WELCOME to Icehouse Earth. It may not feel like it but, right now, our planet is in an ice age. It started about 2.6 million years ago and, until recently, showed little sign of letting up. In the 1970s, scientists were even worried that we were about to plunge into another full-blown icy spell.
Today, those fears have evaporated into a fog of greenhouse gases. Unless we do something, fast, the exact opposite is going to happen. If emissions continue to accumulate in the atmosphere, Earth will blow its cool, with potentially disastrous consequences for humanity and the living world.
As the climate hots up, so does the race to understand what really happens when we crank up the thermostat. The standard approach is computer modelling, but we need every insight we can get, which is why some climatologists are turning their attention to the deep past, searching for global warming events to help predict the future. The good news is that the biosphere has endured some very hot periods and lived to tell the tale. The bad news is that the next hothouse may be more extreme than anything Earth has experienced before. In which case, it really is goodbye, cool world.
Advertisement
The earliest inklings that Earth’s climate was radically different in the past came in the 1800s, when geologists were stumped by phenomena such as glacial deposits and desert sandstones at temperate latitudes. In the early 20th century, the theory of continental drift appeared to offer an explanation – maybe the deposits had been laid down elsewhere and inched their way to their current positions. But in the absence of accurate temperature records, nobody could be entirely sure.
“Reliable temperature records only go back a couple of hundred years, so we have to find proxies,” says Tracy Aze, a marine micropalaeontologist at the University of Leeds, UK. Proxies are things that are affected by the temperature at the time and that are preserved over very long periods, creating a record of ancient temperatures. “These are typically chemical traces captured in sedimentary rock or ice, which allow us to make inferences about [past] ambient temperatures,” says Aze.
The first such measure was invented in the late 1940s by chemist Harold Urey, who also won a Nobel prize and worked on the first atomic bomb as part of the Manhattan Project. He realised that the , known as isotopes, in rocks and fossils was a proxy for ancient temperatures. In general, the more of a heavier isotope of the element – oxygen-18 – in oceanic sediment, the colder the temperature at the time it formed. Since then, numerous similar techniques have been invented. Aze, for example, analyses the fossilised shells of marine organisms called foraminifera. Oxygen isotope ratios in the calcium carbonate from the shells reveal the temperature of the water in which they formed. Another technique is leaf margin analysis, which gives a proxy for temperatures on land. Put simply, the smoother the edge of a fossil leaf, the warmer the air it grew in. “We try to use as many different lines of evidence as possible,” says Aze.
Using these techniques and others, palaeoclimatologists have painstakingly assembled a reconstruction of Earth’s climate stretching back tens of millions of years. The further back you go, the less accurate the estimates, says Aze. However, we have a pretty good handle on the past 65 million years, a period that runs from just before the sizzling time of the Eocene “hothouse”. Things have, by and large, cooled steadily since then (see “Hot histories”).
What started life as a backward-looking historical science has quickly become much more future-facing. Geologists have a mantra: the present is the key to the past. For palaeoclimatologists, the past is key to the future. They see periods of intense warming in the past as potential analogues of a world warmed by increasing greenhouse gas emissions. “The geological past is a guide to what future climates might be like,” says Alan Haywood, an earth scientist at the University of Leeds in the UK.
“In the last interglacial, the rate of sea level rise peaked at 2.5 metres per century”
Our carbon emissions have already heated the planet by 1°C since pre-industrial times and we are on course to add at least another 0.5°C – or more likely 2°C or 3°C – by the end of this century. Understanding what effect this will have on Earth, on life and on human civilisation is one of the most urgent scientific questions we face. It is also very difficult to answer. The climate system is devilishly complex: an intricate and non-linear interplay of solar radiation, atmospheric dynamics, geography, geology, ocean circulation, ice and the biosphere.
The standard approach is to use computer models that simulate the Earth system in as much detail as possible, and then toggle various parameters – such as adding loads of carbon dioxide – to predict what might happen. The models, which don’t rely on ancient data, are increasingly realistic, as long as you don’t want to go beyond the next 300 years or so. But they fail to capture the full, long-term effects of warming. “There are feedbacks that only work on longer timescales,” says Hubertus Fischer, a palaeoclimatologist at the University of Bern in Switzerland. “Ocean overturn has a timescale of 500 to 1000 years; ice sheets even longer. You need a couple of thousand years for Earth to reach a new equilibrium.”
That is where palaeoclimates come in. “It’s really only the palaeo data that has those timescales,” says Fischer. “It gives us good observational evidence of the long-term processes – not just things like temperature, precipitation and sea ice that respond quickly, but how all the components of the Earth system respond to warming.”
So what do past warming episodes tell us about what is in store?
First, a warning. Each past warming period has unique features, so there is no such thing as the perfect analogue for future change. What’s more, current global warming is happening way faster than anything Earth has ever mustered of its own accord. For example, the warmest period of the past 65 million years, the Eocene hothouse, seems to have been caused by pulses of CO2 emitted over thousands of years. Our pulses are happening over decades.
