- FEMA Crisis Triggers Worry Going Into Hurricane Season. - June 5, 2025
- Antarctica’s Growing Ice Sheet Surprises Scientists - June 4, 2025
- 10 Major Earthquakes That Went Unnoticed - June 1, 2025
The Prediction Problem That’s Bigger Than Weather Forecasts
You’ve probably heard the joke about weather forecasters being wrong half the time and still keeping their jobs. But what happens when the stakes aren’t just whether you need an umbrella, but the future of civilization itself? Fifty years into the project of modeling Earth’s future climate, we still don’t really know what’s coming. Some places are warming with more ferocity than expected. Extreme events are taking scientists by surprise. Right now, as the bald reality of climate change bears down on human life, scientists are seeing more clearly the limits of our ability to predict the exact future we face. It’s like trying to predict where a ping-pong ball will land in a hurricane – except the hurricane is our entire planet’s weather system, and we’re all standing in the path of that ball.
When Models Meet Reality: The Accuracy Problem
The authors found no evidence that the climate models evaluated either systematically overestimated or underestimated warming over the period of their projections. “The results of this study of past climate models bolster scientists’ confidence that both they as well as today’s more advanced climate models are skillfully projecting global warming,” said study co-author Gavin Schmidt, director of NASA’s Goddard Institute of Space Studies in New York. However, this confidence comes with a major caveat. But the recent findings about temperature extremes point in the other direction: The models may be underestimating future climate risks across several regions because of a yet-unclear limitation. And, Rohde said, underestimating risk is far more dangerous than overestimating it. It’s the difference between packing an extra umbrella and forgetting flood insurance – one inconveniences you, the other can destroy your life.
The Hot Model Controversy: When Climate Gets Too Spicy
CMIP6 includes several “hot” climate models whose sensitivity to greenhouse gas forcings exceeds the likely range inferred from multiple lines of evidence. CMIP6 includes several “hot” climate models whose sensitivity to greenhouse gas forcings exceeds the likely range inferred from multiple lines of evidence. Global warming estimates assessed in the Sixth Assessment Report (AR6) of the Intergovernmental Panel on Climate Change (IPCC) were reduced by applying observational constraints on the historical rate of warming to the CMIP6 ensemble. Think of these hot models as the overachievers in a physics class who always predict the most dramatic results. Around a fifth of the new CMIP6 models lie outside the very likely (90th percentile) equilibrium climate sensitivity (ECS) range adopted by the IPCC AR6, with 18% of CMIP6 models having an ECS above 5C per doubling CO2 and 27% of CMIP6 models having an ECS higher than the most sensitive model in the prior generation (CMIP5). Scientists are basically having to decide whether these models are being realistic or just really, really pessimistic.
The Data Quality Dilemma: Garbage In, Garbage Out
Climate models are only as good as the data that feeds them, and that’s where things get messy. “We have to approximate cloud formation because we don’t have the small scales necessary to resolve individual water droplets coming together,” Robert Rohde, the chief scientist at the open-source environmental-data nonprofit Berkeley Earth, told me. Similarly, models approximate topography, because the scale at which mountain ranges undulate is smaller than the resolution of global climate models, which tend to represent Earth in, at best, 100-square-kilometer pixels. Imagine trying to paint the Mona Lisa with a paintbrush the size of a house – you’ll get the general idea, but you’re going to miss some pretty important details. Models simply can’t function on the scale at which people live, because assessing the impact of current emissions on the future world requires hundreds of years of simulations. Modeling the Earth at one-square-kilometer pixels would take “like a hundred thousand times more computation than we currently have,” Schmidt, of NASA, told me.
The Cloud Conundrum: Why Sky Puffs Crash Computer Models
Clouds are the comedians of climate science – they look simple but they’re incredibly hard to understand. The Intergovernmental Panel on Climate Change (IPCC), the established global authority on climate change, acknowledges this in its most recent Assessment report, from 2013: The simulation of clouds in climate models remains challenging. There is very high confidence that uncertainties in cloud processes explain much of the spread in modelled climate sensitivity. Here’s the kicker: This contributes an average uncertainty of ±4.0 Wm–2 to the atmospheric thermal energy budget of a simulated atmosphere during a projection of global temperature. This thermal uncertainty is 110 times as large as the estimated annual extra energy from excess CO2. If our climate model’s calculation of clouds were off by just 0.9 percent—0.036 is 0.9 percent of 4.0—that error would swamp the estimated extra energy from excess CO2. It’s like trying to measure the weight of a feather while standing on a scale that’s bouncing up and down.
