Chapter 6

Climate surprises

We are changing the composition of the atmosphere beyond anything that has been experienced over the past 3 million years. We are headed for unknown territory and therefore scientific uncertainty can be great. We know from the study of past records that the climate system can switch into a new state very quickly once a threshold has been passed. For example, ice-core records suggest that half the warming in Greenland at the end of the last ice age was achieved in only a few decades. This chapter examines the possibility that there are thresholds or tipping points in the climate system that may occur as we warm the planet. Figure 29 shows the main tipping points which scientists have been concerned about over the last three decades. Irreversible melting of the Greenland and/or western Antarctic ice sheet, slowing down of the North Atlantic deep-ocean circulation, massive release of CH₄ from melting gas hydrates, and the Amazon rainforest dieback will all be discussed.

29. Tipping points in the climate system.

Thresholds and tipping points

The relationship between a climate forcing factor such as GHGs and the climate response is complicated. In an ideal world it would be a simple relationship with little or no delay, but we already know that there is inertia in the climate system, so that it responds to GHG forcing with a 10- to 20-year delay, depending on how much is being emitted. So we can examine the way different parts of the climate system respond to climate change with four scenarios (see Figure 30):

30. Greenhouse gas forcing and the response of the climate system.

(a) Linear but delayed response (Figure 30a). In this case, the increase in GHGs produces a delayed but direct response in the climate system whose magnitude is in proportion to the additional forcing. This can be equated to pushing a car along a flat road. At first nothing happens as friction must be overcome before the car will start moving. Once that has happened most of the energy put into pushing is used to move the car forward. An example of this is the heating up of the oceans which has a delay due to the inertia of heating such a large body of water.

(b) Muted or limited response (Figure 30b). In this case, the GHG forcing may be strong, but the relevant part of the climate system is in some way buffered and therefore gives very little response. This is analogous to pushing the car up a hill: you can spend as much energy as you like trying to push the car, but it will not move very far. An example of this is the east Antarctic ice sheet, which has been stable at much warmer temperatures than today.

(c) Delayed and non-linear response (Figure 30c). In this case, the climate system may have an initial slow response to the GHG forcing but then respond in a non-linear way. This is a real possibility when it comes to climate change if we have underestimated the positive feedback in the system. This scenario can be equated to the car near the top of a hill: it takes some effort and thus time to push the car to the very top of the hill; this is the buffering effect. Once the car has reached the peak, it takes very little effort to push the car over it, and then it accelerates down the hill with or without help. Once it reaches the bottom, the car then continues for some time—the overshoot—and then it slows down of its own accord and settles into a new state.

(d) Threshold response (Figure 30d). In this case, initially, there is very little response in that part of the climate system to the GHG forcing. However, on reaching a threshold, all the response takes place in a very short period of time, in a single, large step. In many cases, the response may be much greater than one would expect from the size of the forcing, and this can be referred to as a ‘response overshoot’. This scenario equates to the bus hanging off the cliff at the end of the original film The Italian Job: as long as there are only very small changes, nothing happens at all. However, a critical point (in this case weight) is reached and the bus (and the gold) plunges off the cliff into the ravine below. An example of this could be the Greenland ice sheet that has started to melt; the melting could suddenly accelerate, causing a catastrophic collapse.

There are also situations in which a threshold becomes a tipping point. You can think of a threshold as a point at which there is change in a system that can be reversed. But a tipping point is a threshold that, when crossed, means the system moves into a new state and this transition is irreversible. An added complication when assessing whether climate change will create a simple threshold or a tipping point are bifurcations within the climate system. This means the forcing required to push the climate system one way across the threshold is different from the reverse. This implies that once a climate threshold has been passed, it is a lot more difficult to reverse it and in some cases it may in effect be irreversible.

The term ‘tipping points’ is used a lot in climate change research and discussions. However, care must be taken as there are two usages of this word. First, there are references to climate tipping points, which are the large-scale, irreversible shifts in the climate system, such as irreversible melting of ice sheets or the release of huge stores of CH₄ from below the oceans. The other usage concerns societal tipping points, which occur when climate change has a major effect on a region or a particular country. For example, a 200 mile (~322 km) shift northward of the South-East Asian monsoonal rainfall belt is, in climatological terms, a small shift, not a fundamental climate tipping point. But for the countries where the rains no longer fall or those where it does for the first time, such a shift is a major climate tipping point, because their weather may have been permanently altered.

