A Pragmatic Guide to the New CMIP7 Climate Scenarios and the Retirement of RCP8.5

New Climate Scenarios, the death of RCP8.5, scenarios schemarios. Yes, the headlines have done their thing, drumming up drama and confusion about these new scenarios. Maybe they have even triggered a bit of an existential crisis: “but what does this mean?” is a question we’ve been asked multiple times over the past few weeks.

So let’s start with that main question or assumption: “Does the retirement of RCP8.5 mean we’re doing ok and can all relax? or to paraphrase the words of a certain world leader “liar liar pants on fire” (aimed at the obviously lucrative industry of academia).

Well, yes and no. Emission projections are reducing largely because of climate policy and the cost of renewables, did this happen faster than anticipated, yes. Is that good news, definitely! Is it all going to be smooth sailing from here? Well no. Climate scenarios and Earth System Models (ESM) are just one tool in the kit. A very well documented shortcoming of these models is that they don’t capture extreme events and tail events very well. Climate by definition is the average of climate, so those tails get averaged away. The retirement of RCP8.5 does not mean things can only get better. It means emission projections have gone down.

If you take one thing away from this blog post it should probably be that scientists are not fortune tellers. They use the best data and understanding available at the time to develop scenarios that can help answer the question: “What if this happens?”

In the rest of this blog post, we’re going to take a step back and focus on the key takeaways of the new scenarios. And that includes discussing the demise of RCP8.5, and of course SSP5-8.5, as it is the equivalent scenario from the current generation of scenarios).

A summary of takeaways is provided below, if you don’t have the time or attention span to read the rest of this article:

• A smaller range of plausible emissions is captured in the new CMIP7 scenarios than previous generations: Changes in climate policy, legislation, the cost and extent of renewable energy have all impacted this reduced plausible range of emissions. More information over time allows us to reduce these ranges of ‘plausible’.
• The new CMIP7 “High” climate scenario represents lower emissions than RCP8.5 or SSP5-8.5. So in essence, RCP8.5, and SSP5-8.5 are retiring. It will likely be at least a year before the new CMIP7 model output is available so they have a few years of life in them.
• For continuity: The mid-high climate scenario SSP3-7.0 from the current generation of scenarios (CMIP6) will still be captured in the next generation of climate scenarios (CMIP7).
• Climate scenarios do not capture all eventualities: Tail events , rare events such as 1 in 500 or 1 in 1000 are not well captured. Similarly, tipping points, critical thresholds in the climate system are not captured in these scenarios or by the current generation of Earth System Models (ESMs).
• Our current pathway: Current emissions and projections of climate change indicate circa 2.6°C of warming above industrial levels by 2100. This places us in the middle of the current CMIP6 scenarios, i.e., SSP2-4.5.
• Climate Sensitivity, the elephant in the room: Climate Sensitivity is typically defined as ”the global temperature rise following a doubling of CO2 concentration in the atmosphere compared to pre-industrial levels.” Current best estimates of Climate Sensitivity indicate the following ranges:

  Likely Range (66–100% probability): 2.5°C to 4.0°C

  Very Likely Range (90–100% probability): 2.0°C to 5.0°C

  Whilst scenarios provide emissions to drive an Earth System Model (ESM). The amount each ESM warms in response to these emissions will differ due to different assumptions and design choices embedded in the ESMs. This is the climate sensitivity, it is an emergent property of the models, it is not prescribed.

If you’re just left wondering, I just don’t know what to do with myself, here is our practical guide to actions:

• Expect a transition period: CMIP7 scenarios are being defined now, but full CMIP7 model outputs may take time — plan for interim approaches (e.g., SSP3-7.0 for continuity + sensitivity/tail overlays for stress tests).
• Don’t equate “lower emissions range” with “lower risk”: A narrower scenario range reflects updated plausibility (policy/technology shifts), but it doesn’t remove tail risks — rare, high-impact events and tipping points still sit outside what scenarios/models capture well or even the range in estimates of climate sensitivity.
• Be explicit about what scenarios are (and aren’t): Scenarios are inputs (emissions pathways), not predictions. Use them to stress-test decisions (“what if?”), not to make single-number forecasts.
• Treat model choice as a first-class decision: Different Earth System Models warm by different amounts under the same scenario (climate sensitivity), so results can vary materially even before you change the scenario.

