When climatologists started standardizing global weather data about 100 years ago, they didn’t know that heat-trapping greenhouse gases were already pushing the planet’s climate inexorably in one direction, off the charts of human experience. 

But people like to measure things in understandable segments, so, based on the data it had at the time, the World Meteorological Organization created three-decade climate reference periods they called “climate normals” against which they could measure daily temperatures, unusual heat waves, cold snaps or big rainstorms.

Scientists still assumed the climate was generally stable. But they also knew about seasonal, annual or multiyear temperature and precipitation swings caused by cyclical changes in ocean currents and winds, like El Niño and La Niña, which sloshed warmer or cooler water about every two to seven years. Averages from a 30-year period could smooth out those variations in the record.

But the climate normals can’t smooth out the one-way trend line of global warming, which has heated the planet by nearly 2 degrees Fahrenheit in the last century, so some experts now question the usefulness of the 30-year climate normal as a reference period in the era of rapid climate change.

“I’d vote to scrub the word ‘normal’. Maybe something like ‘recent average,’” Rick Thoman, a climate specialist at the University of Alaska, Fairbanks, said via Twitter. Other climate scientists trying to communicate global warming risks to the public also say the terminology could be confusing.

In the U.S., the data for the climate normals come from about 8,700 weather stations. The recent update for the first time includes precipitation readings from 770 automated snow measurement sites and a network of 5,400 citizen science observation stations. The new normals are much wetter across a big swath of the eastern United States, especially from the Upper Midwest down to the Gulf Coast, and the Great Lakes region east to New England.

Overall, when the new climate normals, from 1991 to 2020, are compared with the previous 30-year period (1981 to 2010), much of the U.S. has warmed. In parts of the Northern Plains and Upper Midwest, the new climate normals are cooler. 

Longer-term, the Midwest is warming and projected to warm by at least 5 degrees Fahrenheit by 2100, according to a report from the Union of Concerned Scientists authored by Texas Tech climate scientist Katharine Hayhoe.

Shifting ‘Climate Normals’ Felt All Over the Country 

The warmer and drier conditions shown by the new climate normals are already affecting communities and ecosystems across the country, including commercial shellfisheries near Galveston, Texas.

The effects are beyond obvious, said Lisa Halili, CEO of Prestige Oyster. 

Oysters are dependent on very specific levels of salinity in the water, about 14 parts per million, so any weather that forces it to become too salty, or not salty enough, is a problem, Halili said. Warmer weather means more evaporation, concentrating salt, and lack of rain means no freshwater replenishment.

 “We’re dealing with the effects of climate change every day,” she said. “It’s part of being in the oyster business.”

In fact, if the trend of warming in the region continues, soon more than just the oyster business might have to change the way they operate. Cities could find themselves planning for more droughts, and workers will have to keep cool outside in the heat.

The updated NOAA data show that the region’s daytime mean temperature through the day increased from 71.2 degrees up to 72.3 degrees Fahrenheit over the past 30 years. Annual average precipitation has declined 3.54 inches, from 50.76 inches a year down to 47.22 inches.

Climate Snapshots Don’t Tell the Full Global Warming Story

The climate normals are useful for making forecasts that are beyond the range of daily and weekly weather predictions, but not in the scope of global climate models. For example, the long-term historical average can give you a pretty good idea of what to pack if you’re planning a trip to San Francisco in mid-October. And the normals are also used to forecast river flows and to regulate power production, to time crop planting, and for construction planning and building design.

The 30-year climate normals provide a decent snapshot of recent historical conditions, so at first, the updates were made simply to account for new weather stations added to the system, or for stations that were removed, said Mike Palecki, who led the National Oceanic and Atmospheric Administration team that compiled the data. 

“But the climate normals are not the best way to measure global warming,” he said. “It kind of mutes the signal a little bit.” 

A day that might have been exceptionally warm for a given season back in 1950 now “looks normal because we’ve moved the baseline,” he said. But the 30-year reference periods are still useful because they let us compare today’s weather with “normal weather” on a relevant timescale.

For example, the normals can help a construction company plan a project considering how many days of rain or hot temperatures to expect, on average, during July and August at a given location, he said. 

But with rapid warming, it might be worth rethinking the concept of normals, at least for public facing climate communications, said Özden Terli, a broadcast meteorologist with ZDF, one of the main German television networks and a member of Climate Without Borders, a global network of weather presenters focusing on communicating scientific climate change information.

The previous climate normal reference period (1981 to 2010) was “relatively far in the past, but now we’re fully in the fast warming phase,” he said. “When you say the temperature is 0.2 degrees Celsius above normal and you’re using that new climate period as the reference, it kind of hides some of the previous warming.”

To fully communicate how much greenhouse gas pollution has warmed the planet, he said it would be best to use a climate reference period from the pre-industrial era, before about 1850. But many countries don’t have extensive and accurate records going back that far, so it’s not a useful reference for some purposes. 

Some experts have advocated for using the 1961 to 1990 period because it was a relatively stable period during the modern climate era, he said. But global warming has accelerated so much since the 1980s that the averaged data from the decades since then don’t really fit into any historical climate context, he added.

‘Shifting Baselines are Bad’

And that’s the core of the problem, said Tobias Schad, a Berlin-based meteorologist currently working on forest health issues. 

“When you compare today’s temperatures against a reference period that includes so many hot years, some people might think, ‘Oh, it’s not so bad,’ so, through the back door, there is a normalization of warming,” he said. “And it’s also a problem that skeptics, or climate deniers, could use it to say it’s not warming dangerously, which is bad at a time when it’s more important than ever to show people that we are coming out of a stable climate, and that we are in an unstable climate.”

Several climate scientists also questioned the terminology via Twitter. Robert Rohde, at Berkeley Earth, said, “I’m not a fan. That we call them ‘normals’ often makes communication worse, in my opinion. The concept of a 30-year baseline is not inherently terrible, but ‘normalizing’ the recent past as the frame of reference can hinder understanding of the important longer-term changes,” he said.

“I think we almost all understand that there no longer is a normal climate,” said physicist Nick Cowern, Emeritus Professor at the Newcastle University School of Engineering. “It’s better to look at dynamic change, which is what actually affects us, or to compare to the preindustrial climate to understand how far we are from equilibrium and civilisational stability,” he said.

“Shifting baselines are bad,” said Stefan Rahmstorf, a climate scientist at the Potsdam Institute for Climate Impact Research. “They obscure climate change.”