Melting ice sheets

The IPCC projections for sea-level rise by 2100, if there are no significant curbs to carbon emissions, are between 0.50 m and 1.3 m. The largest uncertainty within these estimates is the contribution that the melting of Greenland and Antarctica will make by the end of the century. At the moment it is estimated that Greenland is losing over 230 Gt of ice per year, a seven-fold increase since the early 1990s. Antarctica is losing about 150 Gt of ice per year, a five-fold increase since the early 1990s. Most of this loss is from the northern Antarctic peninsula and the Amundsen sea sector of west Antarctica. Greenland and Antarctica together constitute one of the most worrying potential climate surprises. If the large ice sheets there melted completely, their contribution to global sea-level rise would be as follows: Greenland, about 7 m; the west Antarctic ice sheet, about 8.5 m; and the east Antarctic ice sheet, about 65 m. These compare with just 0.3 m if all the mountain glaciers melted. Palaeoclimate data show that the huge east Antarctic ice sheet developed 35 million years ago due to the progressive tectonic isolation of Antarctica and that it has in fact remained stable in much warmer climates. So climate scientists have a very high degree of confidence that the east Antarctic ice sheet will remain stable in this century.

However, scientists are very worried that the melting of Greenland or west Antarctica could significantly accelerate in the next 100 years. Even if we have already started the processes of melting the whole of these ice sheets, there is a physical constraint to the speed at which the ice can melt. This is due to the time it takes for heat to penetrate the ice sheets. Imagine dropping an ice cube into a hot cup of coffee. You know it will melt entirely, but it takes time for the heat to penetrate to the middle of the ice cube. Most of the ice from ice sheets flows through ice streams to get to the sea, and there is a limit on how much ice these ice streams can transport. The worst-case scenario, according to leading glaciologists, is that these ice sheets could add between 1 m and 1.5 m to the sea level by the end of the century, which would threaten many coastal populations around the world. There is also scientific debate about what happens to both the Greenland and Antarctic ice sheets beyond the next 100 years. Even if significant melting does not occur this century, we may have started a process that causes irreversible melting during the next one. Our carbon emissions over the next few decades could determine the long-term future of the ice sheets and the livelihoods of billions of people who live close the coast.

Deep-ocean circulation

The circulation of the ocean is one of the major controls on our global climate. In fact, the deep ocean is the only candidate for driving and sustaining internal long-term climate change (of hundreds to thousands of years) because of its volume, heat capacity, and inertia. In the North Atlantic, the north-east trending Gulf Stream carries warm and salty surface water from the Gulf of Mexico up to the Nordic seas. The increased saltiness, or salinity, in the Gulf Stream is due to the huge amount of evaporation that occurs in the Caribbean, which removes moisture from the surface waters and concentrates the salts in the sea water. As the Gulf Stream flows northward, it cools down. The combination of a high salt content and low temperature increases the density of the surface water. Hence, when it reaches the relatively low saline oceans north of Iceland, the surface water has cooled sufficiently to become dense enough to sink into the deep ocean. The ‘pull’ exerted by the sinking of this dense water mass helps maintain the strength of the warm Gulf Stream, ensuring a current of warm tropical water continues to flow into the north-east Atlantic, sending mild air masses across to the European continent. It has been calculated that the Gulf Stream delivers 27,000 times the energy of all of Britain’s power stations put together. If you are in any doubt about how good the Gulf Stream is for the European climate, compare the winters at the same latitude on either side of the Atlantic Ocean, for example London with Labrador, or Lisbon with New York. Or a better comparison is between Western Europe and the west coast of North America, which have a similar geographical relationship between the ocean and continent—so think of Alaska and Scotland, which are at about the same latitude.

The newly formed deep water sinks to a depth of between 2,000 m and 3,500 m in the ocean and flows southward down the Atlantic Ocean, as the North Atlantic Deep Water (NADW). In the South Atlantic Ocean, it meets a second type of deep water, which is formed in the Southern Ocean and is called the Antarctic Bottom Water (AABW). This is formed in a different way to NADW. Antarctica is surrounded by sea ice and deep water forms in coast polynyas, or large holes in the sea ice. Out-blowing Antarctic winds push sea ice away from the continental edge to produce these holes. The winds are so cold that they super-cool the exposed surface waters. This leads to more sea-ice formation and salt rejection, producing the coldest and saltiest water in the world. AABW flows around the Antarctic and penetrates the North Atlantic, flowing under the warmer, and thus somewhat lighter, NADW (see Figure 31a). The AABW also flows into both the Indian and Pacific Oceans.

31. Deep-ocean circulation changes depending on freshwater inputs.

This balance between the NADW and AABW is extremely important in maintaining our present climate, as not only does it keep the Gulf Stream flowing past Europe, but it also maintains the right amount of heat exchange between the Northern and Southern Hemispheres. Scientists have shown that the circulation of deep water can be weakened or ‘switched off’ if there is sufficient input of fresh water to make the surface water too light to sink. This evidence has come from both computer models and the study of past climates. Scientists have coined the phrase ‘dedensification’ to mean the removal of density by the addition of fresh water and/or warming of the water, both of which prevent sea water from being dense enough to sink. As we have seen, there is already concern that global warming will cause significant melting of the polar ice caps. This will lead to more fresh water being added to the polar oceans. Climate change could, therefore, cause the collapse of the NADW, and a weakening of the warm Gulf Stream (Figure 31b). This would cause much colder European winters and more severe weather. However, the influence of the warm Gulf Stream is mainly seen in the winter and has only a small effect on summer temperatures. So, if the Gulf Stream fails, global warming would still cause European summers to heat up. Europe would end up with extreme seasonal weather, very similar to that of Alaska.

