The release of the CMIP7 ScenarioMIP framework has prompted renewed discussion about the future of climate scenarios, particularly the reduced prominence of high-emissions pathways such as RCP8.5 and SSP5-8.5. This raises an important question for users of climate risk information: if these scenarios are now considered less plausible than they once were, should severe climate stress tests based on them be abandoned?
Here, Paul Young, Head of Science at JBA Risk Management, explains why the answer is nuanced, what CMIP7 means for physical climate risk analysis, and why less plausible emissions scenarios do not automatically imply less severe physical climate risk.
The full CMIP7 Earth-system model ensemble is still emerging, so discussion at this stage is based primarily on the newly published scenario framework. In that framework, the highest-emissions pathways used heavily in CMIP5 and CMIP6 are now viewed as less plausible representations of socio-economic development than they once were, reflecting changes in observed emissions trends, energy systems, technology deployment, and climate policy over the past decade.
Yet, emissions scenarios describe how society might evolve and how much greenhouse gas might be emitted. They do not uniquely determine the resulting climate. While higher-emissions pathways generally increase the likelihood of higher warming, the relationship is not one-to-one, and a given scenario label does not correspond to a single physical climate outcome.
As a result, the reduced plausibility of some emissions pathways does not automatically imply that the physical climate conditions explored under them are no longer relevant for risk analysis. However, it does have important implications for how those analyses are interpreted and used.
Summary: What should users do now?
• Do not assume that existing RCP8.5- or SSP5-8.5-based physical risk assessments are obsolete.
• Existing analyses may remain useful as severe physical climate stress tests, but they should not automatically be interpreted as central estimates of future socio-economic development.
• Be explicit about whether a scenario is being used as a forecast or a stress test.
• Expect future frameworks to place greater emphasis on warming levels and physical climate storylines alongside traditional scenario labels.
• Focus on decision robustness across a range of plausible climates rather than finding a single "correct" scenario.
Scenario labels do not uniquely determine climate outcomes
The key point is that emissions scenarios do not uniquely determine physical climate outcomes. Scenario labels, such as RCP8.5 or SSP2-4.5, are shorthand descriptions of greenhouse gas concentrations, emissions trajectories, socio-economic assumptions, and land-use change. They are useful as a common language across climate science, impacts research, finance, and policy. They are not deterministic descriptions of a single future climate.
The physical response of the climate system to a given emissions trajectory remains substantially uncertain, despite advances in our understanding. Different climate models can produce substantially different warming outcomes for the same forcing pathway because they represent key physical and Earth-system processes differently.
Figure 1 illustrates this point using assessed warming projections from CMIP6 simulations, which informed the IPCC’s Sixth Assessment Report. Even when the scenario is fixed, the range of plausible warming outcomes remains substantial. By the end of the century, the warming ranges associated with different scenarios overlap considerably, meaning that some outcomes under lower-emissions pathways overlap with outcomes under higher-emissions pathways.
The key point is not that the scenarios are interchangeable. Higher-emissions scenarios remain more likely to produce higher warming outcomes. Rather, the figure illustrates that the relationship between scenario label and the physical climate outcome is not one-to-one.
The ScenarioMIP framework explicitly recognises this uncertainty and places greater emphasis on “emission-driven” modelling to better capture uncertainty in carbon-cycle and Earth-system responses.
At the same time, climate models remain imperfect representations of the Earth system. Important uncertainties remain around carbon-cycle feedbacks, regional circulation changes, extremes, and tipping points. In practice, uncertainty grows as we move from global mean temperature to regional hazard variables such as rainfall intensity, drought persistence, or flood-generating weather patterns.
This has an important implication for climate risk analysis. Even if very high emissions futures are now considered less plausible than they were a decade ago, the range of plausible physical climate outcomes under lower-emissions pathways remains substantial. The move away from RCP8.5 and SSP5-8.5 does not eliminate the possibility of severe physical climate outcomes or the need to assess resilience under them.
Why this matters for stress testing
Climate stress testing is not primarily an exercise in forecasting the most likely socio-economic future. It is an exercise in understanding vulnerability to severe but plausible physical conditions. For that reason, we expect climate risk analysis to continue evolving away from heavy reliance on scenario labels alone and toward more explicit physical climate storylines.
For example, one might explore a UK climate storyline associated with approximately 3°C of global warming, warmer North Atlantic sea-surface temperatures, and a tendency toward more positive winter NAO conditions, and then assess the implications for rainfall extremes, drought, or wildfire weather and the resulting implications for risk.
This kind of framing is often more transparent and decision-relevant than treating a scenario label as though it uniquely determines a future climate.
In that sense, the ScenarioMIP framework should not be read as narrowing the range of climate risks that organisations should consider. Rather, it is part of a broader shift toward clearer separation between socio-economic pathways, physical climate uncertainty, and the information needed to support adaptation, resilience, and investment decisions.
For users of climate risk information, the practical implication is not to abandon severe climate stress tests, but to become more explicit about what those stress tests represent and why they are relevant. The central challenge in climate risk analysis remains understanding physical uncertainty and decision robustness, rather than relabelling scenario sets.
The move toward more careful treatment of scenario plausibility is welcome. But severe physical climate outcomes – and the need to test resilience against them – remain central to robust climate risk management.