PART THREE
(4.5–4.1 billion years before the present)
MARS AT THE START
You can’t stay long here, but there is a kind of awful majesty to a landscape that shimmers with so much heat; it’s like staring into the heart of a furnace. Spacesuits are useless: not only is the ground hot enough to melt anything resembling a boot, the air pressure could crush you like an empty can of pop. You’re in a heavily insulated bathysphere-type structure; the walls are half a metre thick but still you can hear the metal around you tick as it tries to simultaneously expand with the temperature and contract under the external pressure. The one saving grace is knowing that if the hull fails, it will fail faster than your nerve endings can carry pain.
The tiny porthole you have to make observations through reveals this: a flat, black plain. The surface is sometimes ropey, as if a liquid has congealed at the same time as flowing, and sometimes blocky, as if a solid crust has been tipped and broken and tipped again. From orbit this seemed to be the safest place to land, but now you’re here you’re not sure how stable the rock is beneath you. There are other places riven by deep rents through which the heat from Mars’s formation still glows. This broad, dark, cinderous plate, though, has remained uncracked long enough for you to risk a momentary touchdown.
A little way distant is what you’ve come to see – and measure – and you can feel it through the landing pads, despite the shock absorbers and all the layers of ceramic foam between you and the surface. The vents that roar out gases into the Martian atmosphere from deep underground – mostly superheated steam, boiling carbon dioxide, oxygen and hydrogen – are so hot and fierce that the already oven-like air twists and braids. You know they’re there, but you can’t actually focus on them. Each vent is a mirage, albeit one that would sear your flesh and then rip it off with the force of its discharge. The newly formed land is scattered with these sites: cracks and tubes that channel pockets of gas from under the kilometre-thick layer of cooling rock and squirt it out into the dense, bright air.
The horizon is a line made inconstant by the rising heat. The sky itself is a luminous white, cathedral-high, with a complete cloud cover that roils like the thickest smoke. The top of the atmosphere is at a ridiculous height, almost 10,000 kilometres up, and it’s only close to halfway up that the water vapour becomes cool enough to condense out. Somewhere up there it’s raining, but that rain will evaporate long before it reaches the ground.
You check the temperature outside, and you have to look twice because you think the sensor might be broken. It’s registering 350°C and a pressure of 200 bars. The probes you want to launch have already been pressurised to this. You don’t expect them to last longer than a few minutes, but you hope it’ll be long enough to take at least some readings.
You’ve been in the vacuum of space. You’ve been at the bottom of the deepest oceans. You’ve been on the flanks of an erupting volcano. But you’ve never before been in a place so implacably hostile in so many different ways. You drop the probes. You want this done and over so you can leave as soon as possible. If the extreme conditions were all you had to deal with, then you wouldn’t be so nervous, but there’s always the possibility of a meteorite strike. The early solar system is still thick with potential impactors: you think you’ve tagged them all and calculated their orbits, but if you’re wrong and a big one hits almost anywhere in the same hemisphere as you, it’s over.
Three probes scuttle across the baking surface, high-stepping on their spindly legs to limit contact with the ground. They look almost comical, like they’re dancing, but this, you’re assured, is the only way they’re going to make it to their target. They skitter and slide, but their robot brains instantly make all the calculations required to keep them on track and upright.
They’re closing in on the vents. All you can see of the probes now is their flickering silver outlines against the dark rock. Then the first one missteps: it trips at the edge of the vent on a piece of forcibly ejected rock, falls into the billowing gas and it’s gone, high into the air until it’s lost from sight. It’s somehow still transmitting, though. You quickly take a sample, even though the probe is only partially surrounded by the gas expelled from the vent. At some point, it will come down again and will probably end up dashed on the rocks below.
You move on to the remaining probes. They extend their proboscises into the almost supersonic jet of gas and sniff. The initial readings come through and you get to see what Mars is made of. The results are useful, but they’re not making up for the peril you feel.
Another probe fails: having stood still for sufficiently long to measure the jet, it has now overheated. It’s flatlined, fried. The third is still hanging on, drinking deep at the vent when your console pings. Something’s coming and you have to go, now.
Take-off is hard. The rockets kick you into your reclining couch and multiples of your own weight hold you there. Outside, through the porthole, you glimpse a distant trail of smoke arcing swiftly downwards, even as your own takes you up. It’s difficult to tell how big that particular meteorite is, or how hard it will strike – the thick atmosphere robs all but the very largest of their energy. That it’s made it this close to the ground indicates that it’s going to hit.
Sure enough, there’s a sudden flash that prints an oval of light against the wall inside your cabin. Then comes the snap as the shockwave heaves through the soup-like air. But you’ve made it: you’re high enough and far enough away that the overpressure doesn’t crack your hull like an egg, even though you feel the course correction kick in as the strong winds shake your ship.
Down on the surface is a bright circular scar, a glowing lake of molten rock contained within newly formed high walls, and a blanket of rubble that is still spreading across the dark rock of Mars in a ballistic arc. Where the rubble falls, it’ll tumble and rattle to a stop, subtly changing sightlines as well as burying the fractured surface under a layer of newly broken, altered rock. The light’s already fading as the fireball dissipates over the nascent crater, and then it’s gone: you’re in the cloud layer, heading back into orbit. You’re relieved to have made it out alive and you won’t be hurrying back.
PLANETARY MELTING
Barely ten million years after the first shockwave rattled the giant molecular cloud into life, Mars had accumulated all but a final few collisions’ worth of its eventual mass. By the measure of deep time, this is only slightly longer than an eye-blink and a little less than a drawn breath.
Mars never experienced what the planets sunwards of it would: repeated, late, massive impacts with other planetary embryos that tore both impactor and target apart, melted and mixed their contents and then coalesced them again into a new, larger whole. Neither did it experience, as is suspected of Mercury, an impact so vast that its outer layers were almost completely stripped away and dumped into the Sun. Mars, stuck at the edge of the rocky-planetesimal-rich band, avoided both the injury and the benefit of collisions with the very largest non-planet objects. Smaller planetesimals still fell on Mars and left their marks, but no great pile of matter was added to it after formation. It might have been left underweight thanks to Jupiter’s pocketing of much of its potential mass, but it was also complete.
As it accumulated through chondrule to boulder to planetesimal to embryo, the stuff of Mars was more or less the same throughout. Most of its material came from the surrounding orbits within the debris disc, and it all had a similar chemical composition. Lying just inside the snowline, Mars found itself poor in volatile gases and ices, and rich in silicates and metals; a cross-section through it would have shown it to be a jumble of rock from surface to core, held in a more or less spherical shape because of gravity.
