While the high levels of fossil fuel emissions mapped in the RCP 8.5 scenario are unlikely, it remains an important scenario for modeling future climate impacts, particularly at local scales.
There has been a lot of confusion over something called “RCP 8.5” and we write here to set the record straight.
Back in 2020, two climate researchers published a comment letter in Nature arguing that the use of the emissions model known to specialists as “RCP 8.5” was “misleading.” These researchers noted that while RCP 8.5 had often been labeled the “business as usual” scenario, it in fact represented an emissions trajectory that our slowly-but-steadily decarbonizing world was unlikely to follow. They advised researchers to either stop using RCP 8.5 altogether, or label it an “unlikely worst case” scenario.
This argument received a lot of attention — it has now been cited 1372 times — and was almost immediately seized upon by right-wingers who transformed it into a conservative talking point that climate scientists are deceiving us into thinking that things are worse than they actually are. Relying on RCP 8.5 at all became categorized as a scare tactic, meant to shock us into action. President Trump took up this line of attack in his Executive Order on “Restoring Gold Standard Science” released last May.
This is all incredibly unfortunate, as RCP 8.5 is in fact a crucial tool to help us understand the climate impacts that lie ahead — even as the emissions trajectory it represents, fortunately, becomes less likely. Why are scientists still using RCP 8.5 in their research? They are not out to shock or deceive; they are simply following the best science. That RCP 8.5 has become unfairly understood to be a “controversial” choice of scenario is a distortion of climate science and we seek here to correct the narrative. The details of why RCP 8.5 remains not only legitimate, but crucial, are going to get technical fast, but bear with us.
First of all, what is RCP 8.5? Starting in 2014, the Intergovernmental Panel on Climate Change began using a series of different possible standardized emissions trajectories that could be used as inputs to global climate models. These trajectories — called “Representative Concentration Pathways,” or RCPs — represented a range of different climate futures based explicitly on projected concentrations of greenhouse gases (GHGs) for the year 2100. There were initially four RCPs: 2.6, 4.5, 6.0, and 8.5, with this last RCP representing the highest level of cumulative greenhouse gas emissions, and therefore also the hottest future.
If you are with us so far, it’s about to get more confusing. RCP 8.5 is meant as an example trajectory that could lead to a certain level of cumulative emissions (and hence warming) by the end of this century. Meaning, the scenario itself traces greenhouse gas concentrations (in the case of RCP 8.5, around 1100 ppm of CO2 by 2100), but the label for the scenario “8.5” is an estimate of the strength of the greenhouse effect resulting from those concentrations of GHGs. These are not the same thing, obviously, and this jump has led to a lot of confusion, in no small part because our scientific understanding of the relationship between atmospheric GHG concentrations and global average temperature rise has always been subject to substantial uncertainty—and debate. For instance, what if the emissions levels mapped in low or mid-tier scenarios like RCP 6.0, in fact lead to temperatures conventionally associated with RCP 8.5?
The extent to which these estimates are “off” is a deeper issue than a confusing nomenclature. There are several reasons why the emissions paths modeled in the RCPs could lead to higher temperatures than they were created to represent. One is simply that the earth’s warming response could be more sensitive to greenhouse gas concentrations than we originally modeled. Just a few years ago, some scientists argued that certain global climate models considered by the IPCC ran “too hot,” meaning they projected unrealistically hot temperatures for certain emissions levels. But the past few (very hot!) years have had these very same scientists forced to bite their tongue… maybe those models weren’t too hot after all. Recently we’ve heard more talk of a potential “cold bias” in the models.
Another reason is that the RCPs model fossil fuel emissions only. 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. Moreover, how each climate model responds to each RCP is strongly influenced by how they represent the varying Earth-system processes and feedbacks (such as those related to the carbon cycle), which are subject to considerable scientific uncertainty.
You might say, well that’s unnecessarily confusing! If RCP 6.0 emissions levels could lead to RCP 8.5 impacts, why don’t you just call the scenarios something else? Or else fix the underlying models themselves so that RCP 6.0 emissions lead to… RCP 6.0 impacts… that would appear to be a better (and clearly less confusing) outcome. The difficulty comes, in part, from the cumbersome multistep process of how climate modeling is done. “Re-running the model” is far easier said than done. Most climate researchers don’t work directly with global climate models, as they are so enormous they can only run on government-funded supercomputers. And each model run takes a significant amount of energy and time to produce. This is part of why we have the IPCC process itself: it’s meant to standardize inputs (like emissions scenarios) so that we can compare outcomes. In fact, one of the primary purposes of the RCPs is to benchmark how each climate model responds to a standardized set of emissions so that we can understand which models might tend to run “hotter” or “colder.” For this purpose, RCP 8.5 is arguably the most valuable scenario due to its higher levels of GHG emissions: any model will have its biases, and the stronger “signal” of RCP 8.5 makes it easier to elicit and identify them, whereas the lower-emissions scenarios, with weaker signals, can have more “noise” — that is, more uncertainty about what could be causing a particular outcome in the modelling. As a result, RCP 8.5 has historically been the scenario run for every climate model which the IPCC uses.