A counter-scenario is that, if the Antarctic ice sheet starts to melt significantly before the Greenland and Arctic ice, things could be very different. If enough melt-water goes into the Southern Ocean, then AABW will be severely curtailed. Since the deep-water system is a balancing act between NADW and AABW, if AABW is reduced then the NADW will increase and expand (Figure 31c). The problem is that NADW is warmer than AABW and, because if you heat up a liquid it expands, the NADW will take up more space. So any increase in NADW could mean a rise in sea level. Computer models by Dan Seidov (now at the National Oceanic and Atmospheric Administration) and myself have suggested that such a scenario would result in an average sea-level increase of over 1 m.

It has been over 30 years since the possibility of a catastrophic shut down of the deep-ocean circulation was suggested, and there has been a huge amount of work on it. Monitoring has shown that the Gulf Stream has weakened by 15% since the middle of the last century. Evidence collated from this ocean monitoring and climate model predictions of the future in the very latest IPCC report suggest that collapse of the Gulf Stream is highly unlikely in the 21st century. The models do, however, show a significant weakening in the overturning of the North Atlantic in this century, especially in the high-emission scenarios, and the problem is that we do not know where a potential tipping point leading to shut down of deep-ocean circulation might be. Moreover, if the melting of Greenland or western Antarctica accelerates, then huge amounts of fresh water could enter the oceans, significantly disrupting deep-ocean circulation.

Gas hydrates

Below the world’s oceans and permafrost is stored a large amount of carbon in the form of CH₄. The CH₄ gas is trapped in a solid cage of water molecules at low temperatures and/or high pressures. The CH₄ gas comes from decaying organic matter found deep in ocean sediments and in soils beneath permafrost (Figure 32). These gas hydrate reservoirs could be unstable, as an increase in temperature or decrease in pressure would cause them to destabilize and to release the trapped CH₄. Climate change is warming up both the oceans and the permafrost, threatening the stability of gas hydrates. CH₄ is a strong GHG, twenty-one times more powerful than CO₂ (see Table 1). If enough were released, it would raise global temperatures, which could lead to the release of even more gas hydrates—producing a runaway effect. Scientists really have no idea how much CH₄ is stored in the gas hydrates beneath our feet: estimates are of between 1,000 and 10,000 Gt (compared with ~800 GtC currently in the atmosphere), which is a huge range. Without a more precise estimate, it is very difficult to assess the risk posed by gas hydrates.

32. Gas hydrates in a marine setting.

The reason why scientists are so worried about this issue is because there is evidence that a super-greenhouse effect occurred 55 million years ago, during what is called the Palaeocene–Eocene Thermal Maximum (PETM). During this hot-house event, scientists think that up to 1,500 Gt of gas hydrates may have been released. This huge injection of CH₄ into the atmosphere accelerated the natural greenhouse effect, producing an extra 5°C of warming. There is still, however, considerable debate over the PETM. For example, was it gas hydrate CH₄ release or CO₂ release from a phase of massive volcanism occurring around the same time that was the main cause of the warming?

The current consensus is that the ocean reserves of gas hydrate are likely to remain stable this century. Gas hydrates form a solid layer at the bottom of the ocean. The depth of this layer is controlled by the geothermal heat gradient—as you go deeper in the sediment, the temperature rises at about 30˚C/km. At a certain depth, it is too warm for gas hydrates to exist and CH₄ collects there as free gas in the sediment. As ocean temperatures change, the temperature change has to be transmitted through the solid gas hydrate layer to the lower boundary for some of it to melt. If this process is slow enough, the gas released migrates upwards in the ocean sediment column and refreezes at a higher level. However, if carbon emissions are not curbed, then by the next century we could see this process speed up, leading to the release of some of the CH₄ stored in the deep ocean.

It is clear that the gas hydrate below what was once permafrost is already melting, with bubbles observed in many Canadian and Siberian lakes. With the Arctic amplification temperature, rises will be nearly twice the global average in the northern polar regions, which will accelerate the gas hydrate melting. But we still do not have an indication of how much CH₄ is stored beneath the world’s permafrost regions. So at the moment, our best estimate suggests a global warming of 3°C could release between 35 and 940 GtC, which could add between 0.02°C to 0.5°C to global temperatures.

