The systems that keep
the world running

Eutrophication — the bloom that kills its own world

Lake Erie, 1953. Or any lake, any summer, whenever the phosphorus arrives.

In the summer of 1953, on the shores of a lake whose name does not matter, a fisherman noticed that the green was different. Not the green of shallow water over limestone, not the green of early spring, but a green with weight to it, a green that moved like a slow animal and smelled of copper and rot. He said nothing, because what was there to say, and anyway the fishing was still good, and anyway the summer was hot.

What he was seeing was a system discovering its own limit from the wrong side. Nitrogen and phosphorus had arrived, invisibly, from the fields, from the sewers, from the soap, and to the algae this was not pollution. It was permission. It was the removal of every constraint. They divided with a speed that had no precedent in the history of the lake, and for two weeks the water was worth more than gold to any organism that could photosynthesise, and then it was worth nothing at all.

What follows is a curious thing, almost cruel. The bloom shades itself. The cells at the surface block the light from the cells beneath, and the cells beneath begin to die, and the dying releases the nutrients they had taken, and there is a moment — no one knows exactly how long — when the whole system is alive and dying at the same time, a green engine consuming itself.

Decay was the second engine. It always is. The bacteria came, and the bacteria breathed, and the oxygen they breathed was the oxygen the fish needed, and by August the bottom of the lake held no oxygen at all, and the fish floated up with their mouths open, in rows, like a sentence trying to be spoken and failing.

Not everything died. That is the part worth remembering. What survived did not need air: anaerobic bacteria, sulphur-cycling microbes, cyanobacteria curled into spores like tiny clenched fists. The lake did not end. It became a different lake, running on a different chemistry, with a different cast. The nutrients settled into the sediment, where they remained, patient, ready.

The lake had been, for a summer, the purest argument for a proposition no one wanted to accept: that life does not exist in an environment. It is the environment, and it rewrites itself without ever meaning to.

The nitrogen cycle — the air that isn't food

Every protein in your body began inside a bacterium you will never meet.

In 1909, in a laboratory in Karlsruhe, a German chemist named Fritz Haber did something that had been, until then, the exclusive province of a few obscure soil bacteria: he pulled nitrogen out of the air. Not metaphorically. He took the gas that fills seventy-eight percent of the atmosphere — the gas that is, from the perspective of almost every living thing, inert, useless, a colourless famine — and he broke its triple bond, and he turned it into ammonia.

The triple bond of N₂ is one of the strongest in chemistry. Breaking it takes more energy than a lightning bolt delivers in a single strike. Only a small set of bacteria — some free-living in soil, some housed inside the root nodules of clover and beans and alder — can do it, and they do it with an enzyme called nitrogenase, and they have been doing it, quietly, for two billion years.

What Haber did was not create life. It was something stranger: he made the atmosphere edible. And in doing so, he made possible a world in which four billion people now exist who would not otherwise have been born.

And yet. The nitrogen that feeds them does not stay where it is put. It leaks. It washes into the groundwater as nitrate, and into the air as nitrous oxide — a gas roughly 270 times more potent than CO₂ as a greenhouse — and into the rivers and lakes, where it becomes, all over again, the fuel of the bloom.

The most important organism in the history of your body is a bacterium you will never see, and the most consequential chemist of the twentieth century was a man who learned to imitate it, and the two facts are the same fact, and they are not entirely a comfort.

The phosphorus cycle — the nutrient with no sky

Carbon flies. Nitrogen flies. Phosphorus can only be carried, and it is tired.

Phosphorus has no atmosphere. This is the fact on which the rest depends. Carbon moves through the sky as CO₂, nitrogen as N₂, oxygen as O₂, sulphur as SO₂ — but phosphorus, the element that builds the backbone of every DNA molecule on Earth, the element without which no cell can store energy, has no gaseous form. It cannot escape the ground. It can only be eroded, slowly, grain by grain, from rock.

In the soil, it is a prisoner. The moment phosphate is released by weathering, it is seized by iron, by aluminium, by calcium, and locked into minerals from which plants cannot extract it. So plants wait, and the phosphorus waits with them, and the cycle turns so slowly that a single atom might spend a million years in rock, a thousand in soil, a season in a leaf, and then a hundred million in the seafloor before the planet's slow tectonic churn lifts it up again.

What humans have done is to accelerate one half of this cycle and leave the other half exactly as it was.