But that would change, and the very start of this story gives us the reason why. The exploding star that shook the molecular cloud and caused it to collapse also pressed exotic matter into it. Forged from the ash of successive nuclear burnings in other, already-dead stars, elements heavier than iron were mixed with the mainly hydrogen–helium cloud. A few of these elements were radioactive, and as these elements spontaneously disintegrated, they gave out energy in the form of fast-moving fragments of atomic nuclei and high-energy gamma rays: radiation. Half-lives are the measure of how quickly a radioactive element splits apart – and an indication of its energy output: one half-life means half the element has gone. Some, like uranium, have half-lives measured in billions of years (and can be used in the radiometric dating of rocks). But others we’d never normally encounter had very short half-lives: a few tens of thousands of years.
A lump of radioactive metal is warm to the touch. That heat is the result of its decay from one unstable element to the next, and the next, until it reaches a stable atomic state. Every little block of material is generating a tiny amount of heat, sufficient to raise the temperature of the whole before it’s radiated away. But if we trapped that heat within tens or hundreds of kilometres of dense, insulating rock – which was at the same time also generating its own heat – that radioactive heat would accumulate faster than it could escape.
There is a minimum size for a rocky planetesimal to trap enough radioactive heat for it to partially melt. That size is somewhere between 20 and 30 kilometres across. At this point, Mars was 3,400 kilometres in diameter. Such was the speed of the accretion process, Mars was completed before the heat started to build. But once it did, the temperature rose so high, so fast, that the rock began to soften and sag. Pockets of molten rock formed. Any metal – radioactive metal included – sank to the bottom of each pocket, and because it was falling, it released more energy, which was turned into yet more heat. The whole process took just a few thousand years – a mere moment. Mars melted, spectacularly and completely, from centre to surface.
When we think of molten rock, we have a particular image in our minds – that of glowing, red-hot lava with dark flecks of crust floating on top, running in rivers down the sides of a volcano. But on a planetary scale it has a different quality. On early Mars, everything at the surface melted. From pole to pole, there was a single global ocean of seething molten rock. Geysers of high-pressure gas roared through this ocean and sent fountains of fire into the sky; any volatile ices that had been caught up during the making of Mars boiled out to form the planet’s first dense atmosphere of water and carbon dioxide, which acted as a greenhouse-like blanket, protecting it from the vacuum of space. And keeping the heat in the rock and making it stay liquid for longer allowed the magma more time to separate out into temperature-dependent layers.
Below the kilometres-deep liquid surface, the pressure was so great that rock hot enough to melt instead flowed like warm putty. Far denser than the surrounding rock, metals like iron and nickel were dragged downwards by gravity until they formed a vast accumulation of liquid metal at the centre of the planet. The huge amount of energy released by this sinking ensured that the warm putty-rock started its own movement: thousands of kilometres of hot, soft rock rose in planet-wide curls, reaching for the surface, cooling, then sinking again under its own weight. The less dense silicates separated out and floated to the very top, where they melted and formed continent-thick liquid rafts that drifted on denser but still incandescent rock.
What we might imagine would happen next – that the magma ocean froze from the top down, and over long years gradually thickened as heat escaped into the atmosphere – didn’t. Certainly, minerals with high melting points started to freeze out as the magma cooled, and the coolest part was at the very top, where the heat escaped into the atmosphere and was carried away. But minerals with lower melting points stayed liquid, and the magma ocean became a kind of hot rock slurry that contained solids but still flowed. And while solids are generally denser than liquids and will sink through them, some liquids are denser than some solids, which can lead to an unstable heavy, cooler solid layer balanced on top of a lighter, hotter semi-molten slurry of crystals.
On other bodies in the solar system, the low-melting-point fraction froze in place as expected and became the crust, but on Mars the conditions were different and we don’t know why. There, the deeper layers of rock stayed fluid for longer, and the layer on top – dense and rich in iron and radioactive elements – sank again, potentially all the way back down towards the core. So Mars recycled its first, ‘primary’ crust, and with it, its payload of heat-creating elements. The result was far-reaching: Mars was going to be able to generate internal energy for far longer and at a higher rate than it would otherwise have been able to do.
Mars turned from an undifferentiated rocky lump, with the same characteristics at its centre as at its surface, into a stratified planet. At the centre was an iron–nickel–sulphur core, rich in radioactive elements, solid in the middle, liquid on the outside; on the outside, a thin secondary crust of frozen rock, laden with iron oxide; and in between, a vast hot, viscous, mobile mantle of silicate rock.
This abrupt, cataclysmic, planet-wide and planet-deep event essentially reset the clock of Mars. Everything that it was before – the records held in its layers of compacted, chondrule-rich rubble and impact-melted rock – was utterly obliterated. This was its year zero. From this time onwards, after the global melting and formation of Mars’s new crust, any collision with material still being swept up from the debris disc stood a chance of being recorded and preserved as an impact crater.
CRATER FORMATION
Meteorites come in every size. Even the very smallest grain of sand can pit a surface it strikes if it’s moving fast enough – that is, in the order of kilometres per second. The particle squashes against the target rock, compressing both. The moment the moving particle stops, all that energy comes bounding back, focused inwards. The particle vaporises and forms a tiny expanding sphere of gas and molten rock. The mark it leaves is the familiar circular indentation with a slightly raised lip that we call a crater.
Scaling up from millimetre- to metre-sized meteorites, all the features of simple craters remain present – they just get more impressive. Parts of the target are vaporised along with the impactor, forming an instantaneous source of molten rock called an impact melt. Most of this will be thrown out of the crater, together with all the pulverised and excavated material, as a chaotic mix of broken rock, cooling melt and fine dust. The debris splashes out radially around the crater as an ejecta blanket, thickest nearest the crater rim and thinning out towards the margins; the size and depth of the ejecta blanket depends on the gravity of the planet and the material it’s made from. The crater floor itself is cracked and crushed, and after the event the crater walls tend to slump inwards, blurring their initially sharp edges.
Kilometre-sized meteorites produce huge, broad craters. The phenomenal pressures in the ground beneath the crater cause the rock below to fracture, at the surface and deep down, and immediately after the initial violence of the compression and rebound that vaporises the meteorite, a lake of molten rock collects inside the crater, filling it and levelling it out. A central mountainous peak often forms, caused by the sudden release of energy. Beyond the crater wall, the surrounding rocks are fractured in concentric circles, which can subsequently slip downwards, slumping into the crater floor and exaggerating the initial diameter of the crater. The ejecta blanket is projected outwards at supersonic speeds and travels as a red-hot wall of debris. The highest-thrown material arcs down with such force that it causes secondary craters.
When meteorites reach tens of kilometres across, the effects of a single impact on a Mars-sized planet are global. The impactor punches down into the crust like the fist of a furious god, breaking the rock, bending the plastic mantle beneath and sending a plume of material outwards across half the planet and potentially beyond: into orbit, and even out of it. The impact forms a basin a hundred kilometres across or more, filled with glowing impact melt that abuts the mountainous internal rings caused by reflected shockwaves. Ejecta buries everything for a thousand kilometres around under a suffocating layer of churned-up debris. Dust blots out the Sun, and secondary craters can occur anywhere over the entire planet as debris re-enters the atmosphere. Even some of the atmosphere itself can be lost, thrown out into space, never to return.