Another important consideration when searching for analogues is that the continents should be roughly in the positions they are today. Physical geography is a key component of the climate system, affecting ocean currents and weather patterns. The placement of Antarctica is especially important: its huge ice sheets, massive power to reflect sunlight off its surface and isolation from other landmasses all exert a disproportionate influence. Antarctica arrived at its current position about 100 million years ago, so anything before that has serious question marks over it. “You can try to correct for the different locations of continents in the past in climate models,” says Fischer. “But the further you go back, the more of a stretch it is.” That all but rules out some very ancient warming periods, notably the late Cretaceous and end-Permian as analogues.
Analogues of the future
Despite these caveats, there is no shortage of hotter eras as candidates. The most recent warming period is the Mid-Holocene Thermal Maximum, which began around 11,000 years ago and lasted 6000 years. This has the advantage of modern physical geography and ample fossil data, but the average temperature was only about 0.7°C above the pre-industrial baseline. Given that we have already overshot this, it isn’t an especially illuminating window on the future.
Going further back, the next spike is the last interglacial, around 129,000 to 116,000 years ago. As the name suggests, interglacials are temporary respites within an ice age. The last interglacial is recent enough to pass the physical geography test, but average global temperature was only about 0.8°C above pre-industrial levels. This is enough to rule it out as an analogue, although it does tell us something hugely important. Towards the end of the last interglacial, sea levels had risen by up to 9 metres, inundating wide areas of what today is dry land. The source of all this water has long been debated, but last year glaciologists at Oregon State University announced that it was almost certainly the covering West Antarctica. This has since grown back, but it gives today’s climatologists sleepless nights: its base is below sea level, which exposes it to warming ocean waters and makes it susceptible to sudden melting. That is apparently what happened in the last interglacial, with geological records suggesting that the rate of sea level rise peaked at 2.5 metres per century.
For the next candidate analogue, you have to go even further back, to a 3-million-year-old episode called the Mid-Pliocene Warm Period. During the Pliocene epoch, gradual cooling from the Eocene was still the dominant direction of travel, but a 300,000-year period significantly bucked that trend. For unknown reasons, CO2 levels in the atmosphere soared and global temperatures followed. The continents weren’t quite in their modern configuration: North and South America had yet to be joined by the Isthmus of Panama – an event that set in motion the Gulf Stream, an Atlantic Ocean current that plays a crucial role in moving heat around the planet. Even so, palaeoclimatologists see the Mid-Pliocene as a pretty good analogue of the future Earth. “It’s the first interval which really works,” says Fischer. “CO2 is about the same as today, so that is a direct comparison.” But, he says, even though it was only 3 million years ago, the palaeo figures are a little sparse – there are “only a couple of handfuls of records”.
Nonetheless, we can say with some confidence what the world was like back then. “The global annual mean temperature was between 2°C and 3°C warmer than pre-industrial, with warming more pronounced in higher latitudes,” says Haywood. “There was a significant reduction in sea ice in both hemispheres. The Arctic may have been completely free of ice in the summer. The Greenland ice sheet may have retreated to a small ice mass and we believe that the West Antarctic ice sheet was not there. Sea level is somewhere between 15 to 25 metres above modern levels.”
“In as little as 11 years, our climate will have shifted to a state not seen for millions of years”
Sound familiar? To a large extent this echoes warnings about where our world is headed. “Yes, you do think, hmmm, I’ve heard this before somewhere…” says Haywood. But the devil is in the detail. Warming at high northern latitudes was around 8°C – much more than our best climate models predict at these CO2 levels. This suggests the models may substantially underestimate how much the Arctic will warm in the long run, says Fischer. The Pliocene also provides some much-needed clarity about what will happen to ice sheets and sea levels as the effects of warming play out on longer timescales. In a nutshell, they melt fast and rise a lot.
For seasoned climate-watchers, this may not come as any surprise. The Pliocene has long been touted as an analogue of the 22nd century and beyond, says Haywood. But when he and his colleagues did a formal statistical comparison of climatic similarities between the Pliocene and various computer models of the future, . Dividing Earth into a grid and focusing on temperature and precipitation, their analysis revealed that under current emissions trajectories “we very quickly arrive at a Pliocene scenario”, he says. How quickly? “Within the next 20 to 30 years. There is a little window of uncertainty, but one of the models shows us getting to the Pliocene as early as 2030. And here we are in 2019… when I first saw those results, I couldn’t speak for a while. It pretty much blew my mind.”
That doesn’t mean the imminent loss of ice sheets and huge sea level rises – those will still take centuries to play out – but it does mean that in as little as 11 years, our climate will have shifted to a state not seen for millions of years.
It isn’t all doom and gloom. “Many of the species in the Pliocene are around today, so the fact that we might go back there doesn’t necessarily mean we’re going to lose a huge number of species,” says Haywood. Nevertheless, he acknowledges that the current rate of change might exceed nature’s adaptive capabilities. Besides, the Pliocene was essentially a pristine natural ecosystem, which isn’t now the case. “We have lost very large areas of natural vegetation and habitat so there are additional stresses that wouldn’t have been there in the Pliocene. That’s where you get into the realm of ecosystem resilience. I know the biologists are deeply concerned about this,” he says.
