This is the third post in a new climatetippingoints.info series fact-checking claims that various climate tipping points have been crossed, and that sudden catastrophic warming is now inevitable. See the Introduction post for an overview.
Fact-Check: do tipping points and feedbacks commit us to rapid catastrophic warming?
Claim: More than 3°C of warming is already locked in over the next ~10 years even if we reduced or stopped emissions now, making catastrophic warming inevitable.
Reality: By 2030 we’ll likely reach an average of ~1.3-1.5oC. If all carbon emissions ceased now we’d reach ~1.4°C by 2100 (and more with reduced aerosol emissions), but our current trajectory is for far higher. Many of the values in committed warming lists are excessive, double-counted, or happen far slower than claimed.
There have been recent claims circulating (e.g. 1,2,3,4) that global warming is already higher than officially stated, and that even if we stopped human emissions now we’re already committed to a catastrophic or even runaway warming of 3 degrees celsius or more by around 2030. These claims often take the form of a list of feedbacks which have either been nearly or definitely triggered, and add up to a far larger number than commonly accepted.
In this post of climatetippingpoint.info‘s new Fact-Check series, we investigate how much warming we’re committed to and what role climate tipping points and feedbacks play. For each claimed source of committed warming, we add up running totals on the claimed warming and the likely actual warming between 2019 and 2030 at the top of each section
Claim: Global warming is higher than officially stated, and is already at 1.2-1.5oC
[Claimed 2030 Running Total: 1.0oC +0.2 = 1.2oC]
Reality: Although total warming from pre-industrial is a bit higher, climate impacts, targets, and tipping points are measured relative to 1850-1900 because of better observational records from then onwards
[Actual 2030 Running Total: 1.0oC +0 = 1.0oC]
A common starting point is disputing the current level of warming on the basis that climate scientists take too late a baseline to measure warming from. The global average surface air temperature change (the key figure used to define “global warming”) from the average values between 1850 and 1900 (the baseline used by the last IPCC report, AR5) stood at about 1 to 1.1oC in 2019, or about 0.8 to 0.9oC above the also-common 1951-1980 baseline:
But some say we should measure this “pre-industrial” baseline from 1750 instead of 1850, and this would make current global warming be slightly higher at 1.2oC to 1.5oC (or even 1.6 to 1.9oC if you cherry-pick just the hot February in the record year of 2016, and not a recent multi-year average).
It’s true that 1850-1900 is after industrial emissions started and so doesn’t technically count as pre-industrial. And a recent scientific paper has made the case for using 1720-1800 as the best pre-industrial baseline definition, while another shows the difference is enough to significantly affect the Paris Target carbon budgets. But the reason this later period is often chosen is because many observational records were started between 1850 and 1880, making temperature estimates more accurate. This is why many of the climate models start from 1850 too, because they need high-quality input data that an earlier start-point can’t provide – it’s not a “little trick” to make warming seem smaller.
To further confuse matters, the Paris Targets of 1.5 to 2oC didn’t actually specify what baseline its targets are measured from. However, as the conference used the IPCC AR5 report it’s assumed to be 1850-1900 as well. This means the climate targets are measured to this baseline, as well as the climate impacts analysed in the IPCC reports and many tipping point estimates.
But more importantly, it doesn’t really matter exactly which baseline we use, as long as it’s consistent. The temperatures at which climate impacts and tipping points are projected to occur are defined relative to a chosen baseline, and using an earlier and lower baseline to measure from doesn’t move the impacts and tipping points as well.
As a result, even if warming is more like 1.2 to 1.5oC versus true pre-industrial rather than 1.0 to 1.1oC, the Paris Targets, IPCC climate impacts & carbon budgets, and most tipping point estimates are defined relative to the latter. For consistency, we’ll keep using the latter values (and for simplicity use 1oC as the 2019 baseline warming for our committed warming sums).
Claim: A large amount of warming from previous emissions hasn’t hit yet
[Claimed 2030 Running Total: 1.2oC +0.7 = 1.9oC]
Reality: There is a warming lag, but it’s smaller than stated as it’s partially balanced out by negative feedbacks
[Actual 2030 Running Total: 1.0oC +0.1, but already counted in the business-as-usual section below = 1.0oC. By 2100 it’d be +0.1-0.2oC, plus each decade of continued emissions with their associated lag]
The next step in the committed warming list is the warming from recently emitted CO2 that has yet had the chance to have its full impact yet. This is called the climate lag, and has been quoted at +0.5 to +0.7oC on top of the existing warming.