We mine phosphate from fossil seabeds — rock laid down over tens of millions of years when the ocean was warm and shallow and full of bones — and we spread it on fields, and the fields leak, and the rivers carry it to the sea, and there it settles into sediment where it will remain, effectively, forever. Every gram we apply is a gram removed from the geological cycle and deposited, permanently, somewhere it cannot come back from on any timescale that matters to us.

There is no substitute for phosphorus. There is no atmosphere to draw it from, no synthetic route to create it. When the mines are exhausted — and they will be, sooner than anyone wants to say — the element that structures every living cell will simply be, in the places where we need it, gone.

The carbon cycle — two cycles wearing one name

There are two carbon cycles. We have spent two centuries trying to make them into one.

Consider the forest. In spring, it inhales. Across the entire northern hemisphere, the trees and grasses and mosses pull CO₂ out of the air with a hunger that has been tuned by three hundred million years of evolution, and for six months the atmosphere of the planet loses carbon, measurably, and then in autumn the leaves fall and decay and the carbon returns, and the whole northern half of the world breathes in and out like a single slow animal, once a year, every year, and the atmospheric CO₂ record shows it clearly, a sawtooth scribed into the air.

That is the fast cycle. It runs on sunlight and seasons. It is not, however, the only carbon cycle, and the other one is so slow that it is nearly invisible to human perception, and yet it is the one that will decide the next ten thousand years.

The slow cycle is made of rock. Carbon in the air dissolves in rain, forms carbonic acid, and over hundreds of thousands of years eats away at granite and basalt. Marine organisms use the resulting ions to build shells of calcium carbonate, and when they die their shells sink and accumulate on the seafloor into limestone, and over hundreds of millions of years the limestone is subducted into the mantle and returned to the surface as volcanic CO₂.

For the entire history of the human species, this slow cycle and the fast one ran at their own speeds, separate, unmixed. Then, in the space of two hundred years, we dug up carbon that had been stored in the slow cycle for three hundred million years — coal, oil, gas — and dumped it into the fast one. The fast cycle has no way to absorb a shock of that size. The carbon is going into the air, and the air is going into the ocean, and the ocean is becoming acid, and the corals are beginning to dissolve.

No one did this on purpose. No one had to. We simply reached into the geological past and pulled the handle, and now the two cycles that were never supposed to meet are meeting, and the result is not a change of climate, it is a change of timescale, and there is no precedent for it that does not end with most of the things we love being dead.

The oxygen cycle — the atmosphere life built

Two billion years ago, oxygen was a poison. Then it was the air. Both things are still true.

In the beginning there was no oxygen in the air. This is almost impossible for us to imagine, because we are oxygen's creatures, we are built to burn it, but for the first two billion years of life on this planet the atmosphere was a mixture of nitrogen and methane and carbon dioxide and water vapour, and free oxygen was a rare and dangerous thing, a waste product, a pollutant produced only by a few obscure bacteria that had learned to split water with sunlight.

Then something happened. Around 2.4 billion years ago, the sinks that had been absorbing all that oxygen — the dissolved iron in the oceans, the volcanic gases, the reduced minerals on the seafloor — filled up. And the oxygen that had been a trace became a flood. The atmosphere went from a fraction of a percent O₂ to several percent in a geological instant, and for the anaerobic bacteria that had dominated the planet for two billion years, this was not a new dawn. It was the first mass extinction in the history of the world, and it was caused by nothing but the waste of a metabolic reaction.

The oxygen you are breathing now is the leftovers of that catastrophe. It is the fraction of photosynthesis that was never consumed, that got buried in sediment as organic carbon before the decomposers could eat it, that has been accumulating in the air for hundreds of millions of years like the balance in an account no one is tending.

Which means that every time you take a breath, you are breathing the waste product of another organism's metabolism, an organism you will never see, and you are breathing it because the earth, over a period of two billion years, failed to consume all of what was produced. There is no design in this. Only residue.

And when a lake dies, when the algae bloom and crash and the bacteria multiply and the oxygen falls to zero, what you are watching is the fast cycle failing. The oxygen debt that had been accumulating since the Great Oxidation Event is, for a few weeks, called in.

The water cycle — the planet's circulatory system

The rain is old. The rivers are old. Only the aquifers are older than the rain.

Picture the sun as a pump. Every second, it lifts water into the air — from the oceans, from the lakes, from the soil, from the leaves of every plant on the surface of the earth. A single mature oak can raise four hundred litres into the sky on a hot day, which is a fact so strange that it took botanists a century to accept it, and yet it is true, and it is happening right now, in every forest on every continent, an unmeasured flux of water moving from the ground to the clouds.