The extraordinary effect and suddenness of these large meteorite strikes is sobering. They don’t just have the power to excavate huge quantities of material to great depth and launch it outwards – they can sculpt entire regions, alter the climate and potentially knock the planet off its axis, all within moments.
However, meteorite impacts also supply us with one of our most useful tools for telling Mars’s story. If we know, even roughly, the rate at which craters are formed, we can not only date the various parts of Mars’s surface relative to one another, but crucially we can give a rough absolute date in terms of a fixed time for a crater’s formation, without ever having to set foot there.
CRATER COUNTING
Back in the eighteenth century, when geology was becoming a proper, systematic science, one of the early breakthroughs was the insight provided by the law of superposition. Simply put, this is the idea that younger rocks lie on top of older rocks. Following that observation, we discovered that we could order our own geology in terms of youngest, oldest and the layers in between, simply by comparing outcrops of rocks in various places – as well as natural cliffs and riverbanks, we could use information taken from railway cuttings, canal bottoms and mines.
After mapping all these out, we were able to tell with reasonable certainty which layers were the youngest, and then go all the way down to the oldest in an iterative process, until all the rocks were accounted for.
The Martian landscape might be different to ours, and our access to it very limited, but we can use the increasingly detailed photographs we’ve taken to map out the various landforms and rock types that are visible, and supplement those images with our knowledge of how these shapes were formed. We can do something similar for Mars as the first geologists did for Earth.
If lava flows into a valley, it has to be younger than that valley – we have no real way of telling how much younger, but if further on the valley appears to cut through a crater, we can tell that the crater is the oldest of these three features. More investigation might show that the crater wall has slumped down later, partially blocking the valley. This event, then, is younger than the valley, but it could be contemporaneous with the lava, or younger still than that.
There are further tells the landscape gives us, and they are all interconnected and overlapping. Two separate lava flows from two separate volcanoes could reasonably appear to be the same age, but by noting the features each of them cross, or are crossed by, we can determine which is older, even though they themselves never meet.
We can tentatively identify other features as being formed by water, and by wind, and by ice, and know how these riverbeds, sand dunes and glaciers are connected. Add these to the structural elements – the areas where the land has been thrust up or brought down, the valleys and mountains, and underlying it all, the ancient Martian crust – and a careful, painstaking examination of the surface geology of Mars reveals a chronology of events.
What would be best of all would be to put exact dates on that relative history. But in the absence of radiometric dating of pristine rock samples, we are left with the less satisfactory but nevertheless powerful tool of crater counting.
Crater counting is a useless exercise on Earth because the planet renews its surface so often: craters here are transient features, eroded not just by the effects of our highly active weather but by the constant motion of the underlying crust. Earth has a system of plate tectonics: its mantle moves in a complex series of gyres, with hot rock rising from the depths and pushing against the crust as it cools, before sinking again, forming vast, ponderous convection cells. The crust in turn travels along the surface, driven by the mantle currents, splitting and moving and colliding, riding up and being dragged down. Mountains rise, basins sink, oceans form and close. Any record of impacts is reduced to the most recent or the very largest, and even then most visible signs become scoured from view.
The situation is different on Mars. While, like Earth, Mars also had a mantle that moved, a full system of plate tectonics was only ever a brief episode – a couple of hundred million years, if that – after which any movement of different parts of Mars’s crust ground to a halt. From an active sheath of rock that split and moved and sank, it became a single, stagnant, solid shell. As a consequence, the whole of Mars has a basement of crust that dates all the way back to when it froze out of the global melting event, 4.5 billion years ago. It undergirds every subsequent structure, and the record of every single impact made on it is potentially preserved from those earliest times.
So how does crater counting work? Imagine a big square of wet cement – this is our early Martian crust – and next to it, a bag of unsorted rocks ranging in size from big half-bricks through lumps of masonry all the way to bits of pea gravel: these are our meteorites. We have a shovel, which, at the risk of stretching the analogy too far, represents the number of meteorites that can impact Mars in a given time. We dig that shovel into our bag of rocks, and whatever is on it we fling at the cement.
What we’ve made is lots of craters. Small craters, medium craters, large craters, in a random pattern over the whole area of once-smooth cement. Let’s take another shovel-load of rock and throw that on too, just for good measure. Some of the craters will overlap; some of the new larger ones will have erased some of the older smaller ones. Large craters from our previous effort will now be pockmarked by gravel-sized ones.
Now, we get some more cement – let’s call it a huge lava flow – and we cover half the original area. We can tell which is the original surface, because it’s covered in craters, and which is the new surface we’ve added on top, because it’s completely unmarked. Now we throw another shovelful of rock high into the air. Some of it comes down on the original surface, scarring it further. Some of it comes down on the new surface, marking it for the first time. But the important thing is that we can still tell the difference between the two – even though the newer surface now has some craters too – by comparing the size and number of craters over both areas.
Let’s do it again. Take some more cement – another, more recent lava flow – and cover half the original layer and half the second layer. We now have three ages of surface, but if we throw more rocks at them, we should still be able to work out which is youngest and which is oldest.
This is how we tell the relative ages of the different surfaces: the more heavily cratered a piece of ground is, the longer it’s been exposed to cratering events. There are caveats here, though. A surface may become so saturated with craters that any new crater might erase the memory of an older one. This means that two very old layers might record the same crater density, even though they’re not the same age. At the other end of the scale, an area might be too young to have any craters at all, but there will be no way to tell how young it is – or if it’s a small area, whether its lack of craters is due to its size or its age. Then there’s the effect of an atmosphere: the thicker the atmosphere, the larger a meteorite has to be to keep enough of its interplanetary speed to make a crater when it strikes the ground. And as we’ll see, the thickness of the Martian atmosphere has changed a very great deal, more than enough to affect our calculations.
But the biggest problem with crater counting is that while we know that the rate of cratering is an uneven process, we don’t know how uneven it is. Look in the bucket of rocks we’ve been shovelling up and throwing down. We can see that in our first few throws we took almost all of the bigger rocks, in the next few we were left with several middling-sized pieces and some gravel, and now we’re scratting around at the bottom to get anything. Going backwards from our analogy, this is precisely what happened in the early solar system.
We just don’t know the make-up of the debris swarm that intersected Mars’s orbit early on. But we do know that, like the rocks in our bucket, meteorites can only ever strike an object once. They’re not a renewable resource. Shortly after formation, Mars was subjected to an intense bombardment of leftover space debris, some of which included kilometre-wide planetesimals. After these were cleared by collisions with planets, both the number of impacts per year and the size of the potential impactors decreased rapidly.