The idea here is that warming from increased CO2 doesn’t happen immediately – it takes time for energy imbalance CO2 causes to drive warming, time for the vast mass of the oceans to warm up, and time for this all to reach an equilibrium. This time lag between cause and most (~60%) of the effect is estimated at about 40 years (although the ocean will continue to slowly warm up for hundreds of years longer). Similarly, it was previously estimated that if human emissions immediately stopped, warming would continue to increase by +0.5 to +0.6oC* by 2100 [*IPCC, AR5, WG1, s12.5]. Based on this, it was thought that the 40 year mark might see ~0.4oC of warming.
We can also look directly at the actual energy imbalance driving warming known as Radiative Forcing (RF; measured in Watts per square metre of Earth’s surface), which in the last IPCC report in 2013 was estimated at +2.3 W/m². Assuming this forcing remains constant (i.e. keeping greenhouse gas and aerosol concentrations fixed) and using a transient climate sensitivity of 2oC per CO2 doubling, this translates to a near-term warming of ~1.2oC within the next few decades, versus the 2013 warming level of ~1oC. But on a longer timescale the equilibrium climate sensitivity is higher at around 3oC per CO2 doubling, leading to an eventual warming of more like ~1.6-1.8oC. On the timescales of thousands of years this would be higher still (the Earth system sensitivity), but for now we’re mostly worried about impacts this century. By 2019 this RF will have increased a bit too, with current warming now around 1.0 to 1.1oC.
However, if emissions were to cease immediately the forcing wouldn’t remain constant like this, with concentrations of short-lived gases like aerosols and methane rapidly dropping and some CO2 being drawn down into carbon sinks (somewhat countering the long-term warming from the ocean’s thermal inertia). The 2018 IPCC special report on 1.5 degrees concluded that these factors sufficiently counter the warming lag to still allow warming to be kept to around 1.5 degrees by 2100 with a gradual CO2 phase out. The report also found that if just CO2 emissions ceased now, there’s be a very gradual warming lag of ~0.1 degrees by 2100 (and would carry on thereafter as result of longer-term climate sensitivity), a lower figure than the ~0.5 degrees previously discussed based on maintaining constant atmospheric concentrations.
So, if we stopped emitting CO2 now we would still expect +0.1-0.2oC of warming over the next decades, with a bit more in store over the next few centuries as well as the climate system approaches it’s longer-term equilibrium. And this assumes an immediate stop to emissions, rather than a realistic peaking and then wind-down from current emission rates to zero. This is part of why plans to keep below the 1.5oC Paris target involve such a rapid and challenging decarbonisation process with a 45% emissions reduction by 2030 (and which if we miss then relies on extensive and controversial negative emission technologies), as this warming lag would require constant countering far into the future.
Claim: Arctic sea ice disappearance will trigger a rapid jump in global warming
[Claimed 2030 Running Total: 1.9oC +0.5 = 2.4oC]
Reality: An ice-free Arctic summer will only lead to a small & gradual warming. So far it has likely added +~0.05oC (by 2007), and we might see a similar trend continue
[Actual 2030 Running Total: 1.0oC +0.05 (max. +0.2oC, but unlikely in next 10 years) = 1.05oC.]
Next up is the decline of Arctic summer sea ice (ASSI), where it’s claimed the loss of ASSI would lead to a rapid global warming of 0.5–1.0oC or more. As discussed in our in-depth Fact-Check on Arctic sea ice, in reality the loss of just summer sea ice would lead to more like a +0.1 to 0.2oC global warming increase, wouldn’t happen in one big jump, and is already captured by some of the last IPCC report’s model projections. Check out our previous post for more details.
Also linked here but not explicitly included in our target committed warming sums is the “Arctic Methane Bomb”, which is claimed to be capable of driving an abrupt 1-5oC jump in warming within only a few years. But as detailed in our in-depth Arctic Methane Fact-Check, the chance of a sudden catastrophic release of pure methane is negligible. A more gradual amplification of human-driven warming (0.15-0.3oC from permafrost by 2100 on low emission trajectories) is the most likely outcome, although even this is unwelcome.
