the latest on snowball Earth: https://www.nature.com/articles/s41561-022-00950-1

396 thoughts on “White Christmas

    • Interesting Lorcan. I turned back the video to see the smoke you were talking about. Definitely an appreciable amount showing up. Then I marked the area on the screen and turned back the video further back to daylight hours and the smoke/ dust/ steam, whatever it was, appeared exactly at the left side of the original cone. Like you I have no idea what it actually was but watching very closely it appeared a couple of times later in a more wispy fashion. I did wonder if that given it was around the time of 3 quakes at fagradalfjall between 19:25 and 19:39 then it might possibly have been dust from a rockfall inside the cone. Those quakes were very small though but it is probably just about possible with them occuring so close together that they might have triggered a rockfall.

      • Here is a short animation showing the smoke.

        Does anybody master smoke signal language?

  1. The MBL panorama camera is showing auroral activity in the middle of the view from 20:30 to now. Nothing is visible on the right side, of course.

  2. If you wind back on one of the cams you can see the aurora. Beautiful.

  3. Maybe is nothing, but right before the eruption in March there was a swarm of quakes out at the very end of Reykjanes, near Eldey. Just now there is a pair of quakes there, looking like more on the way.

    All of the quakes at Fagradalsfjall in the past days have been under 2 km deep too, several under 1. If quake lines trace the bottom of the dike then magma is very shallow. Must be getting really close now.

    • There has been an east to west trend in all the earthquake swarms since the onset of the volcano-tectonic episode November / December 2019. Must be something to do with how the crust responds to magma ascent, i.e. the magma intrusion at Mount Þorbjörn and the dike formation resulting in the eruption at Geldingadalir.

  4. Really very odd moment right now…just *what* is FAF going to do?

    Besides that, better now than never:
    Blessed and Happy New Year to all of you!

  5. Looking at one of the static photos, the old vent is smoking a bit and there was a helicopter having a look.
    Sadly the main web camera is still covered with the results of Thor’s sneeze. I hope they clean them up, soon.

    • I agree Clive. It is very difficult to see what is actually smoke or gas or what is my eyes playing tricks because of the poor visibility on that cam. Frustrating really. Interesting though that today the ‘smoke/gas’ has been in the same area as last nights smoke. Again I think it is possibly rockfalls on the somewhat crumbly porous surface of the cone causing what we are seeing. The bad weather of Sat/Sun could have contributes to areas shifting enough to allow for more offgassing.

      • Currently there is more smoke/gas in the area than has been since the eruption finished. Is it the start of another phase? Who knows without a drone or helicopter.

        • There’s more visible coming from the top left of the crater at the moment; can’t see the “fumarole” that was visible lower down yesterday (change in wind direction or just stopped).

        • It is the dark green double bump, much longer than the earthquakes.

          • Thanks for the reply Albert. I see it now. I did think my ‘beetle’ signature was a bit short for a helicopter buzzing around. Amazing what seismographs pick up.

  6. Calling it now, eruption tomorrow (5th Jan east AUS). There is a swarm beginning this hour out on the end of the peninsula, which happened within the day of the last eruption starting. There would also be expected degassing of the magma if it goes shallower than 1 mile, so rise should be faster than was said before.

    • Here are the most recent quakes according to the preliminary list on the IMO website. Grouped according to loacation.

  7. I am glad to see the IMO checking the smoke and vapor emissions out.

  8. A very large Asteorid impact winter maybe able to cause a snowball Earth today, given the current ice – house climate with our current cold oceans. If Chicxulub hit during the Last Glacial Maximum it coud have triggerd a snowball with the impact winter?


    • I don’t think so. Not that it would make any difference to us, as a dinosaur-extinction size impact would leave us pretty extinct too. But there is a lot of heat capacity in the oceans, and the ocean currents are now very good at redistributing the equatorial heat to high latitudes. A decade of nuclear winter would freeze the temperate regions but I expect that the much of the oceans would remain above freezing. I do recommend not to try this out. We live in funny times where national security against humans is very high priority, but people are happy to take inordinate chances with nature

  9. So… we’re all waiting.
    Do something interesting, FAF!

    Maybe Jester needs to awake it first, like he did to the volcano on La Palma 😀

  10. Hmm, a deep swarm of earthquakes seems to have occured north of Etna over the past few days it seems near Randazzo.

    • Hope it waits a bit; there are some visitors on the left hand side of the lava flow.

  11. Natthagi steams quite a lot – especially at the foot of the flow in. It reacts to rainwater. I very much doubt anything unusual is happening and there’s no tremor right now anyway.

    • The steaming is quite pronounced. It is certainly heated water and it may get its heat from the lava field and the water from a subsurface stream from above. But why now? There is a chance that some of the heat is coming from outgassing from the new dike 2 kilometers down. This is approximately the area of the shallowest earthquakes, in my recollection.

      • Magma is much shallower than 2 km, already on Dec 30 it was estimated at 1600 meters deep and rising at 130 meters a day. Nothing reported otherwise, so presumably this holds, by now it is probably less than 1 km deep, and degassing would be taking place. The forecast is for Jan 11 but degassing should expand the magma and accelerate it.

        If I recall this area actually was one where there was significant quake activity earlier, much more than anywhere else and so much it was several times falsely reported for imminent eruption. In general an eruption seems likely south of the first cone.

        • The hypocenter of the strongest earthquake in the sequence was almost right below this spot. Could it be a ground crack from the earthquake that acts as an outlet for the steam?

    • My memory matchs Clives; this is an area which was particularly prone to steaming last summer. At the same time, it does seem quite pronounced at this moment.