Thus far, we’ve focused on global average temperatures, the typical subject of a lot of this debate. This is what people are talking about when they mention policy goals like 1.5°C or 2°C of warming. But global average temperature is an abstraction, as climate scientist R. Saravanan memorably remarked: No one lives in global average temperature land. Some places will warm much, much more than 2° or 3°C degrees. And temperatures on a particular day, in a particular place, could reach extremes far, far, higher than these averages. There’s a lot of important science involved in trying to predict those local impacts—not only of temperature rise, but of precipitation events, storms, wildfires, etc.
Assessing these impacts and risks starts to get complex very quickly. Even if we are roughly right about the Earth’s sensitivity to atmospheric greenhouse gas concentrations at the global level, there are still additional reasons why RCP 6.0 emissions couldlead to RCP 8.5-type impacts at the local level. Some climate models may have biases towards smaller or larger impacts in different areas of even the United States, let alone the rest of the planet. Moreover, due to their computational complexity, global climate models use a very coarse resolution of the Earth’s surface, which can result in missing certain mechanisms which can influence extremes (such as the role of vegetation in influencing moisture) As a result, more local impacts require downscaling to “fill in” the gaps for most locations. But different downscaling methods (some of which can struggle to extrapolate how extremes might change under future warming conditions) can also result in considerable differences in projected impacts. And the problem of model biases shows up here as well: different models have different levels of skill in different areas, which propagate to downscaled assessments.
Due to these uncertainties, the gold standard for assessing regional or local risks is to use ensembles of downscaled climate model simulations. But we don’t have unlimited computational capacity or bandwidth to downscale or use all of the projections corresponding to every combination of climate scenario, climate model, internal variability, and downscaling method. This raises the question of what runs to prioritize for downscaling and analysis. Given all of these nested uncertainties, using high-emissions runs like RCP 8.5 can capture what local extremes could look like even under lower emissions scenarios. And for climate-proofing large investments like infrastructure, RCP 8.5 provides the ultimate stress test: if your plan can survive RCP 8.5-like impacts (which might result from an amplification of extremes under RCP 6.0 or similar warming), it is likely to perform well under less severe outcomes. And if it can’t, understanding how it fails under these more extreme conditions can be informative about any potential adaptations or improvements.
So from the perspectives of both model benchmarking and local risk assessment, RCP 8.5 serves a critical role. Then what is all of the controversy about? The answer lies in yet another use of these scenarios: assessing the plausibility of future climate outcomes. At its core, the idea is that the plausibility of the underlying emissions scenario reflects the plausibility of the range of outcomes from that scenario. This was the framing used by the authors of the original comment in Nature in 2020 that kicked all of this off: that the emissions associated with the specific RCP 8.5 trajectory were effectively impossible given the progress the world had made towards decarbonization and the future pledges made to help achieve the Paris Agreement targets. Prior work had shown that the specific RCP 8.5 trajectory was based on burning what seems like an unrealistic amount of coal. From this one lens, the argument about not using RCP 8.5 seems reasonable.
But, as previously noted, these emissions (and the specific fuel consumption used to generate them) are just one example, from one simulation of a model, of how we might get to that level of GHG concentrations. The models that produce these emission trajectories are also notoriously bad at capturing emissions from land use change or may miss other emissions such as methane flaring from oil and gas wells, which would increase GHG concentrations. And we might get to projected RCP 8.5 levels of warming from lower emissions scenarios, depending on how the Earth-system uncertainties break. Or we might achieve RCP 8.5 levels of warming with lower levels of emissions. In any event, RCP 8.5 may track emissions more closely over the next few decades than other scenarios. All of these are arguments that, even if we grant the underlying premise that RCP 8.5 as a scenario of emissions from the combustion of fossil fuels through 2100 is implausible, it serves an important role in our understanding of and planning for climate change, and its use is not misleading.