“Humans have the potential to push Earth beyond any extreme that we know of from the past”
However, the Pliocene may not be our final destination. If we don’t substantially curb emissions, says Haywood, we soon leave that epoch behind and strike out for the Eocene, an extreme hothouse episode when average temperatures were 14°C above pre-industrial and sea levels were more than 70 metres higher. That is a long way off – the Eocene becomes the closest analogue of Earth’s climate no sooner than 2150. But the journey won’t be comfortable. “If we stay below 2°C of warming, we can most likely avoid uncontrollable thresholds,” says Fischer. “But if we go to 4.5°C or 6°C, we can’t say we avoid them.” Judging from what happened in the past, he says, that might include catastrophic sea level rise, natural releases of even more greenhouse gases and a sudden collapse of the Gulf Stream. “This will shake our climate system so much we cannot foresee the consequences,” he says.
Novel climates
If you think that sounds bad, the palaeoclimatologists still have a joker up their sleeve. “There’s a growing realisation that our climate trajectory is taking us into places that are not well understood,” says Haywood. “We pretty quickly get into the realms where temperature and precipitation do not match any of our geological references.” These hot and saturated “geologically novel climates” don’t cover the entire Earth, but by the 23rd century are found across east and South-East Asia, northern Australia and the coastal Americas.
In other words, human influence on the climate has the potential to push Earth beyond any extreme that we know of from the past. “We see that as a bit of a wake-up call,” says Haywood. So does the Intergovernmental Panel on Climate Change (IPCC), which looked at palaeoclimatology for its recent report on the chances of keeping warming below 1.5°C. It was alarmed by what it saw. “Palaeoclimatology is an important source of evidence,” says Sonia Seneviratne, a climate scientist at the Swiss Federal Institute of Technology in Zurich and lead author of the IPCC report. “This evidence shows that the present rate of warming is unprecedented on the scale of millennia.”
In human terms, 2150 isn’t far off. Nobody alive today can expect to see it, but our grandchildren may. In geological terms, it is even shorter. It took 50 million years for natural processes to undo the warming of the Eocene. We may reverse that in two centuries, with terrible consequences. Call it a lesson from prehistory.
Cycles of change
Palaeoclimatologists look for clues about past temperatures in rocks and fossils. The story these tell is of a starting around 50 million years ago (see “Hot histories“). At that time – the early Eocene, 15 million years after the dinosaurs went extinct – Earth was sweltering. Average temperatures were . There was no permanent ice anywhere on the planet and sea levels were more than 70 metres higher than today. The cause of this extended heatwave appears to have been a massive injection of greenhouse gases into the atmosphere. The source remains unclear – volcanic eruptions and comet impacts have been proposed – but whatever caused it, carbon dioxide levels (ppm). For comparison, they are 414 ppm today and were about 280 ppm in pre-industrial times.
From that high point, Earth gradually cooled, as slow-moving geological processes scrubbed CO2 out of the atmosphere. One key factor was mountain building. New mountains expose silicate rocks, which extract CO2 from the atmosphere as they weather. Around 34 million years ago, it was chilly enough for ice sheets to appear at the South Pole. Another 13 million years on, the North Pole iced up too. The downward trend bottomed out just over 1.2 million years ago. Since then, the globe has warmed a little, but remains firmly in icehouse conditions, with ice caps at both poles.
However, this is far from the whole story. Superimposed on the general cooling trend is a constant and much more rapid oscillation from warm to cold and back again. This is mainly driven by natural fluctuations in Earth’s orbit around the sun, which alter the distribution of solar radiation hitting the planet’s surface. The variations are small, but can have big long-term effects. If less sunlight hits northern latitudes, for example, regional temperatures fall and winter snowfall rises. Weaker sunlight also means that the snow survives the summer months, which boosts the reflectivity – or albedo of the surface, bouncing more sunlight back into space and amplifying the cooling. “That’s a self-reinforcing effect,” says Hubertus Fischer at the University of Bern, Switzerland. “It gets colder, you get more snow, you build up ice masses and that leads to a glaciation.”
Cooling induced by orbital variations also causes CO2 levels in the atmosphere to fall. “The greenhouse effect becomes weaker, leading to further cooling,” says Fischer. Over tens of thousands of years, this self-reinforcing feedback is enough to flip Earth from warm to cool. But the planet’s orbit eventually swings back and the sequence of events reverses.
This process nudges global average temperatures up and down, with a periodicity of around 100,000 years. These are the Milankovitch cycles, named after Serbian astronomer Milutin Milankovitch who proposed them in the 1920s. Occasionally, something else kicks in, turbocharging the process. That may be the confluence of several different orbital eccentricities, or an injection of greenhouse gases. Whatever the cause, Earth experiences unusually strong warming. These are the periods of particular interest to people trying to predict what Earth’s climate will be like in the future.