That water comes down again as rain, and it does so unevenly, in bursts, in storms, and some of it runs off into streams and reaches the ocean within days. But some of it — a fraction, sometimes a large fraction — does not run off. It sinks. It goes down through the pores and the cracks and the fissures of the soil, past the root zone, past the reach of any plant, into the deep rock, where it accumulates in aquifers that were filled by rainfall before the last ice age.

There is water under the Sahara that fell as rain twenty thousand years ago. There is water under the American plains that fell during the Pleistocene. It has been sitting there, unmoving, through every human civilisation, and we are pumping it out in decades.

When you pump an aquifer faster than it recharges, you are not just removing water. You are removing the slow loop that keeps rivers alive through drought. The stream that once ran year-round because it was fed by groundwater now dries up, not because the rain has stopped, but because the reservoir beneath it has been emptied. The stream and the aquifer are one system, and the system has two directions, and we have reversed one of them.

The water in your glass was once in a dinosaur, once in a glacier, once in a Roman aqueduct, once in a well dug by someone whose name no one remembers. It is the same water. There is no new water. There is only the old water, moving, and some of it is being spent.

Thermohaline circulation — the ocean's global conveyor

The ocean has a heartbeat. It takes a thousand years to beat once.

There is a place in the North Atlantic, between Greenland and Iceland, where the surface of the ocean begins to sink. It is not a dramatic place. There are no whirlpools, no visible signs. But it is here, in the cold months, that the warm water that has been flowing north from the tropics finally loses its heat, becomes dense, and — loaded with salt, made heavier still by the ice that has excluded it — begins to fall.

It falls for kilometres. It falls for months. And when it reaches the bottom, it begins a journey that will take it south along the Atlantic, east around Antarctica, into the Indian and Pacific Oceans, and eventually — after a thousand years — back to the surface and back to the north, to sink again.

This is the global conveyor. It carries heat from the equator to the poles, and it carries carbon and nutrients from the surface to the deep, and it delivers oxygen to the seafloor. Without it, Western Europe would be as cold as Labrador, and the deep ocean would slowly suffocate.

The vulnerability is fresh water. As Greenland melts — and it is melting, faster every year — it releases fresh water into the North Atlantic, and fresh water is less dense than salt water, so it floats, and it dilutes the surface layer that is supposed to become dense enough to sink. If the surface stops sinking, the pump stops, and the pump has stopped before: during the last ice age, a pulse of fresh water appears to have shut it down, and the result was a drop in North Atlantic temperatures of more than ten degrees within a few decades.

We do not know the exact threshold. We may not know it until we have crossed it. And when we have crossed it, the ocean will have to be re-plumbed, over the course of centuries, before the pump can start again, and by then the world that depended on it will have become a different world.

El Niño and the Southern Oscillation — the Pacific's mood swings

Peruvian fishermen noticed it first. They called it the Christ child, because it arrived at Christmas.

On the coast of Peru, in the nineteenth century, the fishermen began to notice a pattern. Every few years, the cold current that ran north along the shore — the current that brought the anchovies, the current on which their entire livelihood depended — disappeared, and the water warmed, and the fish went elsewhere, and the rain came, and for a season the desert bloomed and then flooded and then the rains stopped and the fishing failed. They called this current El Niño, the Christ child, because it usually arrived in December.

What they were watching, without knowing it, was one of the largest and most consequential climate oscillations on the surface of the earth. The tropical Pacific is not, as it appears, a stable body of water. It is the site of a coupled feedback loop between ocean and atmosphere so delicately tuned that a small change in one can amplify into a full flip of the other, and when it flips, the weather of the entire planet reorganises around it.

Drought in Australia. Floods in Peru. Failure of the monsoon in India. Mild winters in Canada. A redistribution of atmospheric energy so vast that it can be measured in ice cores taken from glaciers on the opposite side of the world.

The oscillation is not periodic. It has no fixed clock. It comes when it comes, every two years, every seven, sometimes with a pause of a decade, and no one can say in advance which year will be a strong one, because the system is chaotic, and chaos is not the absence of order, it is order of a kind no forecast can follow.

The fishermen of Peru named it after a child, which was an intuition of a profound kind, because what they were seeing was not a weather event but a birth, an arrival, something that comes into the world and changes it and then leaves, and returns, and never quite returns the same.

Ice-albedo feedback — the loop that melts itself

White reflects. Dark absorbs. This is the entire physics of the end of the Arctic.

