Bad tidings for crystal ball enthusiasts: the future is unknowable. Even our best, data-based attempts to predict exactly what’s to come are far from certain, as highlighted in new research. Climate models that offer timing predictions for critical tipping points are woefully imprecise, according to a study published August 2 in the journal Science Advances.
There are many certainties in climate sciences: It’s certain that climate change is occurring and that it’s having serious negative impacts on Earth’s equilibrium and ecosystems. As human-caused greenhouse gas emissions continue and global warming unfolds, it’s also well-established that we’re getting closer to crossing important thresholds, or tipping points, past which we face inevitable collapse of planetary stabilizing systems like the major forests or ice sheets. At a tipping point, a small change in one variable like atmospheric temperature or precipitation triggers a big fall-out, such as the irreversible changeover of the world’s largest tropical rainforest into dry grassland or unstoppable polar melt and sea level rise. We know that by continuing to burn fossil fuels, we’re increasing the risk that these things could happen.
But what’s far less definite is when, exactly, we’ll plummet off those precipices, according to the new analysis. A subset of recent climate science has focused on developing models to predict the timing of future major events. Yet these projections are subject to too much variance to be meaningful — in large part because of the many assumptions modelers have to make to come to any one number or range.
“The uncertainties are too large to reliably estimate tipping times.”
“Even under the assumption that a given Earth system component has an approaching tipping point, the uncertainties are too large to reliably estimate tipping times,” write the study authors.
The researchers focus in on the particular example of the gulf stream, known to scientists as the Atlantic meridional overturning circulation (AMOC). Broadly: the AMOC is a massive marine current that moves cold, deep water south and warm, surface water north on either side of the Atlantic Ocean. The thermal conveyor belt helps to equalize global temperatures, rainfall, and sea levels. Without it, Europe would be much colder, the southern hemisphere would be much warmer with longer dry seasons, and parts of the U.S. eastern seaboard would be under water.
Worrying data on sea surface temperature and ocean sediments published in 2018 indicates that AMOC has slowed to a 1,500 year low. Other research indicates AMOC is showing early warning signs of destabilization. Most alarming of all was a 2023 study published in the journal Nature Communications that predicted AMOC would collapse around the middle of this century (i.e. within decades). That paper prompted lots of media coverage and both supportive and skeptical responses from the scientific community.
The new study is, in a way, another response to the 2023 article — but also to the field of tipping point modeling broadly. The researchers identified four uncertainties inherent in current modeling processes, which they say makes homing in on accurate time estimates impossible. First, models are simplified and we don’t have a perfect understanding of all of the variables involved in these systems. Second, models often rely on assuming past warming and emissions will continue following the same pattern as the past, which isn’t reflective of reality. Third, indirect indicators like sea surface temperature might not be 1:1 assessments of the stability of complex systems like AMOC. Finally, sparse historic data or contemporary data from multiple sources often leaves scientists filling in gaps and processing data to align — inevitably introducing errors.
They revisited the models used to come up with that mid-century AMOC collapse prediction published in 2023, and made adjustments in accordance with the above uncertainties, testing how slightly different parameters or inputs would change the outcome. The scientists found that, under certain scenarios, their altered models predict AMOC collapse by 2050. But in others, their timing predictions stretched more than 6,000 years in the future to an AMOC collapse forecast in the year 8065.
“The conclusions of this study are certainly in line with my understanding of the current state of the art,” says Gavin Schmidt, a climate scientist and professor at Columbia University and the director of NASA’s Goddard Institute for Space Studies (GISS). Schmidt was not involved in the new work, but has extensively researched climate variability and systems like AMOC. “I have not been impressed by previous or recent efforts to predict upcoming tipping points in either AMOC or ice sheets — there is more going on than just patterns in time series and we still don’t have sufficiently complex and calibrated models to have a robust idea of what will happen,” he tells Inverse.
“I have not been impressed by previous or recent efforts to predict upcoming tipping points,” says Gavin Schmidt, director of NASA’s GISS
Additional research and longer-term data collection could improve modeling moving forward, the study authors note, but there will always be some degree of uncertainty. And in the present, models that can tell us an Earth-changing tipping point will come tomorrow or in 6,000 years aren’t particularly useful.
“Our research is both a wake-up call and a cautionary tale,”said Maya Ben-Yami, lead study author and a doctoral researcher at the Potsdam Institute for Climate Impact Research in Germany, in a press statement. “There are things we still can’t predict, and we need to invest in better data and a more in-depth understanding of the systems in question. The stakes are too high to rely on shaky predictions.”
At the base of climate modeling is the idea that, if we could know for sure we’re headed towards imminent catastrophe, we might redouble our efforts to shift course. Yet we already know plenty to understand that moving away from fossil fuels is critical to avoid ever-worsening consequences.
The new study doesn’t mean that AMOC won’t collapse this century, nor that the Amazon rainforest isn’t on track to become a savannah, nor that Greenland and Antarctica aren’t headed towards runaway melting. All it means is that any of those things could happen and we aren’t able to know when.
“The large uncertainties imply that we need to be even more cautious than if we were able to precisely estimate a tipping time,” said Niklas Boers, study co-author and a professor of Earth system modeling at the Technical University of Munich, in the press release. “We still need to do everything we can to reduce our impact on the climate, first and foremost by cutting greenhouse gas emissions. Even if we can’t predict tipping times, the probability for key Earth system components to tip still increases with every tenth of a degree of warming.”
