RCP8.5 (and its CMIP6 successor SSP5-8.5) has been retired from the core set of scenarios for CMIP7 and the IPCC’s upcoming AR7.

This follows the publication of the van Vuuren et al. (2026) paper in Geoscientific Model Development, which outlines the new ScenarioMIP framework.

RCP8.5 was a high-end radiative forcing pathway (~8.5 W/m² by 2100) designed as an upper-bound “no or minimal policy” scenario. It assumed massive coal expansion, high population growth, and slow clean energy progress. It was never meant as the most probable “business-as-usual” outcome but was frequently presented or interpreted that way in research, media, and policy discussions, leading to projections of ~4–4.5°C (or higher) warming by 2100 under high climate sensitivity.

It was never intended as the “most likely” or business-as-usual outcome — rather, an upper-bound exploratory case. However, its frequent use in research and communication led to debates about overstatement of risks.

The retirement aligns with arguments made years ago (e.g., Hausfather & Peters 2020) that extreme scenarios were being overused. Real-world emissions have tracked closer to lower-to-medium pathways than RCP8.5. However, forcing (effective radiative influence) has sometimes tracked higher due to other factors, and uncertainties in climate sensitivity remain.

This is a positive step toward more realistic scenario planning.

The CERES-Science team welcomes the retirement of RCP8.5/SSP5-8.5 from the core scenarios for CMIP7 and IPCC AR7 (as outlined in van Vuuren et al. 2026) but argues it’s only a partial improvement.

_____________________________________________________________________________________

RCP8.5 Has Been Dropped: What It Means — and What It Doesn’t

“The United Nations TOP Climate Committee just admitted that its own projections (RCP8.5) were WRONG! WRONG! WRONG!” – US President, Donald J. Trump, May 16, 2026 – https://truthsocial.com/@realDonaldTrump/posts/116586488927495029

The developers of the greenhouse gas emissions scenarios for the IPCC’s next assessment report (AR7) have officially dropped RCP8.5 – the high-emissions pathway long treated as a de facto “business-as-usual” case (Van Vuuren et al. 2026). The CERES website has the story.

We at CERES were among the first to criticize the use of RCP8.5 in the peer-reviewed literature (Connolly et al. 2020) and welcome this decision. But, while President Trump’s description of RCP8.5 as “WRONG! WRONG! WRONG!” has a kernel of truth beneath the hyperbole, the real story is more nuanced – less dramatic, but scientifically important.

Dropping RCP8.5 is a good step in the right direction. However, the IPCC still look to be set to continue their overreliance on computer model projections that are highly sensitive to changes in greenhouse gas concentrations. So, while their “business-as-usual” projections will probably be a bit less “hot” than before, unfortunately they still look set to be producing more “alarm-as-usual”.

In this essay, we discuss what RCP8.5 actually was, why it was problematic from the start, and what this change does – and does not – mean.

 

What exactly is RCP8.5 and why was it so significant?

According to current climate computer models, the main driver of past and future climate change is the concentration of “greenhouse gases” in the atmosphere.

Most of these greenhouse gases are naturally-occurring. However, many key human activities (energy usage, agriculture, etc.) also are associated with greenhouse gas emissions. For this reason, the UN and others – especially the computer modellers – have been insisting for decades that humans should be causing “anthropogenic global warming” (AGW) from our greenhouse gas emissions. [“anthropogenic” = is a more ‘sciencey’-sounding word for human-caused].

In fact, as Dr. Roger Pielke Jr. pointed out in a key 2005 paper (R. A. Pielke 2005), through their “Framework Convention on Climate Change” (UN FCCC), the UN has – at each of their annual UN Conference of the Parties (COP) meetings – been effectively defining all “climate change” as human-caused by definition for more than 30 years. These COP meetings are the annual UN gatherings where politicians and NGOs gather from around the world in a chosen city to discuss climate change policies.

This brings us to the so-called RCP scenarios – “Representative Concentration Pathways” (Meinshausen et al. 2011). For the IPCC’s 5^th Assessment Report, the computer modellers had been given a choice of 4 different “projections” for what greenhouse gas emissions (and concentrations) might be over the coming centuries. Each scenario was given a number which represented the theoretical extra “radiative forcing” that greenhouse effect models said the extra greenhouse gases would have by the end of the 21^st century – in units of Watts per square meter: 8.5, 6.0, 4.5 or 2.6.