      I scrolled back to when the camera had a larger viewing angle. There is steam coming from many points along the foot of the hill, but always from where there is currently lava. It the steam came from an area which was not covered by lava, this would help clear up the situation. But that’s OK. I think we all like a good mystery.

  12. For some reason the FAF tremor chart has not been updated since around 6 AM this morning. I wonder what is going on.

    • I noticed that also Vito. Difficult to say what the problem is when we dont know exactly where the seismo is located. I just wish they could get it sorted asap but as I would myself be reluctant to be going out in horribly cold weather I shall not pass judgement on any delay for sure.

      • Does anyone know which other drumplot would be the next nearest to fagadsadalir?

        • The best map that I could find shows LAT as the nearest station with a tremor chart. I don’t check the drumplots, but there are other close seismic stations on the map: ?itok=25J9rxoi.

    • Looks like it was very similar to 1977, even in almost the same place too.2002 was more extensive. The volume reported for all 3 is quite similar though, just in 2002 the eruption was longer.

      I think this puts the 100 km/hr flow speed to bed though, maybe it flowed that fast inside the flow but the flow front was not moving that fast. Still longest flow made it over 10 km, if the flow that went into Goma was a more dominant vent it could have been a lot worse.

    • The Aa fronts wont be flowing very fast because they are viscous ..
      But the flood up in the fissures will be extremely fast, its surface is neither Aa or pahoehoe. But perhaps ”flood pahoehoe” the entire flow turned to Aa lava because of the quick eruption rates. The eruption is so fast and so brief it does not even have lava channels. The lava is rather very brown.. the Aa looks like horse poop everywhere in ground photos 🐎

      • Colour is probably from rapid weathering, lots of iron in the rock which is turning to iron III oxide.

        I have noticed when lava flows over a surface that is either wet or not made of lava it tends to speed up, the leidenfrost effect from water under it. Same thing can happen on steep slopes with a lot of vegetation too. Flow still is fluid but it is instead moving with no base friction, more like a pyroclastic flow, so high speed makes more sense as it is basically a dense substance falling under gravity from a great height.

        Maybe most vivid example of this is from Ambrym in 1913. Major fissure eruption and lava flood event, lava was observed flowing through dense jungle at speeds of 10 km/hr on flat ground, flowing into the ocean so fast it didnt even cool down initially. Was also a phreatomagmatic maar forming eruption on the coast, which was the more damaging part and what most reports were on. Really I think this eruption deserves a whole post on here. But main point is that for really fluid lava erupted at high rate it could flow with low friction ove wet ground.

    • https://apps.sentinel-hub.com/sentinel-playground/?source=S2L2A&lat=-1.4097233616484053&lng=29.205329418182373&zoom=16&preset=1_TRUE_COLOR&layers=B01,B02,B03&maxcc=20&gain=1.0&gamma=1.0&time=2021-07-01%7C2022-01-04&atmFilter=&showDates=false

      Looks like Nyiramuragiras caldera pit have filled again since it collapsed in 2020. Shield building is ongoing there.
      Nyiramuragira haves a serious magma supply thats larger than Etnas supply.
      Very very Impressive ..

  13. Someone posted somewhere on here about the Hunga Tonga eruption. A new video about it has been posted on YouTube today.

  14. Billions and billions of eyes staring at Reykjanes right now.
    What will happen?

  15. And one may ask yourself why is the tremor plot damaged/outdated just now?
    Do we even have deformation enough for another eruption to occur??

  16. Interesting video. Interesting speculation.

    I expect if magma surfaces it will likely take the route established in 2021. Likely, but not definitely. Don’t forget we have the MAR just a wee bit south of Fagradalir vent. This may have prevented the dike from progressing south which is why I feel Natthagi is unlikely to be a place where the lava reaches the surface. I also do not believe the area is aseismic. It’s old, cold volcanic outflow, ground down by glacial ice. Its as faulted as the rest, just lost in the lower noise.

    The steaming seen in the valley is from accumulated lava ponding, particularly towards the base of the steep inflow. It shows more on wet days.

    That’s my threepence! (No doubt about to be proven wrong…)

    • Not sure why there is the idea the dike cant cross the fault, all the other fissure swarms do and past eruptions at Fagradalsfjall seem to as well. I know it was proposed for the first eruption but that could have been a bit of a coincidence, and now the topography has changed too creating a mountain where before there was a valley, a new vent might prefer a lower location.

      • I am one of the proponents of that theory, though it’s not so much that the dike CAN’T cross, just that it is HARDER for it to cross.
        The main reason for this is that the lines of weakness that the dike follows are not likely to line up across the MAR.

        However, it is entirely possible that another line of weakness is close enough that the dike will cross, or that magma may move along the MAR for a short distance to reach another line of weakness it can exploit.

        All in all, I believe the MAR forms a boundary where the easiest path is no longer across, but up.

  17. There sure is a lot of steam/smoke at Geldingadalir at the moment . Any chance it is due to unusual weather conditions? Like recent rain/cold weather etc??

    • I have seen it steaming a lot after bad weather before now, however the steaming is more pronounced today. Whether that is because the ground there has warmed up or because the air is wetter and colder I doubt if anyone could say for sure.

    • Steam only seems to be rising from areas that were covered by lava last year, suggesting that it’s formed by residual heat. If it were coming from areas that were not lava covered, that would be more convincing that a new intrusion is nearing the surface. However, there is no rule which says that a new eruption can’t occur under last year’s lava flow.

    • I can’t see much, if anything at all on the webcams due to rain!
      So yes, hot lava will be steaming with rain pouring on it.
      However, visibility is all but zero!

Comments are closed.