Flattening the Curve of Future Emissions

A decade ago the world seemed on track for a particularly grim climate future. China was building a new coal plant every three days; global emissions were increasing at a rate of 3% per year and increased by 31% between 2001 and 2010. Scenarios where global carbon emissions tripled by the end of the 21st century with coal use increasing sixfold seemed plausible to many. Researchers argued that “business as usual” would likely lead to a world 4ºC or 5ºC above pre-industrial levels by 2100.

Today, the world is a very different place. We are succeeding in making clean energy cheap, with solar power and battery storage costs falling 10-fold since 2009. The world produced more electricity from clean energy — solar, wind, hydro, and nuclear — than from coal over the past two years. The International Energy Agency (IEA) now argues global coal use is in structural decline, unlikely to ever surpass its 2013 peak, while the odds of a 21st century dominated by coal now seem vanishingly small.

Global emissions only rose 1% per year over the past decade, and are projected to plateau in coming years under current policies and commitments. Thirty-two countries have absolutely decoupled their emissions from economic growth, and global emissions may begin to fall this decade as countries are increasingly committing to more ambitious near-term and long-term climate policies. As we point out in this analysis, the world is now on track for a best estimate of 3ºC above preindustrial levels by 2100 in a current policy world, and 2.4ºC when near-term stated policies and commitments – such as those made under the Paris Agreement – are taken into account.

The world will continue to warm as long as emissions remain above net-zero, and peaking global emissions is only one of many stepping stones on the journey to deep decarbonization. We are still far short of what would be needed to meet the Paris Agreement goal of limiting global temperatures to well-below-2ºC above pre-industrial levels. But it is clear that a very high emissions future is far less likely today.

Unfortunately the spectre of that very high emissions future is still haunting much of the climate impacts literature – with a sizeable portion of studies on future climate impacts focused on a warming scenario called “ RCP8.5” or its successor “ SSP5-8.5.” These high-emissions scenarios are frequently referred to as “ business as usual,” suggesting they are a likely outcome if society does not make concerted efforts to cut greenhouse gas emissions. For example, The New York Times referred to this outcome specifically as “business as usual” earlier this month.

However, according to the researchers who developed them, these scenarios were never intended to be a most-likely “business as usual” outcome, but rather a worst-case baseline outcome in a world of no climate policy “consistent with the highest emissions scenarios in the literature.” A decade later, we no longer live in a baseline world and have increasing clarity on what emissions scenarios are more likely than others, given technological progress, climate policies in place today, and both near- and long-term commitments. This change is reflected in the new IPCC 6th Assessment Report, which says “the likelihood of high emission scenarios such as RCP8.5 or SSP5-8.5 is considered low in light of recent developments in the energy sector.”

No longer in a “business as usual” world

Back in 2019, we argued that “ a 3ºC world is now business as usual,” using 2040 projections from the International Energy Agency (IEA). We suggested the most likely outcome of current policies would be 2.9ºC warming relative to preindustrial levels by 2100, though uncertainties in how sensitive the climate is to our emissions mean this could end up anywhere between 2ºC and 3.8C. A number of other groups have published their own estimates of likely climate outcomes in a current policy outcomes scenario — as shown in Figure 1, below — including many new estimates that have come out in the past two years.

Figure 1: Estimates of 2100 warming in current policy scenarios relative to preindustrial (1850-1900) temperatures from Hausfather and Riche 2019, Rogelj et al 2016, Climate Action Tracker (May 2021 update), Liu and Raftery 2021, and the integrated assessment models (MESSAGE, REMIND, and GCAM) participating in the 2021 NGFS Climate Scenarios.

These studies use a wide variety of independent approaches to arrive at largely similar results: the best estimate of global warming by 2100 under current policies is 3ºC (2.8ºC to 3.2C) with an uncertainty range of 2ºC to 4.4ºC reflecting roughly the 10% to 90% uncertainty range in climate sensitivity outcomes.

The results of the new IAM runs done as part of the NGFS Climate Scenarios are particularly noteworthy. Two of the three IAMs used (MESSAGE and REMIND) were also used to create the high-end emissions scenarios ( RCP8.5 and SSP5-8.5, respectively). The fact that the same models used to generate 8.5 outcomes suggest that they are implausible in a current policy world should serve as a strong signal that these outcomes are not “business as usual”. These current policy outcomes are most analogous to the SSP4-6.0 scenario in terms of expected 2100 climate outcomes.