In section 5.3 of Connolly et al. (2020), we discuss what scientific merit these theoretical “radiative forcing” calculations have for understanding the climate. But, within the climate modelling community, the size of this final theoretical value is the primary determinant of how much “global warming” would occur. So, the various scenarios were named based on that number: RCP8.5, RCP6.0, RCP4.5 and RCP2.6.

When you looked in detail at the emissions scenarios, RCP6.0, RCP4.5 and RCP2.6 all implied that there were major policy interventions to reduce greenhouse gas emissions during the 21^st century. In contrast, RCP8.5 was the only scenario that didn’t assume major interventions.

Perhaps for this reason, the climate modellers began relying on the RCP8.5 scenario as the “business-as-usual” scenario. This meant that thousands of papers describing simulations of “the climate changes that will happen if we don’t do something urgently” ended up being published based on this RCP8.5 scenario – i.e., the one that has now been dropped.

 

Why was RCP8.5 unrealistic from the beginning and why has it been officially dropped?

 

When we wrote our paper in 2020, we noted that a few others had criticised the use of RCP8.5 as “BAU”. Ritchie and Dowlatabadi (2017b, 2017a) dug into the economics of the scenario and discovered the RCP8.5 assumed that societies would dramatically increase the use of coal over the next century – so much so that the scenario suggested that coal usage would burn more than 5 times the known reserves of coal!

They warned that this was ludicrous – but since they were writing in academic journals, the exact language they used was a lot milder:

“Accounting for this bias indicates RCP8.5 and other ‘business-as-usual scenarios’ consistent with high CO2 forcing from vast future coal combustion are exceptionally unlikely.” (Ritchie and Dowlatabadi 2017b).

Prof. Ritchie was joined in his criticism by others, including Prof. Pielke Jr. (who we met earlier), e.g., (R. A. Jr. Pielke and Ritchie 2021; Burgess et al. 2020).

Meanwhile, the scenario was also criticised from a completely different angle: Hausfather and Peters (2020) argued that the scenario was unrealistic because they believed governments were transitioning to lower emissions policies. They therefore insisted that “business-as-usual” should consider the recent trends in policies to be “business-as-usual”.

Ultimately, the recent decision to drop RCP8.5 from the next report seems to have taken Hausfather & Peter’s arguments as their primary justification:

“the CMIP6 high emission levels (quantified by SSP5-8.5) have become implausible, based on trends in the costs of renewables, the emergence of climate policy and recent emission trends” – Van Vuuren et al. (2026)

Although, they did also cite Ritchie & Dowlatabadi’s criticisms as a key factor in their decision.

However, our criticism of RCP8.5 in Connolly et al. (2020) came from a different angle…

 

What was our goal in Connolly et al. (2020)?

After the 2015 COP meeting in Paris, France, a highly influential agreement was signed called the Paris Agreement. All of the nations agreed to,

“…holding the increase in the global average temperature to well below 2 °C above preindustrial levels and pursuing efforts to limit the temperature increase to 1.5 °C above preindustrial levels” – UN Paris Agreement (UNFCCC 2016)

Parsing this convoluted statement and what it means was the main theme of our Connolly et al. (2020) paper. Essentially, they seemed to be saying that governments around the world would have to collectively implement major policy changes to keep global warming “below 2 °C above preindustrial levels”. But what did that mean?

What policies did they need to implement? How much “global warming” would be expected without these policies? What did they mean by “pre-industrial levels” – there have been many warm and cold periods over the millennia – which one did they consider “pre-industrial”?

We realised that this widely-promoted “agreement” didn’t mean anything unless you had some sort of answer for those questions.

So, for Connolly et al. (2020), we decided to try and answer a fundamental question that the Paris Agreement needed answered before the agreement could mean anything other than waffle:

“How much human-caused global warming should we expect with Business-As-Usual (BAU) climate policies?”

If we don’t know the answer to that, then how could policymakers figure out what to do to meet their nebulous “goal”.

When we tried to answer that question, we realised that the answer actually depends on four sub-questions:

1. How much greenhouse gases will we emit if we continue current trends?
2. How much of those emissions will remain in the atmosphere?
3. How much of the warming already observed is caused by humans?
4. How strongly do greenhouse gases warm the planet?

 

In our paper – very long, but methodical and systematic – we went through each of these questions, step-by-step. For each step, we surveyed the scientific literature to identify the different perspectives out there, looked at the available data, considered the various assumptions that could be made. We then compiled a best-estimate of what we could expect if trends continued “business-as-usual”, i.e., based on current trends.

