Why think of the worst? Scenarios of Climate Pathways at Resilens

Climate skeptics blamed scientists for faulty modeling when scientists retired the worst-case high-emissions scenarios. Have we successfully avoided the worst?

• The retirement of RCP8.5 was a scenario-hygiene move; climate risk continues to grow.

• Even under medium-emissions scenarios, temperatures of 4 to 5°C by 2100 remain an unlikely but not impossible outcome that adaptation planning has to account for.

• An adaptation plan is only fundable and defensible when the scenario behind it is documented and transparent.

The worst case scenario is dead: what changes for adaptation planning?

More than a few eyebrows were raised when the climate modeling community retired the extreme future emissions pathways from the climate models that simulate future rises in temperature. (They also retired the best-case scenario - but that caused less flutter). 

Needless to say, there was discourse. The retirement was quickly weaponized as evidence that climate scientists had overreached beyond what could be considered plausible. Pioneers of the move - Zeke Hausfather, Glen Peters and Piers Forster - had been clear from the start  that the worst-case scenario was in fact never meant to predict the probable outcome of a world that did nothing to prevent climate change, and has been unnecessarily used as the business-as-usual scenario everywhere. They also credited the progress of global decarbonization as the reason why the extreme case scenario was more unlikely than ever. 

It’s true that successful climate change efforts that reduced our dependence on fossil fuels are to be credited for the change. Thanks to the increasing use of renewable energy sources and slowing global population growth, we were no longer headed for the dire case where temperatures could plausibly increase in the range of 3.5-5.5 degrees by 2100. This is a sign of progress in climate action. It does not necessarily correspond to a reduction in the risk that climate adaptation professionals have to handle.

How does this impact what scenarios are used in climate adaptation decisions?

Human activities affect all the major climate system components, with some responding over decades and others over centuries

Global surface temperature change relative to 1850–1900

IPCC AR6 WG1 SPM.8: Global surface temperature projections. Source: Figure AR6 WG1 | Climate Change 2021: The Physical Science Basis

The response of Hausfather and others was in fact more complex than what appears at first glance. While it’s incorrect to say that the worst-case scenario was reflective of reality, the question of what reality should look like, is much harder to answer. All scenarios are “counterfactual” worlds - we do not actually know where we are headed. 

Hausfather suggests looking to the scenario literature to verify how the RCP8.5 compares to the average of historical baselines for predicting global temperature increases. The data there is clear - the RCP8.5 (Representative Concentration Pathway) lies on the extreme higher end of the range of historical baselines, making it an implausible candidate for the business as usual scenario.

CO₂ Emissions Scenarios and Projected Warming

RCP8.5 was the high end of baselines, not the most likely no-policy outcome

Global CO2 emissions (left) and 2100 warming relative to preindustrial (right) for RCP8.5, the range of IPCC AR6 WG3 baseline scenarios, and the new CMIP7 medium illustrative scenario from van Vuuren et al. Future warming ranges based on FaIR model calculations.
Image from Climate Brink.

However, Hausfather - and other climate experts and government bodies - make it clear that we are not doing enough to prevent climate change. Scenarios provide ranges for increases in temperatures, and single point predictions (such as a 3.5 degree increase in temperature as predicted by the IPCC baseline range) hide a wide variation in temperature increases. In other words, the chances of a 5 degree increase in temperature by 2100 even under RCP4.5 are small, but never zero. 

Moreover, as emissions continue to grow, we are nowhere close to stabilizing global temperatures. The earth continues to warm, and even the IPCC mean baseline scenario leads to a 5 degree increase in temperature by 2150 - a date that is less far off in the future now, than it was in the 1990s, when climate models were first conceived. As Detlef van Vuuren, the team lead of the scientists that retired RCP8.5 himself succinctly described: “we are in a much worse situation than we were in 15 years ago.”

Experts therefore agree on a couple of things. The IPCC-assessed scenarios were never meant to reflect absolute reality, and the lowered temperature predictions do not translate into lower climate risk. They’re also clear on something more nuanced: scenarios used by the many players in climate change analysis need to be transparent about where and how scenarios are applied in analysis. Transparency, not the rejection of emission pathways, is the recommendation. 

How should this transparency work in practice? 

Scenarios at Resilens

At Resilens, we implement climate scenarios with the practice of transparency in mind. Using single point estimates for estimating future climate risk may be misleading given the wrong context. Moreover, physical climate risk often combines with socio-economic factors (such as national environmental policy) to shape pathways across a range of future conditions. 

We therefore use two explicit, Resilens-defined scenarios that clarify the end use of the scenario in our platform, the emissions pathways used in the scenario, as well as the underlying socio-economic assumptions that accompany the physical risk model used. The Resilens-defined scenarios explicitly align different climate scenarios with the different use cases of the Resilens platform. 

Our climate data at Resilens uses CMIP6 projections. We consider four temperature planning periods at Resilens: historical, reference, present and projected periods. According to CMIP6 data availability, we use bias-corrected historical temperatures for the historical and reference planning periods. We use temperature projections for predicting future temperatures depending upon two different scenarios considered: baseline and risk-assessment. 

Baseline: Following recommendations from the latest Scenario Model Intercomparison Project and the scientific community, the baseline scenario at Resilens corresponds to the business-as-usual contexts covered by the Resilens platform. Baseline corresponds to the use cases where we expect middle-of-the road trends that correspond to policy baselines that are similar to current policies: it therefore covers the use cases of calculating costs and benefits from different adaptation measures at Resilens.

To this end, we use the CMIP6 scenario for future temperature projections that corresponds to SSP (shared socio-economic pathways) SSP2-4.5 and RCP 4.5. SSP2-4.5 describes policy pathways where challenges to mitigation stay medium to low, while challenges for adaptation stay medium to high, making it relevant for baseline adaptation planning. 

Risk assessment: We use the risk-assessment scenario at Resilens to cover the use cases where we are assessing heat risk, such as the Resilens Heat Risk Score. This corresponds to cases where we want to evaluate the hazard sites that could potentially be faced with. For this evaluation, the worst case scenario still holds merit, as long as it is applied with transparency. 

Our reasoning for this is twofold: first, contrary to popular belief, the IPCC does not have official recommendations yet limiting use of scenarios to RCP4.5, making the worst case scenario still relevant for climate analysis depending upon the use case. 

Second, as shown below, for CMIP6 data and the time periods we consider for temperature projections (Resilens Present and Projected Climate, or 2015-2055), the differences for average temperatures between CMIP6 scenarios SSP2-4.5 and SSP 5-8.5 are not in fact as large as one may imagine - however, the full range of probable high temperatures is higher under SSP5-8.5, which comes into relevance for risk analysis. High-risk heat events such as heatwaves faced in May and June earlier this year can lead to local spikes in temperatures; and since future temperature projections cannot account for heat-risk related weather events, considering the worst-case takes these extreme heat events into account better than the baseline scenario when considering heat risk.

Surface Air Temperature (land only)

Source: Ocean Health Index

At Resilens, adaptation planning happens with transparency in mind. We inform adaptation practitioners and policy makers about which scenario fits which adaptation question instead of covering all use cases with a one-scenario-fits-all approach. That means looking at the use case, the overall risk, and the socio-economic pathway behind the numbers before deciding on our choice of a climate scenario. The world is still warming, and as the need for climate action continues, the business-as-usual case can no longer be applied by default. As Zeke Hausfather wrote about the "absolutely gobsmackingly bananas" temperatures following the El Niño, "there is likely a great deal of intensification still to come."