In addition to studies of current policy outcomes, a number of groups have produced estimates of likely warming under “stated policies” or “pledges and commitments.” This is a bit more of a moving target, as what specific pledges are included or excluded will differ by modeling group; some include existing 2030 Paris Agreement nationally determined contributions (NDCs) only, while others include some additional 2030 commitments that have yet to be codified in Paris NDCs (but none include longer-term 2050 or 2060 net-zero commitments). The range of warming outcomes projected under near-term stated policies are shown in Figure 2, below.

Figure 2: Estimates of 2100 warming in stated policy scenarios relative to preindustrial (1850-1900) temperatures from Hausfather and Riche 2019, Rogelj et al 2016, Climate Action Tracker (May 2021 update), Liu and Raftery 2021, Pielke et al 2021, and the integrated assessment models (MESSAGE, REMIND, and GCAM) participating in the 2021 NGFS Climate Scenarios. Note that two scenarios are shown from Liu and Raftery 2021: the “adj” scenario where emissions revert to their pre-2015 trends after 2030, and the “con” scenario where countries continue to improve their carbon intensity at the same rate after 2030.

Differing choices in what is included in stated policies — and what is assumed to happen after the 2030 Paris commitment period — result in a wider range of future warming estimates than in a current policy world. But there is broad agreement that stated policies would result in notably lower warming than current policies, with best estimates ranging from 2.2ºC to 2.7ºC and a median of 2.4C. Including uncertainties in future warming outcomes associated with climate sensitivity, we end up with a range of between 1.5ºC and 3.7C. These stated policy outcomes are most analogous to the SSP2-4.5 scenario in terms of expected 2100 climate outcomes.

There are many different potential uses for future emissions scenarios. Current policy outcomes can be a useful counterfactual — what would happen if policies in place today remained unchanged for the rest of the century — to calculate the benefits of additional emissions mitigation. At the same time, we are unlikely to remain in a current policy world; while countries may not meet all their stated climate commitments, they will likely meet at least some of them.

The past decade has been characterized by strengthening climate policy and rapid declines in clean energy costs, and these trends appear likely to continue. A stated policy world is probably a more likely outcome from the vantage of the present than a current policy world. At the same time, the future is deeply uncertain. We cannot rule out a world where some current policies are rolled back, where countries directly subsidize coal consumption as the recent US administration unsuccessfully attempted to do. Current and stated policies are neither a ceiling nor a floor on possible future emissions outcomes.

In addition to near-term stated policies, numerous countries including the US, China, EU, UK, Japan, South Korea, and others representing over two-thirds of global CO2 emissions have made commitments to reach net-zero emissions by 2050 or 2060. These commitments are in some cases stated goals rather than codified in legislation, and the translation of long-term commitments into near-term goals will go a long way toward demonstrating how seriously we should take these targets. There are fewer assessments of the impacts of longer-term net-zero commitments on global temperatures, but one by Climate Action Tracker estimates that – if followed through – these commitments would in aggregate limit warming to 2ºC above pre-industrial levels by 2100 (with an uncertainty range of 1.6ºC to 2.6C).

Emissions are only one uncertainty

Were our emissions the only determinant of global warming, we would likely end up somewhere around 3ºC above preindustrial levels by 2100 in a current policy world, and 2.4ºC in a stated policy world. However, emissions are just one of three major uncertainties the world faces when projecting future warming; the other two are climate sensitivity and carbon cycle feedbacks.

Climate sensitivity refers to the amount of warming the world will experience as CO2 in the atmosphere increases; it is typically expressed using a simple metric of how much the world will warm over the long term if atmospheric concentrations of CO2 are doubled.

Climate sensitivity has long been a “holy grail” of sorts for the climate science community but has been difficult to narrow down. Back in 1979, Dr. Jules Charney led a National Academy of Sciences report suggesting if atmospheric concentrations of CO2 were to double (e.g. from their preindustrial value of 280 parts per million to 560 parts per million), the world would likely warm by somewhere between 1.5ºC and 4.5ºC.

The prior IPCºC assessment report (AR5), published 34 years after Charney’s report, gave the same “likely” range of 1.5ºC to 4.5ºC warming per doubling of CO2. Thankfully some meaningful progress has been made on the question of climate sensitivity in the past few years. A recent assessment of climate sensitivity undertaken under the auspices of the World Climate Research Programme – where I was a coauthor – provided the first comprehensive case for narrowing the range of climate sensitivity based on multiple lines of evidence. The assessment suggested climate sensitivity is likely to be between 2.5ºC and 4ºC per doubling CO2, and very unlikely sensitivity is below 2ºC or above 5ºC.

