Have a question?

Will climate change make weather forecasting less accurate?

No—because modern weather forecasts use atmospheric physics instead of historical data, they should be reliable even in a world with a changing climate.

 

January 30, 2023

To predict the future, one must understand the past. Part of forecasting the weather is knowing a place’s meteorological history, so we can check whether the atmospheric conditions we’re seeing today led to sun or rain in the past. Because of climate change, however, today’s weather patterns look different than those of even a few decades ago. Does this mean our weather forecasts will grow less reliable?

Luckily, no, says Kerry Emanuel, professor emeritus of atmospheric science at MIT. The reason is that modern weather forecasting uses totally different methods.

Up until about the 1950s, weather forecasting relied on a kind of modeling based on statistics from the historical record. That approach would be problematic in a world with a rapidly changing climate. But with today’s far more powerful computers, modern meteorologists can take real-time atmospheric measurements and plug them into a model of the Earth’s surface and atmosphere, to simulate conditions days into the future.

“It is basically an algorithm for solving differential equations that govern the behavior of fluids, radiation, oceans, the atmosphere, cloud physics, and more,” Emanuel says. Even if the Earth’s climate changes, the physics that govern these forces do not. “The model is solving physical equations that should be valid no matter what the climate is.”

It is possible, Emanuel says, that climate change could affect the weather forecast, but not because we’ll have a harder time comparing today’s temperature to the historical record. "There is an interesting question about how climate change might affect the predictability horizon,” he says. “In other words, does the weather intrinsically become more predictable or less predictable?"

So far, answers to that question have been a mixed bag. For example, Emanuel’s own study of hurricanes found that it may be harder to predict hurricane intensity in a warmer world.1 The reason is that a hurricane can weaken or strengthen more rapidly than before, and things that change quickly are inherently hard to predict. On the other hand, winter storms like the thunderheads that roll across the American Midwest might grow more predictable as climate change makes them unfold more slowly.

While this remains an interesting and under-explored area of research, Emanuel says, what matters most is the dramatic improvement in the science of weather forecasting. Today’s 7-day forecast is as accurate as the 3-day forecast was in the 1980s, an improvement that blows away any smaller uncertainties arising from climate change. The credit goes to bigger data sets, better real-time data, and, most importantly, the continual rise in computational power, which lets forecasters simulate the exceedingly complex weather system with ever-better accuracy.

That trend won’t continue forever. Emanuel says there is a theoretical limit to how far into the future weather can be predicted, perhaps on the order of two weeks. Beyond that point, the chaotic nature of weather means that two simulations starting from the exact same conditions may diverge so far from each other that they become useless, rather than the basis for predicting a range of temperatures and precipitation chances.

Nevertheless, he says, the distance we’ve come is remarkable.

“Weather prediction is one of the most fantastic unsung success stories in science,” he says. “It's been a huge advance and very, very few people realize it.”

 

Thank you to Jimmie Zinn of Florence, Oregon, for the question. You can submit your own question to Ask MIT Climate here.

Read more Ask MIT Climate

Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license (CC BY-NC-SA 4.0).
Footnotes

1 Emanuel, Kerry. "Will global warming make hurricane forecasting more difficult?" Bulletin of the American Meteorological Society, Vol. 98, Issue 3 (2017), doi:10.1175/BAMS-D-16-0134.1.