The number is 0.85. That is the albedo of fresh snow — the fraction of sunlight it reflects back into space, unchanged. The number for open ocean is 0.06. The difference between these two numbers is not a curiosity. It is a mechanism, and the mechanism has been running for forty years, and it is one of the reasons the Arctic is now warming four times faster than the rest of the planet.

Here is what happens. Sunlight strikes the ice in summer, and the ice reflects it, and the ice stays cold, and everything is as it was. But as the climate warms slightly — a fraction of a degree, nothing dramatic — some of the ice melts, and the melt exposes dark water, and the dark water absorbs the sunlight instead of reflecting it, and the absorbed sunlight becomes heat, and the heat melts more ice at the edge of the remaining pack, and the newly exposed water absorbs more sunlight, and the loop runs, and it runs faster every year, and there is no brake.

What makes this so strange is that the system is not being pushed from outside. It is pushing itself. The melting is not just a result of the warming; it is a cause of more warming. The effect has become the cause, and the cause is producing more of the effect, and the two of them have formed a circuit, and the circuit is running, and there is no one at the switch.

There are countervailing loops. Melting ice adds fresh water to the North Atlantic, which can slow the thermohaline circulation, which can cool Europe — an old story, one that has happened before, at the end of the last ice age. Clouds may increase, and reflect more sunlight. But these are speculations, and the ice-albedo loop is not a speculation, it is arithmetic. Reflectivity is a number. The number is falling. The number will continue to fall as long as there is ice to melt.

The Arctic has been white for three million years. It is becoming blue. No one has to be blamed for this, and no one can be spared, and the process will not stop until the ice is gone, and then it will stop, and the planet will be a different planet.

Atmospheric circulation — the planet's heat engine

The Sahara is a desert not because it is dry but because the sky above it is falling.

Here is a fact that took meteorologists two centuries to fully accept: the Sahara is not a desert because it is far from water. It is a desert because the air above it is sinking. The same is true of the Arabian, the Mojave, the Atacama, the Australian outback. Every one of them sits at roughly thirty degrees latitude, and at thirty degrees latitude the great atmospheric cells that carry heat from the equator to the poles are descending, and descending air is compressing, and compressing air is warming, and warming air is dry, and so the deserts of the world are not accidents of geography. They are the necessary consequences of a heat engine.

At the equator, the sun heats the surface, the air rises, and as it rises it cools, and as it cools its moisture condenses, and so it rains, and the tropics are wet. But the air that has risen has lost its water, and it has to go somewhere, and so it moves poleward at high altitude, cooling further, until around thirty degrees it has become dense enough to sink. And as it sinks, it warms, and as it warms it holds more moisture, and it does not rain. So the desert forms. So the Sahara forms, and the Kalahari, and the Sonoran, and the Atacama. They are all the same desert. They are all the dry falling air.

Once you understand this, you understand that the map of the world's wet and dry places is not a fixed map. It is a map of where the air happens to be rising, and where it happens to be sinking, and if you change the temperature of the poles, the air will rise and fall in different places, and the map will change.

We are changing the temperature of the poles. The Arctic is warming three to four times faster than the tropics, and that means the temperature difference that drives the jet stream — the boundary between the warm and cold air masses — is shrinking, and the jet stream is weakening, and as it weakens it becomes more wavy, and a wavy jet stream stalls, it stops moving the weather, and a stalled weather pattern means a heat wave that lasts for weeks, a drought that does not end, a flood that will not stop.

The deserts will move. The rainforests will move. Nothing about the map of the living world is permanent. It is all just air, rising and falling, and the places where it rises have never been guaranteed to stay the same places.

The carbonate–silicate thermostat — the planet's slow climate control

The Earth has a thermostat. It takes a million years to turn on.

For two hundred years, geologists could not understand why the Earth had not frozen. The sun, over the course of the planet's history, has brightened by thirty percent — the early sun was a third dimmer than it is now — and yet the geological record shows liquid water on the surface of the Earth almost continuously for four billion years. Something must have been keeping the planet warm when the sun was weak, and cool when the sun was strong, and the something, it turned out, was rock.

Here is how it works. Carbon dioxide in the atmosphere dissolves in rain and forms a weak acid. That acid eats away at silicate rocks — granite, basalt — in a chemical reaction that consumes CO₂ and releases calcium and bicarbonate ions into rivers. The rivers carry the ions to the ocean, where marine organisms use them to build shells of calcium carbonate. When the organisms die, their shells sink and become limestone, and over hundreds of millions of years the limestone is subducted into the mantle and returned to the atmosphere as volcanic CO₂. The cycle closes.