Amazon dieback

In 1542, Francisco de Orellana led the first European voyage down the Amazon River. During this intrepid journey the expedition met with a lot of resistance from the local Indians; in one particular tribe the women warriors were so fierce that they drove their male warriors in front of them with spears. Thus the river was named after the famous women warriors of the Greek myths, the Amazons. This makes Francisco dexOrellana one of the unluckiest explorers of that age, as ordinarily the river would have been named after him. The Amazon River discharges approximately 20% of all fresh water carried to the oceans. The Amazon drainage basin is the world’s largest, covering an area of 7,050,000 km²—about the size of Europe. The river is a product of the Amazon monsoon, which every summer brings huge rains. This also produces the spectacular expanse of rainforest, which supports the greatest diversity and largest number of species of any area in the world.

The Amazon rainforest is important when it comes to climate change as it is a huge natural store of carbon. Originally it was thought that established rainforests such as the Amazon had reached maturity. Detailed surveys of all the rainforests of the world over the past four decades show this is incorrect. In the 1990s, intact tropical forests—those unaffected by logging or fires—removed roughly 46 billion tonnes of CO₂ from the atmosphere. The sting in the tail is that this removal had diminished to an estimated 25 billion tonnes in the 2010s. The lost sink capacity is 21 billion tonnes of CO₂, equivalent to a decade of fossil-fuel emissions from the UK, Germany, France, and Canada combined. All this data is compiled by the African Tropical Rainforest Observatory Network and the Amazon Rainforest Inventory Network. Over 300,000 trees are being tracked, with more than 1 million diameter measurements in seventeen countries, and the data is standardized and managed by the University of Leeds (at ForestPlots.net).

The concern about a possible Amazon rainforest dieback came from a seminal paper published in 2000 by colleagues at the UK Meteorological Office’s Hadley Centre. Their climate model was the first to include vegetation–climate feedback and suggests that global warming by 2050 could increase the winter dry season in Amazonia. For the Amazon rainforest to survive, it requires not only a large amount of rain during the wet season but a relatively short dry season so as not to dry out. According to the Hadley Centre model, climate change could cause the global climate to shift towards a more El Niño-like state with a much longer South American dry season. Kim Stanley Robinson, in his novel Forty Signs of Rain, uses the term ‘Hyperniño’ to refer to a new climate state. The Amazon rainforest could not survive this longer dry season and would be replaced by savannah (dry grassland), which is found both to the east and south of the Amazon basin today. This replacement would occur because the extended dry periods would lead to forest fires destroying large parts of the rainforest. This is exactly what was seen during the two extreme Amazon droughts of 2005 and 2010. The wildfires also return the carbon stored in the rainforest back into the atmosphere, accelerating climate change. The savannah would then take over those burnt areas, as it is adapted to coping with the long dry season, but savannah has a much lower carbon storage potential per square kilometre than rainforest does.

Modelling the Amazon forest response to climate change is complicated because there are positive and negative feedbacks. For example, higher levels of atmospheric CO₂ have a ‘fertilization’ effect on plants and trees, boosting photosynthesis and promoting growth. They also make plants more water efficient and hence more drought tolerant, offsetting some of the effects of the longer predicted dry season. Other climate models have not found such a profound dieback and the current IPCC review suggests a sustained dieback of the Amazon rainforest is unlikely this century—if the Amazon rainforest stays intact. It is that last part which is the biggest issue, because, under the leadership of Brazilian President Jair Bolsonaro, deforestation rates have been on the rise accompanied by a significant increase in forest fires, many occurring in areas which do not usually suffer from them, indicating that many are being started deliberately. The deforestation and fragmentation of the Amazon and other rainforests around the world make them more vulnerable to climate change and hence make the likelihood of a catastrophic dieback more likely.

Human-induced climate change has already affected our planet and could have even more radical impact over the next 80 years. In addition, scientists worry constantly about potential surprises in the global climate system that could exacerbate future climate change. As discussed above, these include the possibility that Greenland and/or the Antarctic could start to melt irreversibly, raising sea level by many metres in the next century. The North-Atlantic-driven deep-ocean circulation could change, producing extreme seasonal weather in Europe. The Amazon rainforest could start to die back due to the combined effects of deforestation and climate change, causing the loss of huge amounts of biodiversity and increasing carbon emissions to the atmosphere, driving further global warming. Finally, there is the threat of additional CH₄ being released from gas hydrates beneath the oceans and permafrost, which could accelerate climate change. One way to ensure we avoid the worst effects of climate change and greatly reduce the likelihood of climate surprise is to keep climate change as small as possible. The aspiration of our global leaders is to try to keep climate change to just 1.5˚C above pre-industrial levels. In Chapter 7, we explore how they came to this decision and how they hope to achieve it.