The biggest aid to dating the Martian surface comes not from Mars itself, but from the Moon. The two aren’t particularly alike, and they formed in different parts of the solar system by different mechanisms, but one vital similarity links them: their craters. Crucially, we have some samples of the Moon, which we can not only date using radiometric dating, but because we know exactly where they came from, we can compare landscapes. With a few adjustments, we can devise a scheme to transfer the whole timescale of cratering rates on the Moon a hundred million kilometres away to Mars, and roughly – sometimes very roughly – date the various ages of its surface. Nothing trumps walking the landscape, examining the structures close up, taking samples and drawing on a paper map with coloured pencils. But these methods are the best we have for now, and they work far better than they have any right to.
Mars’s landscape might appear alien at first, but by spending time looking at the folds of the land – the craters, ridges and valleys – we start to appreciate that the processes that sculpted Mars are explicable – each age dominated by impacts, water, fire and ice. Reading the shape of Mars gives us the words to describe both its past and its present.
MARTIAN METEORITES
Much of the purpose of sending sophisticated probes, whether they are intended to stay in orbit or to land, is to take our laboratory to Mars. Spaceborne lasers can measure not just the height of the land but also the composition of the air, exposed rocks and ice. Photography has revealed weather patterns, dust storms, the seasonal variations in landforms, the growth and retreat of the ice caps and glaciers, the movement of sand dunes and tantalising hints of surface water. Magnetometers detect the residual magnetic fields imprinted on rock formations.
Down on the planet, rovers have tasted the air, drilled the rock and driven kilometres across the surface in search of new vistas. They’ve carried weather stations to record the daily rise and fall of temperature, pressure and wind speed. Their cameras have looked up at Mars’s moons transiting the sky and the pale-yellow disc of the Sun. They’ve even taken pictures of the faint dot of our own planet. But we’ve been unable to send anything that might prepare and then analyse a sample for radiometric dating. The equipment is simply too big and energy-intensive to take.
The plan is to ultimately gather up a bag of carefully labelled rocks from the Martian surface and carry them back to Earth – the newly arrived Perseverance rover will collect samples, a second rover will load them on a small rocket, and a third mission will retrieve them from Mars orbit, sometime in the 2030s. Back on Earth, we can subject the samples to the full panoply of scientific tests. We’ll have the advantage of knowing exactly where the samples came from and other contextual information about them: how weathered they are, what types of rocks were above and below them, whether they show signs of alteration or impact shock, whether they were from a crater wall or floor or from a rock bed outside, whether they were taken from a contiguous formation or a broken block of ejecta, and even their north, south, east and west orientation. All of this will help us to interpret the raw physical data.
But even without that extra information, any verifiable piece of Mars is going to be fantastically valuable for research. We’ve already discussed how a very large impact on Mars can throw material into orbit. Beyond that, it is only a short step to realising that some of that Martian ejecta might have already found its way, through the clockwork of orbital mechanics, to other planets and landed as meteorites there.
Identifying particular meteorites as having come from Mars has been a difficult process and it was only recently settled using information from the Martian landers. What was known beforehand was that some rare meteorites weren’t like any of the others. The majority of meteorites are chondritic meteorites: that is, mainly composed of chondrules. The majority of the others are from broken planetesimals which, like Mars, have either completely or partially melted: these are the meteorites which are all iron, mixtures of iron and rock, or now-frozen magma.
The Martian meteorites were originally assumed to be from the last of these groups, but they had unusual chemical characteristics that set them apart. The iron minerals were oxidised. Some were affected by recent low-temperature alteration with carbonates, salts and clays. The magmas from which they formed showed the presence of water. The mineral crystals in the frozen magma itself were different to those found in other silicate meteorites. When analysed for their radiometric ages, they didn’t all home in on 4.6 billion years, like almost every other meteorite. Some were very old, but still not that old, and some were considerably younger.
Crucially, there were minute traces of gases trapped within the rock, in crystals that were decoded to be shock-glass, which is created by instantaneous heating as a pressure shockwave passes through the rock. It was realised that these gases could only be pushed into the glass if, at the moment of melting, it had been exposed to air. After the Viking lander provided a basic analysis of the current Martian atmosphere, the similarities between this and the gases trapped within the meteorites were too great to ignore. Every difference between an asteroid-derived meteorite and these other, stranger ones could be accounted for by proposing that they came from Mars.
Do we find meteorites that come from the Moon too? Yes, and the chemical signatures of these are identical to the samples retrieved from the Moon by the Apollo missions – in fact, their recognition smoothed the way for accepting that unaltered, unvaporised debris from a meteorite impact on another celestial body can end up on Earth. Once we get curated samples back from the Martian surface, bagged and tagged, any lingering doubts will be laid to rest, but for now it’s accepted that these meteorites – numbering 266 out of more than 60,000 known space-rocks – are examples of Mars.
These rocks tell us a great deal, but they also come without context. They are random pages torn from different books, and while we can tell where in their stories they come from, we don’t know what happened before or after, or where that particular passage was written. They are all igneous rocks of some kind – either lavas that flowed on the surface or deeper emplacements of magma that were later excavated. All have recorded more than one stage of alteration, though, before the final event that saw them launched into space, to fall to Earth, where they were collected by us.
These few rocks give us a benchmark of chemical compositions, some scattered ages and tiny insights into what was happening on Mars at the times they were formed. One particular Martian meteorite is ancient, at 4.5 billion years old. But some of the meteorites appear very young, around 180 million years old, and most troublingly for our understanding of the planet, they suggest both that Mars was geologically active far later than previously thought, and that free water was present on the surface well beyond the time when it was presumed to have vanished. Like ghosts, these two ideas will haunt the rest of this story. Whenever we’re tempted to say that Mars has finally fallen asleep, we can hear these rocks whispering, not yet, not yet.
THE GREAT DICHOTOMY
Dividing the history of Mars into stages is largely a matter of established convention, but our first major reference point is going to be the formation of the Hellas crater. Its creation resurfaced most of Mars with its planet-spanning ejecta blanket – the evidence for almost everything that happened earlier was buried.
Parts of the Martian crust resemble the most heavily cratered surfaces on the Moon. Typical of these earliest surviving post-Hellas landscapes is Noachis Terra, a section of the southern highlands, and it’s from there that we take the names of the oldest Martian eras: the pre-Noachian, to denote everything that happened before Hellas, and the Noachian, for events that happened afterwards. The date we put on the Hellas event is somewhere around the four-billion-year mark, but there’s some debate about that – if we call it 4.1 billion, we’ll be roughly right. It’s older than any rock formation on Earth, for certain.
Conditions on pre-Noachian Mars are difficult to ascertain, if not unknowable. Most of the information we might want to look at has been destroyed, reworked or covered over. We can deduce little about the planet’s atmosphere, its surface conditions or its major landforms, and we can only tentatively discuss its internal processes, the state of the core and mantle or how vigorous its volcanic activity was. Most likely, though, pre-Noachian Mars was a place of almost unimaginable horror, a genuine representation of what we might think of as Hell.