The beauty of the system is that it is temperature-dependent. Warmer air holds more moisture, and more moisture means more rain, and more rain means more weathering, and more weathering means less CO₂, and less CO₂ means a cooler planet. Colder air holds less moisture, less rain, less weathering, more CO₂, warmer planet. The system is self-correcting. It is a thermostat.

But it is a thermostat made of stone, and stone moves slowly. The weathering that draws down CO₂ takes hundreds of thousands of years, sometimes millions. The full cycle from CO₂ to rock to CO₂ takes a hundred million years. Against a change that unfolds over centuries, the thermostat is powerless. It is not a thermostat in the human sense. It is a flywheel, and it will not save us, because it was never designed to respond at our speed.

The Earth will eventually recover from whatever we do to it. The carbon we release will be drawn down by rock, and buried, and the climate will return to something like a normal range, and the planet will be habitable again. This will take a million years, and everything we love will be gone, and the thermostat will still be there, slow, patient, indifferent, doing what it has always done.

Trophic cascades — the predator effect that reaches everything

In 1995, fourteen wolves were released into Yellowstone. The rivers changed shape.

In the winter of 1995, fourteen grey wolves were released into Yellowstone National Park. They had been absent from the park for seventy years, killed off in the early twentieth century by a government policy that considered them vermin, and their absence had been felt, though no one had realised it, in the shape of the rivers.

Without wolves, the elk had grown numerous. But the more important effect was not on the number of elk — it was on the behaviour of the elk. Wolves are ambush predators. They hunt in the open valleys and along the stream banks, where the elk cannot hide. In the absence of wolves, the elk had stopped avoiding these places. They had stayed in the valleys, and they had browsed, and they had browsed heavily on the willow and aspen saplings along the streams, and they had eaten them down to nothing.

What followed was a chain of consequences that no one had predicted. Without willows, the beavers disappeared, because beavers eat willows and use them to build dams. Without beaver dams, the streams lost their complexity — they straightened, they cut into their banks, the water ran faster and hotter. Without willows, songbirds that nest in willow thickets declined. The entire riparian ecosystem simplified, and no one noticed, because no one had been watching the willows.

When the wolves came back, the elk returned to the high country, where they could see them coming. The willows and aspen grew back. The beavers returned. The streams re-meandered. The songbirds came back. The wolf had done all of this not by eating the elk, but by changing where the elk ate.

This is what a trophic cascade is. It is not a food chain. It is a network of influence, in which a change at one level propagates through all the others, and the most powerful influence is often the one that is hardest to see — a change in behaviour, a shift in where an animal eats, an invisible alteration in the geometry of fear.

Ecological succession — how a system rebuilds itself

After the ice retreats, the rock is bare, and then, over centuries, it is a forest.

When a glacier retreats, it leaves behind a landscape that is, from the perspective of life, essentially sterile. Bare rock, glacial till, no organic matter, no nitrogen. In such a place, nothing should be able to live, and yet within a century the same landscape can be covered in alder and willow, and within three centuries it can be a mature forest, and the mechanism by which this happens is one of the most beautiful processes in biology, and it is called succession.

The first arrivals are lichens — a symbiosis of algae and fungi that can grow directly on rock. They exude acids, they trap dust, they die, and their remains become the first thin soil. Then mosses, which hold water and trap more dust. Then grasses and fireweed and small shrubs, whose roots break the soil further, whose leaves add organic matter, whose bodies create shade. Each stage creates the conditions for the next, and each stage eventually creates the conditions for its own replacement.

The pioneer grasses are outcompeted by the shrubs that grow in the soil the grasses built. The shrubs are shaded out by the fast-growing trees — birch, pine, aspen — that grow in the soil the shrubs enriched. And the fast trees are eventually outcompeted by the slow ones — oak, maple, beech — that can grow in the shade of the fast trees, and that will outlive them, and that will be the last stage, the "climax," although the climax is not one thing, it is whatever the climate and the soil and the disturbance regime will allow.

What is most remarkable about succession is that it is a system that engineers its own transformation. The stage that exists is always creating the conditions for the stage that will replace it. This is not competition in the usual sense. It is a form of ecological time, in which the present is always making the future, and the future is always erasing the present, and the whole process is driven not by any external force but by the accumulated logic of the organisms themselves.

When the process is disrupted — by fire, by flood, by landslide, by invasive species, by human intervention — it can be reset to an earlier stage. And when it is reset, it starts again, from the beginning, and the forest that returns is not always the forest that was there before.

