RCP vs SSP: a simplified introduction for the general public

IPCC scenarios, in two minutes

To imagine tomorrow's climate, experts from the IPCC (the UN-mandated climate panel) use "scenarios". A scenario is a possible story: "What if the world did this or that, what climate would we have in 50 or 80 years?"

There are two main families of scenarios: RCPs, which are older, and SSPs, which are more recent. They are often confused. Here are the essentials in two minutes. Then, a point that almost no one looks at: the starting date of each scenario.

RCP: the insulating power of the blanket around the Earth

An RCP represents, metaphorically, the insulating power of the blanket of greenhouse gases surrounding the Earth at the end of the century. Imagine the Earth under a blanket: this blanket consists of greenhouse gases (CO₂, methane, etc.). The more there are, the more heat the blanket traps, and the more the planet warms.

The number in the name of an RCP indicates the intensity of this blanket's effect on the climate (in watts per square meter, but just remember this: the higher it is, the hotter it gets). There are four RCPs: RCP2.6, RCP4.5, RCP6.0, and RCP8.5. They were created for the IPCC's 5th assessment report (2013-2014).

An RCP does not say how we get there: nothing about the economy, energy, or political decisions. It gives the destination, not the path.

SSP: the story of the world that goes with it

An SSP adds what was missing from the RCPs: a story. For the IPCC's 6th assessment report (2021), experts envisioned five ways the world could evolve:

• SSP1: a world that chooses sustainability and restraint;

• SSP2: a world that continues roughly as it is today;

• SSP3: a divided world, where each region looks out for itself first;

• SSP4: a highly unequal world;

• SSP5: a world that bets everything on growth and fossil fuels (oil, gas, coal).

The name of an SSP is read in two parts. Take SSP2-4.5: the "2" is the story (the middle-of-the-road world), the "4.5" is the effect of the blanket on the climate in 2100. An SSP therefore provides both the path and the destination.

Who resembles whom?

SSPs and RCPs are cousins: when the blanket has the same insulating power in 2100, the two scenarios resemble each other. But they are not twins, because the baseline assumptions and the route taken up to 2100 are not the same. Here are four of the IPCC's five reference scenarios:

* Average over 2081-2100, relative to 1850-1900. Best estimate from the IPCC (AR6, Working Group I), which also provides a range around each figure. The fifth reference scenario, SSP1-1.9, the most restrained of all, is not included here.

For a business, however, the real question lies elsewhere: where does each scenario start from? That is the subject of the next section.

The point that gets overlooked: where does the scenario start?

Each scenario begins at a specific date, and that is what almost nobody checks. Before that date, computers recalculate the past climate based on what actually happened (measured emissions). After that, they follow the scenario's story, like a movie whose ending is unknown.

Think of a road map. If it dates from 2006, it does not show the roads built since. If it dates from 2015, fewer are missing, but some are still missing.

RCPs are used by a generation of climate models called CMIP5, and SSPs by the next generation, CMIP6. (A climate model is a computer program that simulates the Earth's climate.) RCPs start from 2006, SSPs from 2015. We are in 2026: an RCP therefore relies on around twenty years of assumptions rather than measurements, and an SSP on about eleven years.

Yet these years have been anything but ordinary. According to the annual scientific study "Indicators of Global Climate Change", warming caused by human activities increased by about 0.27 °C per decade over 2015-2024, an unprecedented pace since temperature records began. Greenhouse gas emissions also reached a record high over 2014-2023. A scenario starting in 2006 does not "see" these years: it replaces them with its own story, which may drift away from reality.

Two clarifications to remain fair. First, switching to SSPs does not solve everything: they are not recalibrated to what has happened since 2015 either. Second, several CMIP6 models are more "sensitive" than those of CMIP5: for the same amount of greenhouse gases, they warm up more, which partly explains hotter projections. The starting date is therefore only one criterion among others, but it is the simplest to verify.

At Tardigrade AI, our assessments rely on CMIP6 and a historical baseline period of 1996-2025, built from observed data (Copernicus ERA5-Land). The recent climate is thus measured, not assumed.

Every year, we update our projection data based on the indicators, values, and thresholds observed in previous years, whereas a published set of projections remains, by design, frozen at its publication date. Whether the scenario starts in 2006 or 2015, the gap with the actually observed climate is corrected every year instead of widening. This represents a huge difference compared to all the solutions available on the market, ensuring our clients use projections that are updated annually rather than frozen at a specific date. Climate is like a movie at 24 frames per second, not a still photo.

What this means for a business

To interpret a climate assessment accurately, ask five questions to whoever provides it to you:

1. Which scenario was used (SSP or RCP, and which one)?

2. Which model generation (CMIP5 or CMIP6)?

3. For what future date (2030, 2050, 2100)?

4. Against which past period was it compared? This is the "baseline period": the past to which the future is compared.

5. When were the projections last updated? That is the year up to which measurements were taken into account.

Two assessments that do not answer these five questions cannot be compared, just like two prices when you don't know whether one is in euros and the other in dollars.

To make decisions, companies generally use a central scenario for planning, and a high (more pessimistic) scenario to stress-test their assets: factories, buildings, inventory. Our article on SSP5-8.5 explains how to consider these high scenarios without turning them into an excuse for inaction.

For Horizon, we can use the full spectrum of SSP scenarios. However, when it comes to adaptation challenges, our recommendation is to use SSP5-8.5 because it unfortunately reflects the concrete effects of climate disruption as of today. Using a more optimistic scenario would therefore mean undersizing the adaptation measures to be implemented.

What comes next? CMIP7 scenarios

A new generation of scenarios is arriving: CMIP7. Seven scenarios were published in 2026 to prepare for the IPCC's 7th assessment report. Their names reflect emissions trajectories, from "very low" to "high", rather than a number representing the insulating power of the blanket.

They start from emissions actually measured up to 2023 and cover a median warming of 1.6 °C to 3.3 °C in 2100, compared to 1.5 °C to 4.7 °C for SSPs. The first model results are expected by the end of the year. The principle remains the same: the more recent the starting date of a scenario, the smaller the gap between the scenario and reality.

Frequently asked questions

CMIP5 or CMIP6: which one to choose in 2026? CMIP6, without a shadow of a doubt: it starts in 2015 rather than 2006 and relies on more recent models. In all cases, ask your provider to specify the model generation, the start date, and the frequency of projection updates based on recent observations.

Do SSPs replace RCPs? For the IPCC's 6th assessment report, SSPs are the ones used. However, RCPs are still present in a lot of existing data, which poses a real problem regarding the quality of the resulting climate projections.

Is SSP5-8.5 identical to RCP8.5? No. They share the same insulating power of the blanket in 2100 (8.5), but not the same assumptions or the same models.

Which scenario should a business choose? If the goal is adaptation, SSP5-8.5 is the only scenario to retain, because its projections are consistent with the climate drift observed so far.

How much warming in 2050? For 2041-2060, the IPCC estimates between 1.6 °C in the most sustainable scenario (SSP1-1.9) and 2.4 °C in the most fossil-heavy one (SSP5-8.5), relative to 1850-1900.

Conclusion

Understanding the difference between RCPs and SSPs is useful. Knowing the starting point of the scenario your assessment relies on is even more so: it tells you whether the last twenty years were measured or assumed. That is why we update our projections every year.

References

• IPCC, AR6 Working Group I, Chapter 4

• Forster et al. (2025), Indicators of Global Climate Change 2024, Earth System Science Data

• Carbon Brief, CMIP7 emissions scenarios (September 1, 2026)

• ClimateData.ca, differences between CMIP5 and CMIP6