Carbon cycle feedbacks represent the other major source of uncertainty when projecting future warming. Today, around half of the CO2 emitted by humans remains in the atmosphere, with the remainder absorbed by the oceans and land. However, as the Earth warms this is expected to change. For example, warming reduces the amount of CO2 absorbed by surface ocean waters and the amount of carbon sequestered in soils. It can also accelerate tree death and the risk of wildfires. Thawing permafrost may release additional carbon into the atmosphere.

Overall, the carbon cycle is expected to weaken as a result of climate change, leading to more emissions remaining in the atmosphere and less being absorbed by the land and oceans. All of these processes introduce uncertainty when translating future CO2 emissions into changes in atmospheric CO2 concentrations.

Future warming scenarios developed by the climate modeling community do consider carbon-cycle feedbacks, but often use single estimates of the feedback strength from previous studies and do not include any of the uncertainties in carbon-cycle feedbacks. The reason scenarios leave out carbon-cycle feedback uncertainties is that about half of the climate modeling groups do not currently include the biogeochemical cycles needed to model carbon-cycle feedback changes. In a 2020 analysis, we found that including uncertainties in future carbon cycle feedbacks results in as much as 13 percent less warming or 25 percent more warming than estimates based on climate sensitivity alone.

All three uncertainties that determine possible future warming outcomes – emissions, climate sensitivity, and carbon cycle feedbacks – are combined in the figure below. The black dots show the best-estimate warming outcome from emissions alone in no-policy baseline scenarios, current policy scenarios, and stated policy scenarios. The colored bars show the range of possible warming due to climate sensitivity uncertainties, while the black whiskers show the additional range of possible warming due to carbon cycle feedback uncertainties.

Figure 3: Estimates of 2100 warming in stated policy scenarios relative to preindustrial (1850-1900) temperatures in different SSP baseline scenarios, current policy scenarios, and stated policy scenarios. Colored bars show uncertainty associated with climate sensitivity, while whiskers show additional carbon cycle feedback uncertainties based on Hausfather and Betts 2020.

There are a number of different “baseline” scenarios produced as part of the Shared Socioeconomic Pathways (SSPs). These suggest that a world with no climate policies after 2010 could end up with a best-estimate of warming anywhere between 4.6ºC warming by 2100 (in the high-end SSP5) or as low as 3ºC (SSP1). Current policy outcomes result in a best-estimate of 3ºC warming, consistent with the low-end of the baseline scenario range, while stated policy leads to 2.4ºC warming.

If we take climate sensitivity and carbon cycle feedbacks into account, we cannot rule out as much as 7.7ºC warming (or as little as 2ºC warming) in a baseline scenario world, 5ºC (or 2ºC) warming in a current policy world, or 4ºC (or 1.7ºC) warming in a stated policy world. This means while we can firmly rule out very high warming outcomes (>5ºC) in a current or stated policy world, it is much harder to fully rule out a risk of 4ºC+ warming. These are not the most likely outcomes, of course, but remain a risk should we get unlucky with a combination of high climate sensitivity and high carbon cycle feedbacks.

A best-estimate current policy outcome of 3ºC by 2100 is hardly a world we want to live in, with substantial impacts on both human and natural systems. Here uncertainty is decidedly not our friend; even if we were willing to accept risks of 2.5ºC or 3ºC warming, the “tail risks” of substantially higher outcomes provide an important motivation to further mitigate emissions. There is a danger of being too deterministic around likely warming outcomes of our emissions if we do not properly account for other climate system uncertainties.

Carbon dioxide accumulates in the atmosphere over time, and until emissions reach net-zero the world will continue to warm. This is the brutal math of climate change, and it means that the full decarbonization of our economy is not a matter of if but when. This means that even if our estimate of 2100 warming is 3ºC (2ºC to 5ºC) under current policies and 2.4ºC (1.7ºC to 4ºC) under stated policies, the world will continue to warm after that point as these emission pathways do not get us to net-zero.

The world has made real progress toward bending down the curve of future emissions. The worst-case outcomes of a decade ago are much less plausible today. At the same time, we have a long way to go to reduce global emissions if we want to meet Paris Agreement goals of limiting warming to well-below-2ºC. We can both acknowledge the progress we have made and how far we still have to go. The justification for limiting warming to well-below-2ºC never required having a 5ºC counterfactual, and it is important that we acknowledge where we are headed today rather than what might have been.

_Image Credit: Photo: Cpl Phil Dye RAF/MOD, OGL v1.0, via Wikimedia Commons_