Despite the vigour of that impossibly distant and most ancient time, evidence of events then has managed to survive until now – chief of which is Mars’s greatest and most implausible feature. When we were finally able to draw an accurate altitude map of Mars, it showed baldly what had been partially known for years: the northern half of the planet is significantly lower than the southern half. This is known simply as the Great Dichotomy.
It’s something that makes Mars different. Not just slightly different, but wildly, strangely different. In detail: the surface of Mars is at least 2.5 kilometres lower in the north than it is in the south, and there are parts where the difference is greater than 6 kilometres – this is an actual vertical difference in height. The junction between the two regions is, for at least half its circumference, marked by an obvious slope, taking the elevation from the north to the south up to between 2.5 and 3.5 kilometres. It’s not a cliff face; instead it’s a gentle but relentless ramp, gaining height along a horizontal travel of hundreds of kilometres.
Over that distance the landscape changes. To the south, the land is high and characterful. Craters are everywhere, crater upon crater upon crater, subsequent impacts having overwritten the record of earlier ones. Ejecta blankets and lava flows interleave each other. Using the technique of crater counting, we can judge the exposed surface of the southern highlands to be an enormous age: it measures as Noachian, but it’s hammered into the same primal pre-Noachian crust that Hellas struck.
By contrast, the north – dominated by sparsely cratered flat plains – is cloaked with deposits that are a billion years younger than those in the south. Deep beneath the plains lie the bones of huge, ancient, eroded impact craters that manifest as quasi-circular depressions. Only their vastness thwarts their complete burial.
The Great Dichotomy is far from skin-deep, though: it’s mirroring something that’s happening deep underground at the base of the crust, where it meets the mantle. We’ve been able to study this thanks to the same satellite that gave us the altitude data – it contributed to another, equally important map: a gravity map.
From first principles, gravity is a result of mass. All physical objects – from the largest supergalactic structures to the tiniest flecks of dirt – have mass, and therefore they all exert a gravitational force. The more mass an object has, the greater its gravitational effect, and the movements of objects are determined by the interactions between them: moons orbit planets, which orbit stars, which orbit the central galactic core. Also, the closer objects are to each other, the more influence they have; gravity is four times stronger at half the distance. Having said that, gravity is an incredibly weak force. Just think of a person standing on Earth: we usually manage to hold our own bodies upright against the pulling power of a billion trillion tonnes of rock.
Gravity is almost, but not quite, the same value across the entirety of Mars. It takes very sensitive instruments to detect the parts-in-a-million differences, and we can also very closely monitor the orbits of satellites around Mars to see if they deviate from their predicted paths. Where gravity is weaker on the planet below, the satellite will rise ever so slightly as it passes. Where it is stronger, the satellite will dip down. The difference is in millimetres, but we can measure it.
Large masses of rock, like mountains and volcanoes, create minute amounts of extra gravity. Likewise, huge craters have a missing mass and there is less gravity. A first-order map of gravity – called ‘free air’ gravity – will often resemble the relief map, where mountains have positive values and craters negative ones. But the power of a gravity map is its ability to dive under the superficial and measure what is happening below. The gravity values can be adjusted for the visible discrepancies in height and a second map emerges. It tells us where the gravity of Mars deviates from what we expect, and among the things it tells us is this: the Great Dichotomy is baked into the structure of Mars.
The average thickness of the light silicate crust that overlies the denser mantle varies as we move from north to south. In the north, it’s an average of 32 kilometres thick. In the south, it’s almost twice that, at 58 kilometres. The distinct difference between the two regions means we can draw a line all the way around Mars, even where the visual signs have been obscured on the surface above.
How did this startling feature, unique among all the planets, come about? We have two main explanations and little to choose between them. Each one has its supporters. Each one has evidence to back it up. Rather than choose which to believe and which to reject, I’m simply going to present both. They can’t both be true. But because they both end up at the same place – with the Great Dichotomy girdling Mars – there’s no harm done. We can entertain both ideas without committing to just one.
THE GREAT DICHOTOMY CONVECTION THEORY
Back when Mars was molten, the iron–nickel component sank to the centre to form the core, and the lighter silicate crust rose to the top. Between them were the denser, heavier silicates of the mantle which, while hot enough to be molten, were under so much pressure that they behaved more like slow-moving putty than a liquid. The crust cooled, forming a rigid lid, but underneath, everything else was still hot and continued to generate heat.
Heat is transferred in three ways. First, radiation: what you feel when you put your hand up to the Sun is the direct effect of radiation – the energy from the light striking your skin. Second, conduction: this is what you feel when you put your hand on a radiator and touch the hot metal, which is heated by the water inside it. Third, convection: this is what you feel when you put your hand above the radiator, as the air heated by the hot metal rises to mix into the air in the room.
All things above absolute zero (–273°C) radiate heat in the form of light, and the hotter they are, the more radiation they produce. The energy of the light depends on its wavelength, and we can measure that to find out how hot an object is. At cooler temperatures, the light is infrared – invisible and long-wavelength – but as a material heats up, it begins to shine in visible light, which has a shorter wavelength. A hot object will first glow dull red, then bright yellow and finally blinding white. The temperature of the material dictates the colour and the intensity of the light coming from it.
Conduction is simply the ability of a material to transfer heat energy from the hot side of itself to the cool side. Metals are very good at this; rock, not so much. Liquids and gases are terrible at conduction, since when they get hot, they expand and, being less dense, they rise up, taking their heat energy with them.
Which brings us to the third mechanism: convection. Convection is why your radiator is warmer at the top than the bottom: inside the radiator, the hottest water is the least dense and it rises as far as it can go. When it can’t rise any further, the trapped water passes over some of its heat energy to the metal of the radiator through conduction and the air by radiation. This cools the water down and it sinks, to be replaced by more rising hot water in a continuous process of circulation.
Convection also applies to hot, plastic mantle rock. Held in a semi-solid state by the overlying pressure, it can still move very slowly. Heat at the bottom of a thousand-kilometre column of mantle rock is initially trapped, but if the rock starts to flow, both the rock and the heat it carries can be transported upwards at the rate of centimetres a year. Given a long enough time span, that rock will arrive at the bottom of the crust, where the heat will be conducted imperfectly away, up and out towards space, and the cooler rock will sink back down again in a huge, slowly stirring gyre.
The greater the difference in temperature between the surface and the core, the faster the mantle moves from deep down near the core to up against the base of the crust. And like a deep and secret current, the mantle starts to push against the crust, dragging it along as it sheds its heat and then plunging down into the depths again.
