the rise and fall of the best and worst weather app ever.

At the end of this year, Dark Sky—the popular, sleek, sometimes accurate weather app—will go dark, its various components ambiguously incorporated into Apple’s revamped weather forecasting tool. The app was created in 2011 by developers Jack Turner and Adam Grossman, and was funded through a Kickstarter campaign that raised just under $40,000. During its run, it built up a loyal following of people who used it to keep ahead of the weather day to day, hour to hour, and even minute to minute. Now, with the Android version already dead, the beloved iOS app will stop working altogether on Jan. 1.

I was a late smartphone adopter, but even I remember when Dark Sky first arrived. Suddenly, everyone I knew was a meteorologist, pulling out their pocket-sized radar display to chime in about whether a baseball game would be postponed or a wedding ceremony moved inside.
The app’s standout was the radar map that allowed users to watch storms—rendered in electric blue and neon magenta—wash across their localities in real time. Maybe we didn’t understand the way that pressure, temperature, and humidity interacted to create weather systems, but with the radar, everyone could play weatherman while we waited under a bus shelter for a system to pass through.

Was it perfect? God, no. The other thing that stood out about Dark Sky was that it wasn’t very good. Dark Sky’s forecasts failed me on hikes, bike trips, park days, beach days, birthdays, and more. You name it, and I’ve had it ruined by the siren song of that neon map. It led me to over-plan my outdoor activities and under-pack for them. Just as often, I stayed home and called off plans because of a storm that wasn’t. Once, ahead of a trip in a boat, its forecast failed to predict an impending microburst—an extreme, hyperlocalized weather system—and I got stuck on the water in 40 knots of wind with two people’s lives in my hands. This app could have killed me.

Still, the allure was real. I kept the app on my phone over the years and continued to consult its forecasts. I did my best to lean on its strengths and work around its weaknesses. On humid summer days, when the sky was full of threatening clouds, it could feel like the software was basically guessing which ones would break. At some point, for forecasts a few days on, it felt so unreliable that I stopped checking it altogether. I diversified my weather news, consulting a nerdly assortment of online tools like Weather Underground and Windfinder for situations where safety considerations called for an accurate forecast. But Dark Sky was there when a storm showed up on the horizon: It was the easiest way I knew of to find out what its shape was and infer how long a coming downpour would last. The real-time radar feature let me know in seconds whether a front was spread out across multiple state lines or a thin band of rain that would pass over quickly.

Meteorologists, it’s worth noting, do not and have never shared my enthusiasm.

“It was processing the images,” said Andrew Blum, author of The Weather Machine: A Journey Inside the Forecast, “rather than forecasting the weather using physics.” The Weather Machine explores, in deep detail, the science and history of the complex systems, processes, and technological innovations that undergird global weather forecasting. Blum’s book excavates the forecast, layer by layer, to put together a picture of how the various components come together to do something seemingly impossible.

Dark Sky, widely disparaged by weather scientists, didn’t make the cut.

“Any weather forecast beyond a couple of hours, any computer forecast beyond a couple of hours,” Blum explained, “is going to depend on the weather models—supercomputer models that work according to the laws of physics. When we talk about anything past a few hours, we’re talking about physics. But when we talk about Dark Sky, all it was doing was taking the visual input of the radar and extrapolating what was going to happen over the next couple of hours.”

Indeed, Dark Sky’s big innovation wasn’t simply that its map was gorgeous and user-friendly: The radar map was the forecast. Instead of pulling information about air pressure and humidity and temperature and calculating all of the messy variables that contribute to the weather—a multi-hundred-billion-dollars-a-year international enterprise of satellites, weather stations, balloons, buoys, and an army of scientists working in tandem around the world (see Blum’s book)—Dark Sky simply monitored changes to the shape, size, speed, and direction of shapes on a radar map and fast-forwarded those images. “It wasn’t meteorology,” Blum said. “It was just graphics practice.”

“If there’s snow, or rain, or some complex storms going on,” said Jack Drake, a meteorologist based in Connecticut, “throw the apps in the garbage.” Drake is not some stodgy old-guard type annoyed at the trends in how people get their forecasts: After graduating from college in 2017, he started his meteorology career in the app era, outside the traditional TV path, by sharing his forecasts for western Connecticut on social media and making local media appearances during more serious storms. Those storms (alas, for an app like Dark Sky) can change their size or shape dramatically in a short period of time. Clouds that look benign at first can become dangerous. The atmosphere is full of invisible variations in temperature, pressure, and moisture that drastically affect the course of weather events. Most of these things are invisible to a simple radar.

“I don’t do a traditional five-day forecast,” Drake said. “I try to cater what I do toward what people find the most useful, and it’s not those 75-degree sunny days. I can make a difference with the complex systems.” Dark Sky fell short of that.

Other meteorologists see the death of the app as an opportunity. Jonathan Porter, AccuWeather’s chief meteorologist, pointed to the group’s own app for soon-to-be weather orphans. Unlike Dark Sky, whose developers more or less boasted when they first launched the app that they weren’t weather scientists (they had previously been working on a project called Tiny Face, an app designed to scientifically measure how small someone could make their face), AccuWeather’s team has more than 100 meteorologists who interpret and explain the output of the weather forecasting models. “The way we do it here,” said Porter, “is by taking all of our world-class weather technology and blending that with the expertise of our meteorologists.”

Why would Apple ever buy Dark Sky, anyway? Well, the app really was attractive—and had a lot of users that Apple is now trying to seduce into Apple Weather by pointing out that some of Dark Sky’s key features have been integrated into its weather operation. In addition to revamping its tool, Apple also recently launched its WeatherKit API with the newest version of iOS, a service that charges other weather apps to power their platforms using Apple’s data. Dark Sky’s best parts—the pretty radar map, and its “next hour precipitation” forecast—are being used to beef up the service as Apple attempts to compete as a more primary provider of forecasting data. Apple, in other words, is looking to become a provider of weather data rather than a customer of those services.

But Dark Sky will always be, to some loyal users, more than these scrap parts. Many will remember it as a cultural event. The app (and, I suppose, its peers) heralded a sea change in how the people around me consumed weather forecasts. It’s easy to forget what things were like in 2010, but most of us used to get our weather once daily, if that, from a professional meteorologist on television, or on buggy websites. The apps that came with early smartphones offered little more than basic percentage chances of rain or snow at a given point during a given day. Precise, minute-by-minute forecasts were the tools of mountain climbers and maritime workers—people whose lives and livelihoods depended on knowing exactly what the sky would do and when. People like me got along fine without so much detail, until we didn’t.

My most satisfying experiences with Dark Sky were probably the occasions when I had no intention of going outside at all. If weather was passing through, especially at night, the real-time radar screen was my favorite place to watch the bands of rain or snow crawl through my neighborhood from the comfort of my home. If the hypnotic blue and purple shapes weren’t the basis for a perfect long-term forecast, they were soothing to take in live. It was cozy to watch the phosphorescent radar images dance across the map while rain drummed at the window. On Dark Sky, the rough edges of a weather forecast felt sanded down to something palatable and easy to digest. Its radar allowed you to zoom in on the block—the house, even—where a rain cloud ended. The storms would move in, change their shape a little, and then pass me by. Avoid the magenta part, and I was safe. Stay out of the way of the blue part, and I was dry. It was a tidy illusion, too good to be true—but an illusion that I, among many others, was happy to indulge.