Fire regimes — when destruction is a form of maintenance

A hundred years of putting out fires gave us the worst fires in a thousand years.

The ponderosa pine is a tree that expects to burn. It has a thick bark, a high canopy, and a habit of dropping its lower branches as it matures, and every five to twenty years, in the forests where it grows, a fire would come through, low and slow, burning the grass and the shrubs but leaving the big trees alive. This is the fire regime the pine evolved with. It is not an accident. It is a relationship, and it has been maintained for millions of years.

In the twentieth century, land managers decided that fire was an enemy. They would extinguish every fire, as quickly as possible, and in doing so they would protect the forests. And they did protect them — from fire — but they also protected them from the periodic burning that cleared the underbrush, and the underbrush grew, and the dead wood accumulated, and the ladder fuels that could carry a fire from the ground into the canopy built up, year after year, decade after decade, until the forest that had once been a fire-adapted system had become a fuel depot.

Then the fire came, and it did not burn low and slow. It burned high and fast, a crown fire, a firestorm, a fire that consumed the canopy and sterilised the soil and left behind not a forest but a moonscape. The ponderosa pine that had evolved with fire could not survive a fire like this. Nothing could.

This is what happens when you suppress a disturbance. The disturbance does not disappear. It accumulates. Every fire you put out is a fire you are saving up, and the interest compounds, and eventually the bill comes due, and the fire that finally burns is not the fire that would have happened in the natural regime. It is something larger, and hotter, and more destructive, and the forest that comes back from it is not the same forest.

Fire is not the enemy of the forest. Fire is one of the ways the forest maintains itself. To suppress fire is to interrupt the maintenance, and a system whose maintenance has been interrupted is not a protected system. It is a system in debt.

Wetlands — the kidneys, sponges and carbon vaults of the landscape

A bog is a landscape that has been building itself upward for ten thousand years.

There is a kind of landscape that is neither land nor water, and that is why it is so important. It is called a wetland, and it exists in the boundary between the two, in the saturated soil where the water table meets the surface, and because it is neither one thing nor the other, it does the work of both.

A wetland filters. The water that moves through it moves slowly, and as it moves the sediment settles out, and the nutrients — the same nitrogen and phosphorus that cause the blooms in system 01 — are taken up by the plants and the microbes, and buried in the sediment, and removed from the water. A wetland is a kidney. It cleans the water that passes through it, and it does so for free.

A wetland absorbs. The saturated soil and the dense vegetation slow the flow of water, and when a flood comes, the wetland spreads it out, and holds it, and releases it slowly over days and weeks. A river with a wetland upstream floods less than a river without one. This is why the removal of wetlands so often makes flooding worse.

And a wetland stores carbon. This is the strangest of its properties. Because waterlogged soil has no oxygen, the decomposition that normally returns dead plant material to the atmosphere is slowed to a crawl, and the peat accumulates, layer by layer, over thousands of years, until a single hectare of bog can hold more carbon than a hundred hectares of forest. When a peatland is drained, this carbon is released, and the process that took millennia to fill can be emptied in a few decades.

The great bogs of the north — the peatlands of Siberia, of Canada, of Ireland and Scotland — are among the largest carbon stores on the surface of the earth. They have been accumulating for ten thousand years. They are the landscape equivalent of a bank account that has not been touched since the last ice age. And we are draining them, and burning them, and building on them, and the carbon they have held for ten thousand years is going back into the air.

The places that matter most are often the places in between — the margin, the boundary, the mud. They do not look like much. They are not dramatic. They are not photogenic. They are simply the places where the system does its most important work.

Mycorrhizal networks — the underground economy of the forest

The forest is not a collection of trees. It is a trading floor, and the traders are fungi.

If you take a tree — a single tree, in a forest — and look at its roots, you will find that the roots are not alone. They are wrapped in, penetrated by, invaded by the filaments of a fungus, and the fungus is not a disease. It is a partner. It is the other half of a bargain that has been running for four hundred million years, since the first plants crawled out of the water and discovered that they could not survive on land without help.

The bargain is this. The fungus extends its filaments — the hyphae — far beyond the reach of the plant's roots, into the smallest pores of the soil, where the plant's roots cannot go, and it mines the phosphorus and the nitrogen and the water and the micronutrients that the plant cannot reach, and it delivers them to the plant. In exchange, the plant gives the fungus the sugars and lipids it has made through photosynthesis. Neither partner is doing the other a favour. It is a trade, and it is priced, and the price is enforced by the ability of each partner to withhold what the other needs.