We warned however that this was a multi-disciplinary problem,

“With each of these questions, there is considerable debate in the scientific literature. However, the relevant literature for each subject comes from quite different academic disciplines. The first question is typically addressed by economists, political scientists, environmental governance researchers, etc. The second question is mostly the realm of biologists, ecologists, geochemists, oceanographers, etc. The third and fourth questions are both climate science problems, but even within these topics, there are separate bodies of literature from, e.g., computer modelling research groups, groups evaluating climate records, statisticians evaluating results from a meta-analysis perspective, etc.

[…] However, we suggest that unless the scientific debate over all four of these questions are simultaneously considered, it is unlikely that truly satisfactory answers will be achieved to the over-arching question, “How much human-caused global warming should we expect with Business-As-Usual (BAU) climate policies?” – Connolly et al. (2020), p3.

 

Q1. How much greenhouse gases will we emit if we continue current trends?

We saw above that much of the debate over RCP8.5 has focused on how policies might or might not change in the future. But, in our opinion, this goes against the very definition of “business-as-usual”. Instead, we reasoned, why don’t we simply define “business-as-usual” emissions as if the emissions were to continue business-as-usual?

Rather than speculating on complicated potential “scenarios” on “what if” and “what if not”, we said, why don’t we simply look at the trends in emissions over the past 50-70 years and assume that these trends would continue “business-as-usual”?

 

What happens when we compare this BAU emissions scenario (grey envelope with black dashed line in the graphs below) to the RCP scenarios? For simplicity, let’s just consider CO2 emissions (in the paper, we also consider the other two main anthropogenic greenhouse gases, i.e., CH4 and N2O).

As soon as we plotted the graph below, it was obvious that the RCP8.5 scenario was never “business-as usual”. Notice how the RCP8.5 scenario rapidly deviates above the actual BAU envelope and stays above it?

For us, RCP8.5 was clearly not a “business-as-usual” scenario. Nor were any of the other RCP scenarios offered to the modellers.

Bizarrely, the very first IPCC report had already developed in 1990 (36 years ago) as “Scenario A” a business-as-usual projection that matched remarkably well with our empirical fit that we developed 30 years later in our 2020 paper. See below:

Why didn’t the IPCC modellers simply stick to that 36 year old “Scenario A” as their “business-as-usual”? We don’t know. Cynics might suggest that it wasn’t giving results that were “hot enough” – maybe the computer models wanted to “spice up” their simulations with some extra “warming”? Or maybe, the older Scenario A just seemed too old-fashioned?

But, whatever the reason, the extra warming the RCP8.5 simulations added to the mix were like crack cocaine to the modellers. It was routinely treated as if the scenario were “business-as-usual”.

For the 6^th Assessment Report (AR6, 2021), the scenario groups tried to add in a bit more balance and introduced intermediate scenarios into the updated “SSP/RCP” scenarios, e.g., “SSP3-70 (Baseline)” with a 2100 Radiative Forcing of 7.0 instead of 8.5. But, they left the 8.5 scenario in as a second “baseline”: “SSP5-85 (Baseline)”. And, so, the 8.5 scenarios continued to dominate the computer model simulations of “business-as-usual”.

However, while much of the recent discussion about RCP8.5 has focused on what emissions we can expect if society continues business-as-usual, in our paper, we emphasised that this was just the first step. The next question is establishing how much of those emissions should we expect to remain in the atmosphere under business-as-usual…

 

Q2. How much of those emissions will remain in the atmosphere?

Almost continuous direct measurements of the average concentration of CO2 in the atmosphere have been carried out at Mauna Loa, Hawai’i since 1958. These measurements have shown a long-term increase from about 0.031% of the atmosphere in 1958 to about 0.043% today.

Since these are such tiny percentages, typically, people refer to CO2 concentrations in terms of parts per million by volume (“ppmv” or sometimes “ppm”). But, we’ve found that people have a better feeling of what the numbers mean when we use percentages. So, in the graph of the data below, we provide the values in ppmv on the left and the equivalent values in % on the right.

There have been several attempts to try and estimate how CO2 has changed in the atmosphere in the years before the Mauna Loa record. The IPCC and many others rely on the Antarctic ice core estimates. These imply that for the last few thousand years, CO2 only varied a tiny amount – between 0.027 and 0.029% – see the grey band in the figure above.

The fact that human-caused CO2 emissions have been increasing and CO2 concentrations have also been increasing has convinced many that we just need to focus on the emissions part. But, when you convert the annual human-caused (“anthropogenic”) emissions into the expected increase in CO2 concentrations, a problem arises.

Read the full story here.