[Geocolor satellite image from NOAA’s GOES-16 of a powerful East Coast storm on Jan. 4, 2018. (NOAA)]
[UC Davis] What kinds of weather do people find remarkable and when does that change? A study led by the University of California, Davis, examined those questions through the lens of more than 2 billion U.S. Twitter posts.
The study, published February 25 in the journal Proceedings of the National Academy of Sciences, indicates that people have short memories when it comes to what they consider “normal” weather. On average, people base their idea of normal weather on what has happened in just the past two to eight years.
“There’s a risk that we’ll quickly normalize conditions we don’t want to normalize,” said lead author Frances C. Moore, an assistant professor in the UC Davis Department of Environmental Science and Policy. “We are experiencing conditions that are historically extreme, but they might not feel particularly unusual if we tend to forget what happened more than about five years ago.”
[Graphical depiction of residual variation in temperature for Cook County IL. Raw temperature values are shown in grey. County fixed-effects remove the mean for each county over the period of twitter data to center the temperatures around zero (green line). State by month-of-year fixed-effects remove the seasonality for the state. This residual variation (purple line), interacted with average temperatures in the reference and recent time periods is used to identify model coefficients.]
Trending on Twitter
To reach their conclusions, the researchers quantified a timeless and universal pastime — talking about the weather — by analyzing posts on Twitter.
They sampled 2.18 billion geolocated tweets created between March 2014 and November 2016 to determine what kind of temperatures generated the most posts about weather. They found that people often tweet when temperatures are unusual for a particular place and time of year — a particularly warm March or unexpectedly freezing winter, for example.
However, if the same weather persisted year after year, it generated less comment on Twitter, indicating that people began to view it as normal in a relatively short amount of time.
The boiling frog
This phenomenon, note the authors, is a classic case of the boiling-frog metaphor: A frog jumps into a pot of boiling hot water and immediately hops out. If, instead, the frog in the pot is slowly warmed to a boiling temperature, it doesn’t hop out and is eventually cooked. While scientifically inaccurate, this metaphor has long been used as a cautionary tale warning against normalizing remarkable weather.
Sentiment analysis tools, which measure the positive or negative association of words, provided evidence for this “boiling-frog effect.” After repeat exposures to historically extreme temperatures, people tweeted less about the weather specifically, but they still expressed negative sentiments overall. Particularly cold or hot conditions still seemed to make people unhappy and grumpy.
“We saw that extreme temperatures still make people miserable, but they stop talking about it,” Moore said. “This is a true boiling-frog effect. People seem to be getting used to changes they’d prefer to avoid. But just because they’re not talking about it doesn’t mean it’s not making them worse off.”
Edited for WeatherNation by Meteorologist Mace Michaels