Could investing in better weather forecasting save lives? Planet Money NPR explores this fascinating link between personal safety and weather prediction.

Richard Bouhet/AFP by Getty Images

Richard Bouhet/AFP/Getty Images

The morning of September 21st, 1938. The New York Times The newspaper published a routine weather forecast which did not raise any alarms among its readers.

“The indications are for rain and cool weather today and for cloudy and continued cool weather, probably with rain, tomorrow, according to the map charted at the United States Weather Bureau at 7:30 o’clock (EST) last night,” the paper wrote.

In the days prior, the U.S. Weather Bureau — the predecessor to the National Weather Service — had been tracking a hurricane that was threatening the coast of Florida. The storm changed course and diverted away from Florida’s coast. The agency, despite warnings from a junior forecaster decided that the system would spin away and die somewhere in the middle Atlantic Ocean. The agency deemed that the cyclone posed no threat.

The New York TimesIn its same edition, on September 21, it praised even the weather agency’s work in tracking the hurricane. “If New York and the rest of the world have been so well informed about the cyclone it is because of an admirably organized meteorological service,” the paper wrote.

The angry vortex of waters was still present off the coasts of Long Island that morning. 120-mph winds It was already roaring towards the land. It made landfall at around 2:30pm. Seismographs recorded the impact of the tidal wave. as far away as Alaska.

The Great Hurricane 1938 or “The Long Island Express” As it was called, this hurricane would go on to be one of most destructive in American history. It caused the destruction of more than 63,000 dwellings. It injured tens of thousands. The number of deaths was more than 600 people. And — because of bad forecasting — many of these victims were taken completely by surprise.

1938 Hurricane damage at Crescent Beach, Connecticut

National Archives

Hide caption

Toggle caption

National Archives

1938 Hurricane damage at Crescent Beach, Connecticut

National Archives

The weather forecasts of the 1930s are a far cry from today. Meteorologists used to forecast the weather back then. “relied on the 16th-century thermometer, the 17th-century mercurial barometer, and the medieval weather vane,” writes the historian William Manchester. The use of newfangled planes was becoming more and more important for making forecasts. However, forecasters still relied heavily on ships at sea to provide them with information about weather patterns like hurricane tracks.

Weather forecasters have a wide range of tools to help them make accurate predictions. Doppler towers can detect wind and precipitation patterns. The radiosondes are attached to weather balloons and float up in the stratosphere. They collect data on temperature, moisture, air pressure, speed and direction of wind, as well as other factors. Automated surface observing systems can provide data in real time about the conditions on the land. Satellites circle our planet, sending back valuable data and imagery. Supercomputers are also used to analyze data. advanced statistical models This data will help forecasters create a picture of our future weather.

With all of this technology, the meteorologists have made incredible progress.

Today’s five-day forecast is as accurate In 1980, a single-day forecast used to be available.

Heavy rain is predicted for the next two days. as good As the same-day weather forecast was in the mid-1990s.

Hurricane predictions are often wrong. half as likely They are still the same as they were only a few years ago.

Forecasters in 1990 could only make a relative accurate forecast of the weather. seven days in advance. Now they are able to make forecasts that are relatively accurate ten days In advance

They may be one of the many things we take for granted in the modern world, but more accurate weather forecasts — and our ability to access them anytime on our smartphones — have tremendous value for our economy. They assist farmers in making decisions about their crops. They can help builders make better decisions. The tourism industry can predict the tourist flow. These tools help people prepare for the future and literally save lives.

It’s been difficult for economists, despite the fact that weather forecasts have a clear value, to quantify their worth. Recently, a group economists made an attempt. A group of economists tried it recently. a new working paper, “Fatal Errors: The Mortality Value of Accurate Weather Forecasts,” The economists Jeffrey G. Shrader (left), Laura Bakkensen (right), and Derek Lemoine (center) focus on the importance of one key aspect in predicting the weather.

What is the value of knowing future temperatures?

Recently, Bakkensen and Lemoine joined me on a Zoom call from Tucson, Arizona, on a day when their city was — quite appropriately for our interview — under an excessive heat warning. Both are economics professors at the University of Arizona.

They estimate that thousands of Americans are killed every year by extreme temperatures. Lemoine admits that he was not confident in the accuracy of temperature forecasts before he conducted this study.

“It’s actually not obvious when forecasts have value,” says Lemoine. “If on some days an error in forecasting means that there are fewer deaths, and other days errors mean that there are more deaths, these things could kind of statistically wash out.” He says that meteorologists have made so many incredible advances in forecasting accuracy in the last few decades, that it’s not clear if any errors still remain.

The economists combined data from the National Weather Service on actual weather and forecasts with fatality data from the Centers for Disease Control and Prevention to see if inaccuracy of temperature forecasts had an impact on deaths. They concentrate on the one-day-ahead temperature forecasts over a 12 year period. “We’re trying to compare the same county, essentially the same people, same temperature day, but this day had an accurate forecast, this day had a slightly inaccurate forecast, and see how that impacts mortality,” Bakkensen says.

The economists have found that forecasting errors can have a big impact on the number of deaths. “We see effects of even errors of just a degree or two,” Lemoine says. “We can see in the data that deaths are higher, and we weren’t expecting it to be that sensitive.”

The economists believe that accurate forecasting is especially important on hot days. Lemoine says that while people can die from the cold as well, it appears that heat is more likely to cause death. So it makes some intuitive sense that a bad forecast in advance of a hot day — in particular a forecast saying it’s going to be colder than it really ends up being — could be particularly deadly.

Bakkensen’s data indicates that people are taking life-saving measures based on forecasts. For instance, they might buy an AC unit, cancel a medical visit, or plan the day so as to avoid direct sunlight. Municipalities could also increase hospital capacity or open public swimming pools.

“Well-forecasted days when they’re hot don’t have that much of an effect on mortality,” Lemoine says. “It’s the inaccurately forecasted hot days that have a big effect. So you can trim a lot of those effects just by having better forecasts.”

The economists calculate this: “making forecasts 50% more accurate would save 2,200 lives per year.” In addition, they estimate. “the public would be willing to pay $112 billion” It will take the rest of the century for this to become a reality. This is only the economic benefit from more accurate weather forecasts resulting in lower deaths. It’s not the overall economic benefit. “I would expect that this number we’re calculating is a big lower bound on the benefits of overall more accurate forecasts,” Bakkensen says.

It is important to note that the word “you” means “you”. annual budget for the National Weather Service Lemoine’s analysis shows that Americans will see considerable benefits if they invest more in this agency in the future.

Both economists agree that it is important to improve weather forecasting, as the climate change is expected to make our country and the entire world hotter and more prone to extreme weather. “As climate change shifts us more toward hot days, it’s implicitly shifting us toward days where accurate forecasts matter more,” Lemoine says. He says that investing in better weather forecasting will be a key part of adapting the climate change.

Bakkensen claims that the National Weather Service forecasts improved by 30% between 2005 and 2017. Therefore, another 50% improvement could be possible. This is especially true since many meteorologists are now using Artificial Intelligence (AI) to improve their forecasts. we may already be seeing the beginnings Another quantum leap in forecast accuracy.