There’s been a lot of media buzz surrounding “the formal retirement of the RCP8.5/SSP5-8.5 scenario,” an event that some are framing as an admission “that climate projections were wrong.” Even a relatively skeptical voice on the subject of climate change has offered, however, a more nuanced view:

The truth is more complicated. The authors of the paper in which the scenarios were dismissed as “implausible” attributed the downgrade to “trends in the costs of renewables, the emergence of climate policy and recent emission trends,” an explanation that can be seen as a justification of current climate policy rather than a rejection of it. On the face of it, that’s fair enough.

The most vigorous dancing on the grave of RCP8.5 appears to be a reaction of “You tried to scare us! We won’t listen to you any more!” There has been some rueful acknowledgment that this isn’t entirely surprising, and that it could have been forestalled by a more careful presentation over the past several years.

For example, writing at Carbon Brief, Zeke Hausfather warned almost seven years ago that “The creators of RCP8.5 had not intended it to represent the most likely ‘business as usual’ outcome, emphasising that ‘no likelihood or preference is attached’ to any of the specific scenarios. Its subsequent use as such represents something of a breakdown in communication between energy systems modellers and the climate modelling community.” That’s putting it politely.

Unfortunately, once the phrase “business as usual” had been spoken by the telegenic Greta Thunberg, it became tempting and even fashionable for others to use those words – often in ways that made RCP8.5 seem like a prediction instead of a possibility.

A view from Canada provides a useful perspective on what has, and what has not, changed in the updated view:

Abandoning what are now outdated scenarios is “truly a good news, bad news story,” said Canadian Climate Institute president Rick Smith. “The worst-case scenario of how terrible things might get in terms of warming by the end of the century is not quite as bad as we thought,” he said. “But the best-case scenario is not quite as good as we hoped for.”

It’s crucial to appreciate that the retirement of RCP8.5 is not the same thing as an all-clear outlook for the planet. The RCP family of models are inherently optimistic in that they do not address, at all, some mechanisms of global warming that are coming to be understood as far more threatening than previously thought to be:

We currently have an incomplete understanding of the Earth’s carbon cycle. As Earth warms, it releases its own emissions, like methane from wetlands. These types of emissions are not well represented in the models which produce the RCPs. The only “tool” that these models have to increase GHG concentrations is by increasing emissions from fossil fuel combustion; but this does not mean that these are the only sources of GHG emissions. These carbon “feedbacks” mean that a pathway with less burning of fossil fuels but greater emissions due to land use change or natural feedbacks could result in higher levels of warming than captured in the RCPs.

From another commentary, “Although the lower ends are comparable in both sets of simulations, the highest climate projections are significantly warmer in the emission-driven simulations because of stronger carbon cycle feedbacks.” I’m not sure that “feedbacks” is the best word for this. Something like “cascades” might be better, but here is a description of one effect that RCP models don’t represent:

Like a warm can of Coke, which loses its carbonation when you warm it up and remove the top, the ocean's CO₂ solubility decreases as the ocean warms… Current projections of future warming may actually underestimate the degree of warming expected from a particular carbon emissions pathway.

If that seems a little vague and “trying to be scary,” let’s be more specific:

Impacts on ice sheets are a very important part of the climate change and have yet to be fully integrated into the standard climate models. So independent efforts with ice sheet models were set up using the output from the CMIP5 and CMIP6 models – they used two scenarios, RCP85 and RCP26 to bracket possibilities in ISMIP6 (Seroussi et al, 2020). Interestingly, they found that, particularly for Greenland, none of the models had melt rates as high as observed even with RCP8.5 forcing. Thus for a situation where the (ice sheet) models are insufficiently sensitive, a higher than expected forcing might give you a more likely outcome. [My italics added]

Thus, RCP8.5 remains a useful tool for anticipation and planning, even if its specific assumptions are now considered “implausible”:

The main scholarly criticisms of RCP8.5 have been that it was misrepresented as a “business as usual” scenario, and that its estimations for coal burning, specifically—as opposed to other sources of emissions—were unrealistically high. Yet the kind of climate impacts that happen when RCP8.5 is plugged into a global climate model are very much not off the table.

This will get to be (even) far(ther) too long unless I defer additional discussion of non-obvious greenhouse gas emissions. I want to end, though, with a vivid analogy from Marshall Shepherd at Forbes that may illuminate, in simple terms, what has really happened in the RCP8.5 withdrawal:

Let’s say a person goes out drinking at the bar. There are a range of outcomes that we could predict a few hours later. They include:

• He just drinks soda all night and watches the game. Drives home safely.
• He has a couple of beers but remains below the legal limit and gets home safely.
• He has 4 to 5 beers and is above the legal limit and takes a risky drive home or calls rideshare.
• He has 10 beers and drives home under extreme impairment or calls a rideshare.
• He has 10 beers and starts a bar fight that results in multiple injuries and several arrests.
• He drinks all of the beers at the bar and in the cooler too with expected dire outcomes.

What the UN panel did, from an analogy standpoint, is take the last scenario off the table. He was never going to drink all of the beer, but there are still very problematic scenarios on the table.

An extreme worst case set of behaviors, globally, is no longer considered likely – but the need to anticipate comparable effects remains.