CMIP, the Coupled Modelling Inter-comparison Project acts as the framework that guides the experimental design and outputs from Earth System models.

CMIP7, the latest iteration of this project, is currently underway. This means that the outputs from a new generation of climate models and a new set of climate scenarios will be available over the next year or so.

ScenarioMIP (Scenario Model Intercomparison Project) is a sub-project under the umbrella of CMIP. Its main objective is to produce climate scenarios such as RCPs, SSPs, and the new CMIP7 scenarios.

The objective of ScenarioMIP is to design and carry out Earth System model simulations driven by alternative plausible futures of emissions and land use.

Scenarios are updated with each iteration of CMIP (we’ve lost the ‘high scenario’ before in previous iterations). They need to evolve to reflect advances in scientific understanding and changes in how we discuss and frame the future. A link to the CMIP7-ScenarioMIP paper, published earlier this year, outlining the new scenarios is here. Climate scenarios have been around for 40 years. Originally used as inputs to drive climate models, their scope and reach have expanded dramatically over time. The first scenarios considered only well-understood greenhouse gases and the solar cycle, and they were developed at a time when there was no climate policy. Gavin Schmidt provides a clear summary of three compelling reasons why climate scenarios change:

1. Actual emission trajectories: Climate scenarios start from real data. Including the current “new” scenarios, there have been five sets of scenarios and therefore five join points over the past 40 years: 1984, 2000, 2005, 2014, and 2023. CMIP7 climate scenarios are being run now, will likely be released in 2027 or later. Their join point is 2023, this means that real emissions are used up to 2023, and projected emissions are used for every year after. Every year after this is an educated prediction that aims to incorporate how changes in policy, legislation, the price of renewables, etc., may evolve.
2. How we view the future: The first climate scenarios were set around 40 years ago, in a world with no climate policy. “Business as usual” was a useful way of contextualising future change. Today, there are in the region of 7,000 climate policies worldwide. As a result, the language we use to discuss the future has changed. High, medium, low, aspirational, overshoot, and regressive is the more appropriate language.
3. Scope: The earliest scenarios provided projected concentrations of well-mixed greenhouse gases, the solar cycle, and a few volcanoes. Today, emissions of a wider range of greenhouse gases, land use changes, human-mediated impacts on dust and fire, and multiple aspects of volcanic forcing are all included. As we understand more about the climate, these are embedded in the models (climate models are now referred to as Earth System Models for this reason) and can be included in scenario development.

For CMIP7, scenarios are named after emission trends, Low to High. The figure below (left) shows the greenhouse gas (GHG) emissions associated with each scenario. The global mean temperature associated with these scenarios will be determined by the modelling experiments and the temperatures shown on the right were used to guide the scenario design. Below the figure is a description of each of the scenarios.

Figure 1 - Proposed scenarios for CMIP7 ScenarioMIP, showing (a) GHG emissions pathways as a function of time for each of the proposed scenarios and (b) the associated global average temperature outcomes using the probabilistic FaIR ensemble used in IPCC AR6 (Smith, 2025; Smith et al., 2024). Scenarios are (H) High, (HL) High-to-Low, (M) Medium, (ML) Medium-to-Low, (L) Low, (LN) Low-to-Negative and (VL) Very Low. Full details are available in Van Vuuren et al., 2026

High (H): Emissions grow to as high as deemed plausibly possible, consistent with a rollback of current climate policies. This scenario will result in strong warming.

High-to-low (HL): Emissions rise as in the high scenario at first, but are cut sharply in the second half of the century to reach net-zero by 2100.

Medium (M): Emissions consistent with current policies, frozen as of 2025, leading to a moderate level of warming.

Medium-to-low (ML): Emissions are slowly reduced, eventually reaching net-zero emissions by the end of the century.

Low (L): Emissions consistent with likely keeping warming below 2C and not returning to 1.5C before the end of the century.