But here is where it gets stranger. The fungal filaments do not stop at the roots of one tree. They connect. A single fungal individual can link the roots of many trees — of different species, of different ages — into a single network. Through this network, carbon, water, nutrients and chemical signals can move from where they are abundant to where they are needed. A seedling in deep shade can receive carbon from a larger tree. A tree under insect attack can send a chemical warning that triggers defensive responses in its neighbours.

This is not altruism. Fungi do not give freely. They charge what the market will bear, and they will withhold nutrients from plants that do not pay enough. Some plants — orchids, for example — cheat, taking carbon from the network without contributing. Some trees seem to punish the cheaters. It is an economy, and it has its own rules, and the rules are enforced by nothing more than the fact that each partner can walk away.

What this means is that a forest is not a collection of individuals. It is, in part, a single distributed system, connected at the roots by a network of fungi, trading in carbon and nitrogen and information, and no tree in the forest is quite as isolated as it appears. The tree is a node. The forest is the network. And the forest was already a network long before we learned to see it.

Coral symbiosis — a city built on a partnership

The largest structures built by any living thing are the work of two organisms, not one.

A coral polyp is a small animal. It looks, to the untrained eye, like a tiny anemone, a few millimetres across, with tentacles and a mouth, and on its own it can do nothing remarkable. It can catch small prey, and it can build, slowly, a small skeleton of calcium carbonate. But it cannot build a reef. It cannot build a structure visible from space. It cannot build the largest biological structure on the surface of the earth.

What allows it to do that is a partner. Inside the tissues of every reef-building coral are single-celled algae called zooxanthellae — millions of them per square centimetre, packed into the coral's cells like a green fog. The algae photosynthesise, using the sunlight that reaches the shallow water, and they make sugar, and they pass most of that sugar to the coral. In return, the coral gives them shelter, and carbon dioxide, and the nitrogenous waste of its own metabolism. It is a trade, and the trade is so efficient that the coral can grow fast enough to build a limestone city, and the city can support a quarter of all the species in the ocean.

The partnership is not, however, unconditional. The zooxanthellae can only photosynthesise within a narrow range of temperatures, and when the water gets too warm — one degree above the summer maximum, two degrees — the algae begin to produce damaging levels of oxygen radicals, and the coral, in a kind of immune response, expels them. The coral turns white. The colour was the algae. The coral is not dead, but it is starving, because it has lost the partner that provided ninety percent of its energy, and if the water does not cool within a few weeks, the coral dies.

In the last thirty years, the water has begun to warm more often, and more severely, and the bleaching events have come closer together, so that the reefs no longer have time to recover between them. The Great Barrier Reef has bleached four times since 1998, and each time the recovery has been slower, and the coral that has come back has been less diverse, and the reef has been, quietly, dying, not from a single blow but from an accumulation of blows, each one of which it could have survived on its own.

When the partnership breaks, the whole structure goes with it. The city, the habitat, the biodiversity, the fishing economy, the tourism, the coastline that the reef protected from storms — all of it depends on the survival of a partnership between an animal and a plant, and the animal and the plant can only remain partners within a narrow band of temperature, and we are leaving that band.

Permafrost carbon feedback — the frozen archive that is waking up

Beneath the tundra is a vault. It holds twice the carbon of the atmosphere. It is opening.

There is a layer of ground beneath the Arctic that has been frozen, without interruption, since the last ice age. It is called permafrost. It covers a quarter of the Northern Hemisphere's land surface — Siberia, Alaska, northern Canada, the Tibetan Plateau — and for tens of thousands of years it has been doing something remarkable: it has been preserving organic matter that would otherwise have rotted away long ago.

When a plant dies in the Arctic, it does not decompose the way a plant in the tropics would. The cold slows the microbes almost to a stop. The organic matter accumulates, layer by layer, over millennia — leaves and roots and bone and the bodies of mammoths and the remains of forests that grew here when the climate was warmer than it is now — and then it freezes, and it stays frozen, and the carbon it contains is locked in the ice.

The amount of carbon in this frozen ground is between 1,400 and 1,600 billion tonnes. That is roughly twice as much as is currently in the atmosphere. It has been accumulated over tens of thousands of years, by the slow process of life and death in the far north, and it has been sitting there, in the ground, undisturbed, since the end of the last ice age.

Now the Arctic is warming, faster than anywhere else on the planet, and the permafrost is thawing. And when it thaws, the microbes that have been dormant in the frozen soil wake up, and they begin to eat. They eat the organic matter that has been preserved for ten thousand years, and they release the carbon as CO₂ and as methane — methane being a gas that is eighty times more potent than CO₂ over a twenty-year period.