This is the exact mechanism that is happening to this day on Earth. Plumes of rising, hot mantle, starting at the core–mantle boundary and ending below the crust, drag plates of crust with them as they turn over and descend. The surface plates pull apart from each other in some places – where we find lowlands, basins, seas and oceans – and collide edge-on in others; where collisions happen, one plate rides over the top of the other – forming mountain ranges on one while the other is forced underneath. Each fracture line between plates is associated with volcanoes and earthquakes, and the motion is constant, gradual and glacial – centimetres a year, at most.
But did this ever happen on Mars? It’s difficult to say, and the answer is probably yes and no. There was certainly a failure to sustain planet-wide mantle convection and a proper system of plate tectonics. If it ever happened at all – and the evidence is scant – then it died away quickly. That might have been due to Mars’s size or the material it was made from (or both, or other factors we haven’t divined), even though the core was still belching out heat and the mantle was hot and soft.
What we do have evidence for, though, is one very large plume – and this is the basis of one theory for the Great Dichotomy. If there was a singular plume of hot mantle material being driven upwards by heat from the core, directed roughly towards what is now the south pole and pushing against the crust until it cycled down again, cool, at the north pole, then we’d get exactly what we see: a high, thick southern highland, where extra material has been pressed onto the base of the crust, and a low, thin northern plain, where the base of the crust has been peeled away, dragged down and entrained in the sinking mantle below.
Computer modelling can show that this concept isn’t just a fever dream, and that a planet of Mars’s size and presumed composition could operate just one plume to the exclusion of others. In fact, as the single plume began to dominate, it would have become so efficient at removing heat from the core that any other plumes that might have challenged it would have faded away.
The most immediate and compelling evidence that suggests this may have happened has to do with the behaviour of Mars’s core, which is one of any planet’s least knowable parts. We can never sample it, never see it and can only infer its presence and composition by indirect measurements. Despite this, we know that Mars has to have a core, on the basis of how big and how heavy the planet is: we can calculate how much of Mars is rock and how much is metal.
The cores of rocky planets – ones made of predominantly iron and nickel – start off as fully liquid, intensely hot but trapped in the centre, crushed by the weight of rock all around. A solid core starts to form soon after at the very middle. This fraction of the core slowly increases as heat escapes through the mantle and crust, until the once-liquid core solidifies completely.
But in the between-times, when there are both solid and liquid metal in the core, the liquid will move in swirling convection currents, just as the mantle above it does but vastly quicker. Those currents are affected by the rotation of the planet, and they all spin in the same direction.
Liquid metal conducts electricity just as solid metal does, and if that conducting fluid moves through a magnetic field – like the one flowing out from the Sun – it produces electricity. Those electric currents in turn create their own secondary magnetic field, which goes on to support the generation of electricity, and so on. A self-sustaining electromagnetic dynamo switches itself on like a light bulb, powered by the rotation of the planet, and it will last as long as the conditions are right – when there is both rotation and a liquid core.
Back at the Martian crust–mantle boundary, a thousand kilometres above the core, the plume of hot mantle added kilometres of rock onto the underside of the crust by injecting semi-molten magma into the lower levels. It simultaneously dragged the crust along as the hot mantle turned ninety degrees and scraped northward along the underside of the crust. Magnetic minerals within the crust – at this point, likely to be in the form of the iron oxide mineral magnetite – were reheated by the new material, and when they cooled again, they preserved a record of the direction and strength of the magnetic field produced by the dynamo in the core, writing the information into the rock like a piece of magnetic tape.
One of the things we know about planetary dynamos is that they flip the direction of north and south at various time intervals. If the magnetic tape of the rock continually records the passage of time as it moves, then it inadvertently stores the north-then-south-then-north-again direction of the magnetic field as a series of stripes. And we can see those stripes from orbit, using a satellite equipped with an instrument that measures the magnetism of the planet below. Although the dynamo no longer works in the present day, the evidence showing that it once did appears as a pattern of differently magnetised bands centred around the south pole like a bull’s eye. Imperfect, battered by later crater-forming and volcanic events, but tantalisingly present all the same.
This left us with what we see on Mars now: half a planet with a thick crust, centred close to the current south pole, that was pushed and squeezed northwards as a creeping tide. The thinner crust was nudged ahead of it and thinned by the sinking mantle near the north pole. But before we take this on trust, we have to consider the other scenario.
THE GREAT DICHOTOMY IMPACT THEORY
If you find the convection theory a bit of a stretch, how about this? The Great Dichotomy resulted from a single, giant collision that left an elliptical impact crater roughly the size of the entire northern hemisphere.
For this theory to work, we’re dependent on the serendipitous coming-together of several factors entirely external to Mars. First, we need a giant impactor, a leftover planetary-embryo-sized one, some 2,000 kilometres across. Secondly, we need it to be moving slowly relative to Mars: somewhere around a 6–10 kilometres per second difference (this is slow in cosmic terms). Thirdly, it has to strike Mars’s northern hemisphere at an angle of around sixty degrees to the vertical.
Conditions outside of these won’t give the results we want. A smaller impactor would leave a regular, albeit huge, crater. A larger one would remelt the entire surface of Mars. A slower impact would add extra material to Mars, not take it away, but a faster impact would strip Mars of much of its mass. A head-on collision would produce a circular shockwave that would travel around the planet and meet at the back, 180 degrees from the impact site, blowing a smaller but still significant hole through the crust there too. A shallower-angled impact wouldn’t give the shape to the northern lowlands that we see.
We’ve already accepted that ‘spare’ planetary embryos existed: that Mercury is almost all core and no mantle is likely an example of a planetary embryo impact, and the formation of Earth’s Moon is strongly believed to have been the result of a Mars-sized embryo colliding with the proto-Earth, 4.5 billion years ago. No physical evidence of either of these impacts remains: they were simply too large and too energetic.
Slow-moving planetary embryos sharing orbits is also not too much of a reach. There are stable points – Lagrange points – one-third of the way around a planet’s orbit, both ahead and behind it, where a decent-sized planetesimal might accumulate material without falling into the main planet. These orbits are stable over hundreds of thousands of years, but we’re operating in time spans of millions of years. Any perturbations caused by, say, wandering gas giants moving into the inner solar system and then back out again would disturb the delicate balance of gravity between the Sun and Mars, and anything at a Lagrange point might be nudged away from it. Because such an embryo shared the same primary orbit around the Sun, Mars would eventually capture it – not in a full orbital-velocity-plus slam, but at a slower, more considered shunt.
Oblique-angle impacts are obviously possible: there’s no law of nature that insists an incoming meteorite has to hit at right angles to the ground. But how does an impactor’s angle of approach affect the shape of the crater it leaves? We’ve already learned that it’s not the actual downward movement of the impactor that excavates the crater – it’s the explosive expansion of the superheated vapour created from both the impactor and the surface directly beneath it. That should mean that all impact craters are circular. And they are… almost. If the initial fireball is elliptical, because the impactor is large enough to have smeared itself across the target before it vaporised, then the resulting crater will be longer in the direction of travel than it is wide.