Claim: Global dimming is masking ~0.7oC of global warming, which will suddenly appear when decarbonisation starts as aerosol emissions decline
[Claimed 2030 Running Total: 2.4oC +0.7 = 3.1oC]
Reality: Global dimming masks around 0.6oC of global warming, but a total disappearance of aerosols in a year is unrealistic
[Actual 2030 Running Total: 1.0oC +0.1 this decade, but only if there’s rapid decarbonisation = 1.05-1.15oC]
Another common claim is that if humans reduced carbon emissions now we’d actually end up causing a rapid jump in warming (by 0.7oC or even 2.5oC in less than a year) as a result of something called global dimming. This refers to the cooling effect of tiny particles called aerosols, which humans emit as well as carbon from sources like power stations, factories, and fires. This has been posed as an unavoidable paradox, as by closing high emitters like coal-fired power stations in order to reduce greenhouse gas emissions we’d end up increasing temperatures and hitting tipping points anyway.
As discussed in more depth in our Fact-Check on Global Dimming, while the total removal of human-made aerosols would lead to a short-term warming of ~0.4oC (and ~0.6oC in the longer term), an abrupt end to all emissions is very unlikely. Decarbonisation will take decades, and aerosols come from a wide variety of sources beyond just the heaviest emitting power stations. Even a challenging target of halving aerosol emissions in the next couple of decades would only lead to ~0.2oC of warming – the equivalent of around a decade more of the current warming trend – which is preferable to carrying on emitting carbon indefinitely instead and is unlikely to trigger any tipping points.
Claim: Business-as-usual human emissions will add +0.5oC in the next decade [Claimed 2030 Running Total: 3.1oC +0.5 = 3.6oC (or 2.9oC if no aerosol reductions)]
Reality: Temperature increases have consistently averaged at ~0.2oC per decade, and there’s no evidence to suggest a significant jump up in the warming rate now
[Actual 2030 Running Total: 1.05-1.15oC +0.2 = 1.25-1.35oC]
The next committed warming claim is that under business-as-usual emissions we might expect an additional 0.5oC of warming in the next decade. A rise this big sits far outside the current average warming rate of ~0.15-0.2oC per decade, and not much evidence has been given to support this claim apart from mentioning warming lags (explained earlier, and a process which is constantly ongoing and not new) and potential feedbacks (which are uncertain and unlikely to be as big as claimed, as discussed elsewhere in this series). Some decades have had slightly higher warming rates counterbalanced by other decades with slightly weaker warming (such as the so-called “global warming pause“), but the longer-term trend has so far remained pretty consistent:
Carbon emissions continue to increase and so we’d expect a similar increase in warming rate too, but there’s no reason to believe the emissions from the last decade would produce more than twice the warming per decade seen so far. So under Business-As-Usual emissions we can continue to expect around 0.2oC of warming per decade (with some decades warming faster and some slower), with the potential for a further gradual upwards creep in this value if emissions grow and extra feedbacks kick in.
Claim: For every 1oC of extra warming commitments, another 1oC is added by the water vapour feedback, so this amplifies the previous warming sources
[Claimed 2030 Running Total: 2.9-3.6oC +1 = 3.9-4.6oC]
Reality: The water vapour feedback is already accounted for by climate sensitivity in models and calculations, so adding an extra 1oC is double-counting
[Actual 2030 Running Total: 1.25-1.35oC +0 = 1.25-1.35oC]
On top of the committed warming proposed so far, some lists add on an extra 1oC of warming from the water vapour feedback per 1oC of extra committed warming (adding up to an extra 1oC or even 2.1oC on the extra warming commitments). The basis of this idea is that as the atmosphere warms up due to CO2, this makes more water evaporate, and as water vapour is itself a powerful greenhouse gas this amplifies the original warming. But the claim of a doubling of the extra committed warming figures quoted so far is flawed.
To understand why, first of all we we need to think about climate sensitivity again. Without any extra feedbacks, a doubling of CO2 concentrations in the atmosphere would directly drive around 1oC of warming. But the longer-term equilibrium climate sensitivity is around 3oC per doubling of CO2. This extra 2oC mostly comes from positive feedbacks, of which around 1oC is due to water vapour (and so is the source of the 1oC extra warming per initial 1oC claim).