Extreme Weather: When Models Get Blindsided
Kai Kornhuber, a climate scientist at Columbia University, and his colleagues recently found that, on every continent except Antarctica, certain regions showed up as mysterious hot spots, suffering repeated heat waves worse than what any model could predict or explain. Across places where a third of humanity lives, actual daily temperature records are outpacing model predictions, according to forthcoming research from Dartmouth’s Alexander Gottlieb and Justin Mankin. It’s as if nature is constantly throwing curveballs that our best predictive technology just can’t see coming. The rationale of our study was therefore to approach the assessment of model ability differently—not through a classical evaluation of impact statistics over several decades—but instead to answer a what-if question: what if these models are used to predict the impacts of a single extreme event like the EHWD?. The answer, according to our results, is that the total impacts of such an event would likely be seriously underestimated.
The Regional Reality Gap: Global Models, Local Problems
Global climate models are like weather apps that tell you it’s sunny while you’re standing in a thunderstorm. Our picture of what is happening and probably will happen on Earth is less hazy than it’s ever been. Still, the exquisitely local scale on which climate change is experienced and the global purview of our best tools to forecast its effects simply do not line up. The world has warmed enough that city planners, public-health officials, insurance companies, farmers, and everyone else in the global economy want to know what’s coming next for their patch of the planet. And telling them would require geographic precision that even the most advanced climate models don’t yet have, as well as computing power that doesn’t yet exist. It’s the difference between knowing there will be rain somewhere in your state versus knowing whether your backyard will flood.
Feedback Loops: The Climate System’s Mood Swings
Climate feedback mechanisms are like that friend who either makes everything better or everything worse – there’s rarely a middle ground. Large biases remain in the representation of clouds in the Arctic region, with models suffering from an underestimation of supercooled liquid water in mixed-phase clouds. Think of the ice-albedo feedback as a cosmic game of mirrors: when ice melts, the dark ocean absorbs more heat, which melts more ice, which exposes more dark ocean. This response bias stems from the models’ inability to reproduce the observed large-scale surface warming pattern and from errors in the atmospheric physics affecting short- and longwave radiation. Models with a better representation of the TOA radiation response to local surface warming have a relatively low equilibrium climate sensitivity. It’s like trying to predict how a room full of people will react to a joke – one person’s laughter might set off a chain reaction, or it might fall completely flat.
Computing Power: The Bottleneck Nobody Talks About
The dirty secret of climate modeling is that we’re essentially trying to simulate the entire planet on computers that, while impressive, are still playing catch-up with the complexity of nature. However, currently feasible horizontal resolutions are limited to O(10 km) because higher resolutions would impede the creation of the ensembles that are needed for model calibration and uncertainty quantification, for sampling atmospheric and oceanic internal variability, and for broadly exploring and quantifying climate risks. A caveat is that climate models are extremely computationally expensive and therefore require the fastest available supercomputers. However, for many types of simulations, even those supercomputers are still not powerful enough to globally resolve several important climate processes (e.g. convection, clouds, atmospheric chemistry). It’s like trying to run the latest video game on a computer from the 1990s – technically possible, but you’re going to miss a lot of important details.
Machine Learning to the Rescue: Teaching Computers to Think Like Clouds
The branch of AI called machine learning — in which computer programs learn by spotting patterns in data sets — has shown promise in weather forecasting and is now stepping in to help with these issues in climate modelling. “The trajectory of machine learning for climate projections is looking really promising,” says computer scientist Aditya Grover at the University of California, Los Angeles. Similar to the early days of weather forecasting, he says, there is a flurry of innovation that promises to transform how scientists model the climate. “Data-driven deep learning models are on the verge of transforming global weather and climate modeling,” the researchers from the University of California San Diego and the Allen Institute for AI, write. One of the researchers’ key insights was that generative AI models, such as diffusion models, could be used for ensemble climate projections. The resulting model starts off with knowledge of climate patterns and then applies a series of transformations based on learned data to predict future patterns. It’s like teaching a computer to be a climate detective, finding patterns that human scientists might miss.
The Uncertainty Principle: Embracing What We Don’t Know
“It should be worrying that we are now moving into a world where we’ve kind of reached the limit of our physical understanding of the Earth system,” Kornhuber said. While models struggle to capture the world we live in now, the planet is growing more alien to us, further from our reference ranges, as the climate keeps changing. This isn’t scientific failure – it’s scientific honesty. While 2023 saw exception levels of warmth – far beyond what we had expected at the start of the year – global temperatures remain consistent with the IPCC’s assessed warming projections that exclude hot models, and last year does not provide any evidence that the climate is more sensitive to our emissions than previously expected. Scientists are essentially admitting that the climate system is like a teenager – just when you think you understand it, it does something completely unexpected.
