Autumn is many people’s favorite season in New England, but as far as changing climate is concerned it is proving to be the most unusual.
That’s one of the surprises from newly updated 30-year “normals” released by the National Weather Service.
The update, which happens every 10 years, recalculates what are considered normal temperatures, rain and snowfall amounts at each official weather station by looking at the most recent three decades.
The updated listing for the weather center at Concord Municipal Airport, the state’s official data site, shows that it has gotten warmer overall. The average temperature over the course of the year from 1991 through 2020 was 47.2 degrees Fahrenheit, which is 0.8 degrees warmer than it had been between 1981 through 2010.
No surprise there amid global warming. But the monthly breakdown for September, October and November did show a surprise.
The new 30-year normals for those months in Concord are much warmer than for other seasons during the daytime yet, unlike any other season, are cooler at night.
“We were a little bit surprised by it,” said Donald Dumont, warning coordination meteorologist at the National Weather Service in Gray, Maine, which covers Concord.
Weather services worldwide have long used a 30-span to determine what is “normal” for temperatures and precipitation in a given area. There’s nothing magic in this number of years but it has become an industry standard.
The weather station at Concord airport is the only National Weather Service site in New Hampshire with an official 30-year normal. A gap in data collection at Manchester’s airport keeps it from having a similar comparison.
This month, U.S. meteorologists switched from using 1981 through 2010 as their comparison period, and began using 1991 through 2020.
Because the change keeps two decades – the 1990s and 2000s are part of both the old normal and the new normal – any changes are actually the result of comparing the 1980s, which were dropped, with the 2010s, which were added.
“What drives the normals just as much is that we lost the ‘80s – it’s not just that the last 10 years is driving all these changes,” Dumont pointed out.
For example, the 1980s were 1.3 degrees colder than the current normal and the 2010s were 0.4 degrees warmer. So the 0.8-degree rise in average was caused mostly by losing the earlier decade.
“If you look at how the normals changed, the fact that we lost the ‘80s is a much bigger factor in these new climate normals,” Dumont said.
Overall, the new normals for Concord are warmer by 0.8 degrees and were wetter by 1.34 inches per year, roughly a 3% increase. Neither is a surprise; warmer air can hold more water and thus can produce more rainfall.
But the monthly analysis shows an autumn surprise. September, October and November saw their average daily high temperatures increase more than any other season – more than 3 degrees – and yet their average daily low temperatures actually fell by 1.5 degrees or more, unlike any other season.
In other words, while other seasons got warmer during the day and also during the night, autumn got much warmer during the day yet colder at night.
There’s no good explanation for this, Dumont said. One hypothesis is that for some reason Concord’s air is getting drier in the fall.
“With drier air, the diurnal temperature spread gets bigger. Think about the desert climate: it can warm up faster in the day and cool down faster in the night,” he said.
But that’s just a hypothesis, or an educated guess on which further research can be built. It’s not obvious why Concord’s air might be drying out in the autumn.
There’s another surprise in the new data: The average annual snowfall shot up a whopping 6.7 inches, to 67.7 inches, even though winters are getting shorter and seem to be less snowy.
This partly reflects the fact that the 1980s had some snow-poor years which are no longer in the averages: The decade as a whole had 8.3 inches less total snowfall than the new normal.
But this figure also shows the limits of using a single number to describe something as complicated as seasonal snowfall.
Total snowfall over the year doesn’t reflect how quickly that snow melts. Last winter, for example, saw two examples of heavy snowfalls that were soon followed by heavy warm rains which removed much or all over the snow cover.
So even though the winter snowfall tally was relatively high, winter didn’t seem all that snowy, as skiers can attest.
“We get more snow but we’re not keeping it,” said Dumont.
Finally, there’s one other thing that the upgrade shows: “We don’t have enough high-order climate sites in the state.”
Considering weather normals at Concord airport to reflect what’s normal on the Seacoast or the Connecticut River Valley or North of the Notches is, of course, ridiculous. But for the moment, it’s the best we’ve got.