But this is per doubling of CO2, not per degree celsius of warming. A doubling of CO2 from the pre-industrial 280 parts per million (ppm) to 560 ppm (an increase of 280 ppm) would lead to ~3oC of global warming, but another doubling to 1120 ppm (an increase of 560 ppm) would also yield another ~3oC. This is because climate is more sensitive in a low CO2 world, like the one we currently live in.
But more importantly, this means the water vapour feedback is already included in all our projections of future climate change, as it’s already a key component of climate sensitivity calculations. The observed and projected ~0.2oC warming per decade implicitly includes the water vapour feedback already, as it is a constantly ongoing process that happens simultaneously with CO2-warming. In the same way, when we calculated the temperature impact of other positive feedbacks like the Arctic sea ice this also used the full climate sensitivity including the water vapour feedback.
As a result, adding an extra degree of warming on top of the extra commitments claimed is double-counting, as the impact of the water vapour feedback is already included in our global warming projections.
Claim: At +3oC warming, the Amazon will burn down and add another +1.5oC [Claimed 2030 Running Total: 3.9-4.6oC +1.5 = 5.4-6.1oC]
Reality: Amazon dieback won’t suddenly happen at +3oC, and it’d add at most +0.2oC [Actual 2030 Running Total: 1.25-1.35oC + 0 (only significant beyond ~2oC) = 1.25-1.35oC]
Once the committed warming sums reach 3oC, it’s then claimed that this is the point when serious carbon cycle feedbacks like the massive release of Arctic Methane (discussed in the sea-ice section above) or the “Amazon burning down“. This latter point refers to the hypothesised dieback of the Amazon rainforest in a hotter and drier climate, and it’s claimed that this will add as much as another 1.5oC of warming. But this tipping point is not as sudden or as massive as is portrayed.
The source of this claim is based on scientists noticing that in many areas the landscape can support either a forest with near total tree cover or a savannah with mixed mostly grassland and some trees. What’s interesting is that the forest state is self-sustaining – it makes a lot of its own rain by recycling moisture, as each tree pumps up water which then evaporates and later becomes new rain upwind. But that means if enough forest is lost – through deforestation or climate change – at some point not enough moisture may get recycled any more, and the whole forest can dieback to a drier savannah instead:
There is evidence that this may at least partially be the case, but there isn’t evidence to suggest that this will happen suddenly when warming hits 3oC or release a huge amount of carbon. In various modelling studies the risk of Amazon dieback gradually grows above2oC and becomes significant in the 3 to 5oC range, but will still take 100-200 years to fully play out:
Then there’s the question of how much carbon this process will release. The video linked to claims 1.5oC of warming is possible, which would imply a CO2 increase of ~200 ppm. The total Amazon carbon sink is estimated to be 150-200 gigatonnes of carbon (GtC), with a maximum forest dieback to savannah estimated to release 53-70 GtC and raise atmospheric CO2 by 25-33 ppm (which is enough to cause up to ~0.15oC of warming if none of the CO2 was re-sequestered). It’s clear then that although this would be significant and unwelcome, it’s nowhere near the CO2 release that 1.5oC of warming would require. Even if all tropical and sub-tropical forests experienced dieback to savannah this would release ~191 GtC / ~90 ppm CO2, but this is even less likely to be a single global tipping point occurring specifically at +3oC. Dieback isn’t projected to be total even in the worst-case models either, with projected Amazon rainforest biomass shrinkage expected to be 24–40% (releasing 15-25 GtC) by 2100.
Overall, although the risk of Amazon dieback grows after reaching 2oC and becomes significant after 3 or 4oC, the warming caused by a total dieback is likely to be more modest (a maximum of ~0.15oC, and more like 0.1oC, instead of 1.5oC) and occur over a timescale of decades to centuries rather than years. A more likely gradual and partial dieback would cause up to ~0.05oC of global warming by 2100. As with many carbon cycle feedbacks, the issue is not that they’ll suddenly release a huge amount of carbon and cause abrupt warming. The issue is that they’ll shift from acting as net carbon sinks to net carbon sources, gradually adding carbon to the atmosphere instead of removing it as they do now.
Claim: Continents will heat up twice as much as the global average, & ruin agriculture
[Claimed 2030 Running Total: 5.4-6.1oC x2 = 10.8-12.2oC experienced in continental interiors]
Reality: Continental interiors will warm up ~+0.5-1oC more than the ocean, not double [Actual 2030 Running Total: 1.25-1.35oC, but slightly more in continental interior (and twice as much in the Arctic), as is the case now]
Finally, a claim often tacked on to the end of extra committed warming lists is that continents will warm up twice as much as the ocean. This means the large committed warming claimed will actually be even more extreme where many people live, and importantly where most crops are grown.