This is not a projection. It is happening now. The ground is collapsing in Siberia, where the thaw has created craters and lakes that did not exist before, and the soil is releasing its carbon into the air, and the air is getting warmer, and the warmer air is thawing more permafrost, and the process feeds itself. It is a vault that has been sealed for an ice age, and the seal is breaking, and once the vault is open it cannot be resealed, not on any timescale that matters to us, not by any means we have.

The biological pump — how tiny life moves carbon to the deep

The deep ocean is filled with the bodies of things too small to see.

In the top hundred metres of the ocean, where the sunlight reaches, there are creatures so small that a single drop of seawater can hold thousands of them. They are called phytoplankton, and they are the base of the marine food web, and they are also the beginning of a machine that moves carbon from the sky to the abyss.

Here is how it works. The phytoplankton photosynthesise, pulling dissolved CO₂ out of the seawater and turning it into organic matter. When they die, or when they are eaten and the faecal pellets are produced, some of that organic carbon begins to sink. It falls through the water column as what oceanographers call marine snow — a slow drift of dead cells, mucus, faecal pellets, and other debris, descending through the dark, one particle at a time.

Most of this marine snow does not reach the bottom. Ninety-nine percent of it is consumed by bacteria and zooplankton on the way down, and the carbon it contains is respired back into dissolved CO₂ in the mid-ocean. But a fraction — somewhere between one and twenty percent — reaches the deep, and that fraction is the pump's net effect.

What it does is this: it takes carbon that was in the surface layer, where it could exchange with the atmosphere, and it moves it into the deep ocean, where it can be stored for centuries or millennia before the circulation brings it back. Without this pump, the atmosphere would hold significantly more CO₂ than it does. The pump is one of the reasons the ocean is the largest carbon sink on the planet, and the pump is made of plankton.

As the ocean warms, the surface layer becomes lighter, and it mixes less with the nutrient-rich water below, and the phytoplankton that depend on those nutrients begin to starve, and the pump weakens, and the carbon that the pump was moving to the deep begins to stay in the surface, and the atmosphere begins to hold more of it, and the warming accelerates. The pump is not a thing we built. It is a thing that was here before us, and it has been running for hundreds of millions of years, and we are slowing it down, not by intention, but by the simple fact of warming the water.

The ozone layer — the shield we damaged and repaired

We tore a hole in the sky, and then we sewed it shut.

In 1985, a team of British scientists published a paper in Nature that should have been impossible. They had been measuring ozone over Antarctica for years, and their instruments had been telling them, since the late 1970s, that the ozone concentration above the South Pole was dropping, and that in the Antarctic spring it was dropping by as much as forty percent, and they had not believed their instruments. They had checked them. They had recalibrated them. They had assumed that there was a fault. And there was no fault. The sky over Antarctica was opening.

The cause, it turned out, was a class of chemicals called chlorofluorocarbons, or CFCs — compounds that had been used, for decades, as refrigerants, as propellants in aerosol cans, as blowing agents for foam. They were, in every ordinary sense, miraculous: non-toxic, non-flammable, chemically inert. They were so inert that nothing in the lower atmosphere could break them down. They rose, slowly, into the stratosphere, where the intense ultraviolet radiation finally tore them apart, and the chlorine atoms that were released — a single atom from a single CFC molecule — began to destroy ozone, catalytically, one chlorine atom destroying tens of thousands of ozone molecules before it was finally removed.

The hole was not a hole in the sky. It was a thinning, a reduction, a place where the shield that had protected the surface of the earth from ultraviolet radiation for a billion years was being dismantled, atom by atom, by a chemical that had been sold as a miracle.

And then something remarkable happened. In 1987, two years after the discovery of the Antarctic hole, the nations of the world signed the Montreal Protocol, an agreement to phase out the production of CFCs and other ozone-depleting substances. It is the only environmental treaty in history that has been ratified by every country on Earth. It worked. The concentration of CFCs in the atmosphere peaked in the 1990s and has been declining ever since. The ozone layer is expected to recover by the middle of this century.

This is the only story of its kind. It is the story of a global environmental problem that was identified, understood, and solved. And the reason it could be solved — and the reason it stands alone — is that the problem was simple, the cause was clear, and there was a single chemical to ban. The other systems in this document do not have that shape. They have causes so distributed that no single act can address them. And that is why the ozone layer is not a template. It is an exception, and the exception proves the rule.