Very low (VL): Emissions are cut to keep temperatures “as low as plausible”, according to the paper. This scenario limits warming to close to 1.5C by the end of the century, with limited overshoot beforehand.

Low-to-negative (LN): Emissions fall slightly slower than in the VL scenario, with temperatures just rising above 1.5C. Emissions then rapidly drop to negative to bring warming back down.

So what do these new climate scenarios look like compared to previous ones? The new CMIP7 emissions scenarios cover a smaller range of forcing than previous ones.

A comparison of the new scenarios against two previous generations, the earliest Hansen-A and Hansen-B and the Representative Concentration Pathways (RCPs) is shown below in Figure 2. SSP scenarios are not included on this plot but the would plot as a slightly wider range than RCPs. Data is plotted as time versus Effective Radiative Forcing (ERF), not temperature or emissions.

ERF is a metric that takes into account both the direct forcing from things like greenhouse gas or pollution as well as the indirect changes that are due to changes in clouds and the atmosphere as a result of the direct changes. ERF is measured in Watts per square meter.

The new highest emission pathway CMIP7 projection (CMIP7-H, deep maroon line) falls below the much older RCP-8.5 scenario (dashed black line) that was developed back in 2011. The Hansen-A scenario (solid black line), which was published in 1988, had an even larger trajectory than RCP8.5. Consequently, the CMIP7 scenarios cover a smaller range of plausible scenarios.

Figure 2 - Comparison of approximate effective radiative forcings across different sets of projections, starting with Hansen et al (1988), SRES (CMIP3), RCPs (+ extensions beyond 2100) (CMIP5) on top of the latest CMIP7 projections. This includes the direct effects of CO2, CH4, N2O and CFCs, but did not include aerosols (-ve) or ozone (+ve) (small effects on this scale) so there is a slight adjustment down to compensate. This may be subject to revision!) Original figure from Van Vuuren et al., 2026, this version from Real Climate.

RCP8.5 or the “business as usual” scenario, as it’s known, will be retired in the next generation of scenarios and model output. Many organisations and companies incorporate climate scenarios in their reporting and decision processes. And perhaps it’s because of this broader audience than the original science and policy crowd that the “what does this mean?” question has been so loud this time.

So, from a practical perspective, RCP8.5 being “retired” means that:

1. In the next generation of climate data (CMIP7) the High (H) scenario will be less extreme than SSP5-8.5 and RCP8.5 the previous worst case scenarios.
2. The worst-case scenario from this type of modelling means using the High (H) scenario in the future, which is less bad. Note the discussion of tail and tipping events later on.
3. If continuity between climate generations of climate scenarios is important, than consider using outputs from the next highest scenario currently available, SSP3-7.0. Bear in mind there is no firm timeline for CMIP7 data to be in production, the scenarios are only in production now.

From a more existential line of questioning, does the end of RCP8.5 mean something more? Does it mean that we’re doing ok climate wise and can all relax? Well, yes and no. Current emissions and policies track us to the middle scenarios today. Which is definitely a positive thing, but that still represents around 2.6°C warming above pre-industrial levels by 2100, as shown in Figure 3.

Figure 3 - Emissions pathways to 2100 from Carbon Tracker, which compiles historical emissions (solid black line) with projected warming scenarios.

So we’re in the clear, but what does that mean for grey swan events or tipping events? Well, this is where it gets a bit more unclear. Scientific understanding of tipping points is an active area of research, but a lot more research is required to understand these better. Watch out for our next blog post on tipping points if you’re interested.

A further complication which compounds confusion around the reduction in plausible emissions scenarios and the existence of extreme, dramatic shifts in climate regimes… is that the Earth System Models is that the that have previously run ScenarioMIP simulations do not fully represent climate tipping points.

As with many things, both things are true. Our current plausible emission trajectories have reduced due to things like climate policy, legislation, cost and availability of renewable energies. But at the same time the International Panel on Climate Change (IPCC) still warns with high confidence that the likelihood of triggering dangerous tipping points increases significantly with every increment of global warming, and some may already be within reach. Whilst at the same time acknowledging that the timing and thresholds of these events are still very uncertain.