While it’s true that the land has warmed up more than the sea due to global warming, observations so far and climate model projections of the future do not show a doubling of temperatures. Under business-as-usual, projected warming by 2100 is expected to be up to 1oC more in continental interiors compared with a ~5oC warming overall – enough to further stress people and crops, but not as extreme as double:
Maps of projected future warming under low (left) and high (right) emission scenarios, showing slightly higher warming in continental interiors relative to the ocean (but not a doubling, except the Arctic). Dots mark areas with strong model agreement. Source: IPCC AR5
There are many positive feedbacks and tipping points which are claimed to be poised to rapidly make global warming far worse within only a few years. In this post we’ve gone through some of the cited sources of these “committed” warming claims, and found that the actual likely total is lower than sometimes stated:
- While the full warming from pre-industrial is probably a bit higher than the “official” 1oC, climate impacts and tipping points are defined relative to the latter
- There is a “climate lag” with ~0.1-0.2oC of warming (not ~0.5oC as often quoted, as this assumes fixed concentrations) yet to arrive by 2100, but this will happen gradually with only a fraction in the next ~10 years
- This figure of +0.2oC is also close to the current decadal warming rate, with no evidence for a jump up to +0.5oC in the next decade
- Global dimming is masking around 0.5-0.6oC of global warming, but even a rapid decarbonisation programme would only expose maybe +0.1oC of this anytime soon
- Arctic sea ice disappearing in the summer is unlikely in the next decade, and will cumulatively only add ~+0.1-0.2oC when it does. A “Methane Bomb” is also unlikely
- The water vapour feedback is already included in these projections, but is often double-counted
- And the Amazon forest won’t burn down at exactly +3oC of global warming, with its potential dieback to savannah gradually releasing around +0.1oC worth of carbon
When the influences of these factors and feedbacks are summed up, we might expect warming o reach around 1.3-1.5oC by ~2030 (maybe a bit more or maybe a bit less, with some years being high or low outliers), rather than the +3< or even +5< oC that is sometimes claimed. Even this is worth avoiding though – continued warming of around 0.2oC per decade of growing emissions plus the extra ~0.1oC lagging behind will put us past the 1.5oC Paris Target in only 10-20 years’ time, and past the 2oC Target around mid-century. Even if all emissions suddenly stopped now, without additional drawdown the warming lag would still push the warming up to ~1.2-1.3oC by 2100, plus a bit more from any aerosol reduction (which depends on extent, but maybe a temporary ~+0.2-0.3oC – see comment section for discussion). And the higher the warming goes, the more that feedbacks like Arctic sea ice, permafrost thaw, and Amazon dieback start to kick in. But these tipping points and feedbacks won’t trigger abrupt warming – instead they’ll act as new sources of extra gradual warming that make sticking to the Paris Targets even more challenging.
This post was written by Dr. David A. McKay, currently a Postdoctoral Researcher at Stockholm Resilience Centre (Stockholm University), where he is part of the Earth Resilience in the Anthropocene Project (funded by the European Research Council) and is researching non-linear climate-biosphere feedbacks. This post was written in his spare time with no funding support for this site, and was proofread and edited by Dr. Rachael Avery.
Post Update Log: 18/5/19 with a little extra information on the likely committed warming by 2100 – see comments for discussion – the Arctic sea-ice section was also updated to refer to the new Arctic Methane Fact-Check; 30/5/19 with additional details on the projected maximum Amazon dieback by 2100, as actual dieback is very unlikely to be total by 2100; 22/8/19 with clarification that the climate lag calculation is based on constant concentrations (not zero emissions), and updated to better reflect more recent lag calculations by IPCC SR1p5 (i.e. CO2 lag values revised down from ~0.5-0.6 to ~0.1-0.2 degrees by 2100); 3/12/19 with clarification that CO2->temperature calculations assume no re-sequestration, and so are maximum values (re-sequestration currently absorbs c. half of all emissions); 28/5/20 to specify 2019 and 1oC as the base year/warming for calculations; 26/10/20 to add extra paper to real pre-industrial discussion.