Enduring Impact of Flawed Climate Science

Welcome to Dispatch Energy! The major energy and climate policy choices of the last decade and a half—deciding which structures to build, determining how financial regulators require banks to stress test against risk, and evaluating how the anticipated costs of carbon dioxide emissions shape regulations—rest on a foundation that has now been officially deemed outdated. Scenarios drive almost every facet of climate policy. Yet only a narrow circle of technically oriented experts understands why they matter or who wields influence over them.

A reference scenario—often called a baseline or business-as-usual path—serves to project the direction the world would take if no policy choices are changed. Other scenarios extend from this baseline to illustrate how the future might unfold if policymakers adopt measures that alter that expected course. The contrasts between the baseline and the policy-focused scenarios are typically employed to weigh the costs and benefits of alternative paths.

In April, the trio of baseline scenarios that international climate research had recognized as official—RCP8.5, SSP5-8.5, and SSP3-7.0—were deemed by the international panel responsible for selecting and ranking scenarios to be “implausible” due to their excessive projections of future carbon emissions from fossil fuel combustion. While the decline in anticipated future emissions is a welcome development, the implications of retiring these scenarios will be substantial, and that is the focus of today’s column.

A climate scenario is not a forecast of what will necessarily happen. It is a deliberate, structured “what if” exercise: clear, internally consistent assumptions about factors such as population growth, economic development, how energy is produced and consumed, and the extent to which anthropogenic and natural processes shape greenhouse gas emissions and other outcomes. The outputs of that exercise feed into earth system models, which project how temperatures, sea levels, and the frequency and intensity of extreme weather may evolve in the future. Policymakers, regulators, and courts then act on those projections and the policy and economic analyses that stem from them.

A small cadre of researchers, working within the Coupled Model Intercomparison Project (CMIP)—an initiative of the World Climate Research Programme—decides which scenarios modeling teams should use and how they should be prioritized in research. Once a scenario is developed and ranked, it becomes the default in nearly all projection-based studies reviewed by the United Nations’ Intergovernmental Panel on Climate Change (IPCC). National climate assessments, central-bank stress tests, lawsuits, and standards for infrastructure design can all hinge substantially on this line of inquiry and on the IPCC’s conclusions.

Yet everything changed in April when CMIP released a new scenario framework called CMIP7 for the forthcoming IPCC assessment. RCP8.5, SSP5-8.5, and SSP3-7.0—the three scenarios that had served as the body’s official reference points through two IPCC assessment cycles spanning nearly twenty years—have been retired, judged no longer to reflect the real world, especially in their energy-technology assumptions. Here are five of the most consequential implications.

National climate assessments in the United States, the United Kingdom, Germany, Canada, Australia, Japan, and the Netherlands all relied on either RCP8.5 (Representative Concentration Pathways) or SSP5-8.5 (Shared Socioeconomic Pathways) as their reference trajectory. The World Bank’s Climate Change Knowledge Portal, which delivers climate projections to more than 100 client nations, did the same. For more than a decade, those reference scenarios formed the basis for climate impact projections, shaping planning assumptions embedded in national energy and infrastructure policy.

A major driver behind the overestimation was a coal-centric assumption. RCP8.5 projected a fivefold surge in global coal use—more coal than the planet’s proven reserves could sustain—and it assumed the commercial deployment of coal-to-liquid technologies to substitute for oil on a scale never achieved before. The now-retired reference scenarios anticipated roughly 5°C of global warming by the year 2100.

Early estimates for the updated middle-ground scenario point to warming of about 2.5 to 2.7°C by the latter part of the century. That gap—shrinking from an anticipated 5°C to around 2.5–2.7°C—is substantial. Recalibrating national assessments to mirror actual trajectories will be an important but challenging task.

The retirement of the implausible reference cases lays bare a structural issue in climate science intended to inform policy: for fifteen years, projection efforts were tested against scenarios whose energy assumptions failed a basic plausibility check, without a formal mechanism to identify or rectify a flaw in scenario construction that everyone has long recognized as a significant weakness.

The committee attributes the retirement to “trends in the costs of renewables, the emergence of climate policy and recent emission trends,” characterizing implausibility as something that developed over time. That framing reverses the actual history, which Justin Ritchie and Hadi Dowlatabadi documented nearly a decade ago.

Reasonable people may disagree about where to set a high-end stress-test scenario to probe the robustness of policies and infrastructure in the face of low-probability, high-impact events. Different users of scenarios will have varying tolerances for risk and other viewpoints on how to construct a “worst-case” scenario. Still, CMIP7’s top-end high scenario—an exploratory case at the apex of their set—continues to assume a dramatic rise in coal use, which could again be unrealistic. Whether anyone consistently checks such scenarios for plausibility remains opaque.

The retirement of the traditional reference paths compels a reassessment of projected climate damages. Damage estimates rely on warming forecasts that themselves come from baseline scenarios.

Projected damages will, all else equal, be substantially lower when anchored to realistic scenarios—a world warming around 2.7°C yields different consequences than one approaching 5°C. The amount of emissions that require mitigation is also far smaller than the old baselines assumed, which has direct implications for the estimated costs of deep decarbonization in the revised baselines. Similarly, social cost of carbon estimates used in federal regulation and damage functions in financial risk models require updating.

A further outcome of CMIP7 is the removal of the scenario aligned with the Paris Agreement’s aspirational aim to cap warming at 1.5°C above pre-industrial levels by 2100. The World Climate Research Programme confirmed last month that its new low-emissions scenario no longer keeps warming under 1.5°C.

More actionable targets remain. For instance, I have long argued that a global agreement to phase out coal—establishing explicit dates, verification mechanisms, and enforceable compliance—offers greater potential for meaningful climate action than an aspirational temperature objective. Coal accounts for more than 40 percent of all carbon dioxide emissions from the energy sector. As of 2023, 84 countries had plans to move away from coal, but collectively they represented only about 30 percent of current coal-fired generation. A framework built around eliminating coal would be verifiable, negotiable among countries with widely different coal dependencies, and directly tied to the emissions cuts that actually influence the temperature trajectory.

Regardless of the ultimate targets chosen, it is now evident that the global community requires new aims in international climate policy.

While the reference scenarios may have originated for scientific purposes, they rapidly assumed pivotal roles in policymaking. The Network for Greening the Financial System, an international coalition coordinating climate stress testing for more than 140 central banks, originally built its “Hot House World” physical-risk scenario to align with RCP8.5. That scenario has since been used in stress testing by the European Central Bank, the Bank of England, the Reserve Bank of New Zealand, the Bank of France, and the U.S. Federal Reserve.

Any regulation, court ruling, or law grounded in the now-retired scenarios remains susceptible to legal challenges, and such challenges would rest on broadly accepted scientific foundations.

Policies that hold up across a range of possible futures make more sense than those optimized for or based on a single scenario. Robust decision-making does not require alarming or extreme projections. Forecasting is inherently difficult, but sound energy policy can accommodate what we do not yet know. Persisting with implausible scenarios risks eroding not only scientific credibility but also the trust in scientific institutions that durable climate policy depends on if it is to endure over the many decades needed to reach deep decarbonization.

CMIP’s decision to move beyond implausible scenarios was the right one. The challenging task of revising projections, renegotiating targets, and managing the legal and regulatory consequences now falls to everyone else.