There’s a sweet spot, then – a Goldilocks scenario that is both entirely possible and highly improbable. It doesn’t matter how unlikely it is, though, because what we see in Mars’s northern hemisphere can be explained by this single, huge impact. Something caused it, so why not this?
In the moments after the collision, an elliptical crater formed, covering almost the entire northern hemisphere. A supersonic shockwave of heat rushed outwards, eventually encompassing the whole globe. Debris from both the impactor and Mars followed as a near-solid wall of material. Some of it was launched into space – as incandescent gas, cold broken rock and every state between. The crater, deep and very broad, filled with a deep sea of lava. Much of the atmosphere of Mars – the dense, steaming, thick, early atmosphere – was blown off the planet and lost forever.
Then the curtain of excavated material – that which hadn’t achieved orbital velocities – started to descend. It was literally a rain of fire from above, one that resurfaced the rest of the planet. A larger, more energetic impact would have melted the planet for a second time, but that didn’t quite happen. Still, the effect was global and complete. Much of the early crust in the north was added to the broken south in a tens-of-kilometres-thick ejecta blanket.
As it cooled, the lava lake settled into the crater floor, producing flatter, lower terrain. Much of the crust had been deeply cracked, not just underneath the crater, but everywhere, from top to bottom. The northern mantle relaxed and rose after much of the weight of the crust was lifted off it. Conversely, in the south, the extra weight pressed down, causing the mantle to heat up. The rim of the crater, initially well defined, slipped and slumped along fault lines as the deep rock underneath moved to accommodate the new situation. As the debris in orbit continued to spiral back down, it cratered and marked the new landscape below.
Mars was left with an abnormally low northern hemisphere and an abnormally high southern one, but the age of the top of the crust beneath each dates to the same period: one was formed by the cooling of impact melt, the other was formed by the shovelling of a wall of hot rock out of one hemisphere and onto the other. In an age where whole protoplanets fell into the Sun and others were torn apart and remade, it seems remarkable that we might still have a record of an event that almost, but not quite, broke Mars.
Two stories, then. Either might be true. Both might be false. And there are other explanations, one of which even suggests that the first theory could have resulted from the second – a giant impact set off mantle convection by providing a thick insulating southern cap, which built up heat beneath it and encouraged the reluctant mantle to move. We’re in the dark here. More information might settle the matter, or it might confuse the theories that we already have.
However it formed, we are left with the Great Dichotomy. Nothing quite like it is seen anywhere else in our solar system. Among other merely spectacular features, it’s unique to Mars. It’s so significant that it dominated the way the planet developed over the next 4.5 billion years. The immediate effect was to leave Mars wildly unbalanced forever. Wherever the initial impact or plume occurred, the difference in crustal thickness meant that the whole planet tipped over to a more stable configuration, leaving the Dichotomy lying around the equator.
The ancient Noachian landscape, with its record of the earliest impacts, now lies completely buried in the north, under later deposits: all that we can detect of it are the deep wounds that manifest as quasi-circular depressions, hidden under layers of lavas and inflowing erosion products from the highlands – sandstones and mudstones and salts. In what became the south, the primordial surface remains exposed, exhibiting its elementally inflicted scars for all to see.
PHOBOS AND DEIMOS
Just as a planet orbits a sun, a moon orbits a planet, bound to it by ties of gravity. Moons can vary enormously, in size, composition and shape. Some, like Ganymede, a moon of Jupiter, and Titan, a moon of Saturn, are worlds in their own right, and only Mercury and Venus do without moons in our solar system. Even Pluto, recently demoted from planet status, has a significant moon called Charon. That Mars has two moons, Phobos and Deimos – named after the twin sons of Ares, the Greek version of the Roman god Mars – is blessedly unremarkable.
This being Mars, though, nothing is ever simple. While these moons are somewhat potato-shaped and very much on the small side – Phobos’s longest dimension is 27 kilometres, Deimos’s is 15 – this is not specifically troubling. The first odd thing is that Phobos’s orbit is really very close to Mars, a mere 6,000 kilometres above the surface. This is so close that it can’t be seen from the surface above seventy degrees north or south – the moon is always below the horizon towards the polar regions. Phobos is so close that it orbits faster than Mars rotates: rather than rising in the east, it rises in the west and apparently tracks backwards across the sky, twice in one day. It moves quickly enough that, if we were to stand on Mars and look up, we’d see it moving, west to east, in a little over four hours. It’s so close that it visibly grows as it passes overhead, then shrinks into the distance.
Yet it’ll move closer still. The same gravitational forces that have locked Phobos into only ever showing one side to Mars are also causing it to gradually fall towards the planet at a rate of just under two centimetres a year. Not that it’ll ever strike the surface – before then, Mars’s gravity will tear Phobos apart and smear it across the heavens. What is now a lumpy but coherent piece of rock will become a long train of debris in some ten million years’ time, which will then spread out further and form a planetary ring. The falling-apart of Phobos will take little more than a month.
Part of the reason for Phobos’s future disintegration is that it is barely there. If it was made from solid rock, its density and surface gravity would be twice what they are. Instead, Phobos seems to contain a lot of nothing. Depending on what we believe Phobos to be made from, the amount of nothing varies from around one-fifth to nearly half its volume. But since we only have ideas as to the composition of the moon, the estimates remain very much in the realm of guesswork.
It’s suspected that Phobos is made out of some sort of accumulated leftover debris. It’s uncertain whether this came from a small, captured rubble-pile asteroid or a small agglomeration of planetary debris thrown up from the surface of Mars by a giant impact early in its history – like the one that may or may not have caused the Great Dichotomy. There are other candidate craters we can see, and potentially ones we cannot.
We need to pause here to consider one of the more bizarre but of-its-time-rational theories: that Phobos is an alien construct, and that its low density is explained by it being nothing more than a hollow metal shell. Russian astrophysicist Iosif Samuilovich Shklovsky seriously proposed this in 1958 – that given its density and size, the idea that Phobos is an iron hull around fifteen centimetres thick surrounding a vast interior space adequately explained everything.
It took a good decade before anyone could prove him wrong, and we know for certain now, having better photographs of Phobos, that it isn’t a relic of a Martian – or anyone else’s – empire. We do know that Phobos is extraordinarily dark, though. Albedo is a measure of reflected light. Fresh snow reflects around 90 per cent of light; Phobos reflects just 7 per cent. The whole of Phobos is covered with a deep, hundred-metre layer of broken rock and dust. It’s almost perfectly black, except that there’s a reddish tinge to its tail and a bluer tinge to its head.
Also at the leading edge of Phobos is Stickney crater, caused by an impact that must have been close to the limit for breaking apart the entire moon. Stickney is 9 kilometres across, on an end-on diameter of only 20 kilometres. At first glance, there are lines, ridges and faults radiating out of Stickney across most of the surface of Phobos, but the lines also cut Stickney’s crater rim and no one is certain what they are or how they got there.
Stickney is not the only crater, though, and now that Phobos has been mapped, an attempt at crater counting – and therefore giving it an age – can be made. If Phobos formed around Mars, then it’s very old: 4.3 billion years old. If it didn’t, and it was captured from elsewhere in the solar system, where impact speeds were a third less and the craters made were comparatively smaller, then it would be a mere 3.5 billion years old.
Deimos, the smaller, more sensible brother, is further away from Mars at 20,000 kilometres distant, and it’s moving away from the planet because it sits on the other side of a gravitational line: in 100 million years it will slip Mars’s orbit completely. It does, at least, have the decency to rise in the east and set in the west, and its orbit takes only slightly longer than a single Mars day – it takes nearly two and a half sols for Deimos to fully cross the sky.
Deimos is made exclusively of the red dirt found on Phobos, but it’s even less dense than its brother. It’s marginally darker too, and smoother. The crushed-rock surface of Deimos might be in the region of a hundred metres thick, like on Phobos, but it’s barely held in place. Because Deimos is so very small and has so little gravity, crater-making impacts will throw almost all their ejecta into space. From the few photographs we have of Deimos, it might just hold the record for the largest crater in relation to the size of the object. The unnamed hollow at the south pole is 10 kilometres across, on a surface that is only 11 kilometres wide at that point.
From the ground, and despite its darkness, its proximity makes Deimos appear as one of the brighter stars in the sky. But neither Martian moon has ever troubled events on the planet’s surface. A Martian ocean has never risen up to greet them. The Martian ground has never flexed beneath them. They are too small, too insignificant to affect the surface. But they are there, and we acknowledge them all the same.
THE EARLY MARTIAN ATMOSPHERE
All of the ices, and all of the gases caught up in them, boiled when Mars melted. Just how much ice and gas ended up as part of early Mars is another question for educated guesswork to answer, but it would have been a lot, and potentially a very great deal indeed.
Mars formed from debris that was close to the snow line around the proto-Sun, at the point where water turned from a gas directly into a solid through the process of deposition. A rough calculation that involves making Mars out of comet-like planetesimals consisting of half water and half rock leads to an estimate of well over a hundred billion tonnes of water included in the material that formed Mars. If all that water was liquid, it would cover a Mars-sized sphere to a depth of between one and two kilometres.
But given the temperature of early Mars – a still-hot, semi-molten planet skinned by a fractured, congealing crust, riven with eruptions of lava and instant heating caused by repeated, giant meteorite impacts – the surface at that time was too hot for liquid water. All of the ocean’s worth of water was steam, and when combined with the carbon dioxide that had streamed out of the rock along with it, Mars most likely had an atmospheric pressure shortly after formation somewhere between 100 and 300 bars. That’s the same pressure as experienced between one and three kilometres down in Earth’s oceans. Today, Mars’s average air pressure is 30,000 times less.
This was a searing, crushing, highly chemically active atmosphere. The temperature fluctuated wildly. When it fell low enough for water to condense out, it would have rained in torrents: temporary lakes, seas and even oceans would have gathered in low areas, with the water draining through cracks in the crust caused by impacts. Here it would have met the intense underground heat and re-emerged as steam: geysers, hydrothermal vents, pools of superheated boiling water under a dense, hot water-and-carbon-dioxide atmosphere. When the temperature rose again – after a large impact – those seas would have boiled off, raising the air pressure even higher. The greenhouse effect – the ability of the atmosphere to absorb and retain heat – would have played a significant part in keeping the surface of Mars hot, too, as both water and carbon dioxide are efficient greenhouse gases.
And yet that early, thick, water-heavy atmosphere is unequivocally gone now, stripped away and replaced by something almost entirely different. Where did it go? It seems to have vanished as quickly as it formed, in a few tens of millions of years, and we have to hunt for potential mechanisms that might have contributed to this act of atmospheric theft.
Giant meteorite impacts can accelerate the column of air above them, punching it into space, never to return. But we also know that meteorites are a net contributor of volatile gases and ices to a planet: if we were to bombard Mars with comets, no matter the wreck they made of the air that was already there, the fact that the comets brought ices to the planet means that we’d ultimately end up with more atmosphere than we started with, not less. So while atmosphere loss from meteorite impact was absolutely a factor, the extent of it depends on what the meteorite was made of. Early in the life of Mars, this doesn’t look like the culprit.
The atmosphere could have simply wandered off the planet. Mars has a low surface gravity, making the column of air above Mars very tall and the escape velocity – the speed at which something needs to be travelling to leave Mars’s orbit – proportionally low. At the very outer reaches of the atmosphere, it only takes a nudge for air molecules just to drift away into space.
The Sun continuously emits a stream of charged particles from its yellow-hot surface, known as the solar wind. These broken parts of atomic matter can strip gas from a planet by colliding with the top of the atmosphere at high speed and, like billiard balls, knocking the gas away. Planets with a dynamo-type magnetic field deflect those solar particles around the atmospheric envelope, while those without suffer dramatic losses of air. But early Mars had such a magnetic field – it was protected. When it failed, solar wind predation became important, but for now, it’s not in the frame.
The most likely answer is that Mars’s atmosphere was inherently unstable because of the planet’s low gravity. Mars’s early hot and energetic atmosphere ballooned to not just hundreds of kilometres above the surface, but thousands, and possibly as far as twice the planet’s radius. This extraordinarily massive gas envelope around a small rock core would still have been protected from the solar wind particles by Mars’s magnetic field, but it would have sat in the full, unflinching gaze of the Sun just as it began to settle into its hydrogen-burning phase. The star’s output of short-wavelength ultraviolet light and X-rays provided more than enough energy to split molecules and send individual atoms over the escape velocity barrier and away. Any gas that found itself at the distant margins of the atmosphere escaped easily into space.
It was just a question of when – not if – Mars’s giant atmosphere would dwindle and fade. From being over 100 bars at its peak, the combination of low gravity, high heat and vigorous sunlight pushed the atmosphere out and away. Some 100–150 million years later, the surface pressure was down to perhaps one or two bars. By that time, almost all the icy planetesimals were gone, having already collided with other objects or settled into stable orbits elsewhere. The only source of fresh atmosphere available was what Mars had managed to hold on to inside itself: carbon dioxide, sulphur dioxide and, critically, water, kept within the crust, where it had percolated down through cracks and into aquifers – great storehouses of liquid below the surface.
The Great Dichotomy had already formed. The northern lands were kilometres lower than the south. The pressure dropped as most of the atmosphere bled away, and with it the surface temperature. The cooling water vapour and carbon dioxide atmosphere wasn’t stable and would, over the next four billion years, diminish to almost nothing. But the conditions at that moment were right: if water was ever going to flow freely on Mars, it was in the early Noachian age.