A machine learning-based weather prediction program developed by DeepMind researchers called “GraphCast” can predict weather variables over the span of 10 days, in under one minute. In a report, scientists highlight that GraphCast has outperformed traditional weather pattern prediction technologies at a 90% verification rate.

The AI-powered weather prediction program works by taking in “the two most recent states of Earth’s weather,” which includes the variables from the time of the test and six hours prior. Using that data, GraphCast can predict what the state of the weather will be in six hours. 

In practice, AI has already showcased its applicability in the real world. The tool predicted the landfall of Hurricane Lee in Long Island 10 days before it happened, while the traditional weather prediction technologies being used by meteorologists at the time lagged behind. Forecasts made by standard weather simulations can take longer because traditionally, models have to account for complicated physics and fluid dynamics to make accurate predictions.

Not only does the weather prediction algorithm outperform traditional technologies to forecast weather patterns in terms of pace and scale, GraphCast can also predict severe weather events, which includes tropical cyclones and waves of extreme temperatures over regions. And because the algorithm can be re-trained with recent data, scientists believe that the tool will only get better at predicting oscillations in weather patterns that coincide with grander changes that align with climate change.

Soon, GraphCast, or at least the basis of the AI algorithm that powers its predictions, might pop up into more mainstream services. According to Wired, Google might be exploring how to integrate GraphCast into its products. The call for better storm modeling has already paved a path for supercomputers in the space. The NOAA (National Oceanic and Atmospheric Administration) says it has been working to develop models that will provide more accurate readings on when severe weather events might occur and importantly, the intensity forecasts for hurricanes.

On Monday, meteorologists documented the hottest day in recorded history, according to US National Centers for Environmental Prediction (via Reuters). July 3rd, 2023 saw average global temperatures edge past 17-degrees Celsius (62.62 Fahrenheit) for the first time since satellite monitoring of global temperatures began in 1979. Scientists believe Monday is also the hottest day on record since humans began using instruments to measure daily temperatures in the late 19th century. The previous record was set in August 2016 when the world’s average temperature climbed to 16.92C (62.45 Fahrenheit).

This week, the southern US is sweltering under a heat dome that has sent local temperatures past the 110 Fahrenheit mark (43C). Even places that normally aren’t known for their warm weather have been unseasonably hot in recent days and weeks, with the Vernadsky Research Base in Antarctica recording a July high of 8.7C.

Scientists attribute the recent heat to a combination of El Niño and ongoing human-driven emissions of greenhouse gases. Studies have shown that climate change is contributing to heat waves that are more frequent, last longer and hotter than ever. "The average global surface air temperature reaching 17C for the first time since we have reliable records available is a significant symbolic milestone in our warming world," climate researcher Leon Simons told BBC News. "Now that the warmer phase of El Niño is starting we can expect a lot more daily, monthly and annual records breaking in the next 1.5 years.”

Amazon told lawmakers it wouldn’t build storm shelters in its warehouses after a December 2021 tornado killed six employees at an Illinois location. Although the company changed its severe-weather response strategy after the incident, it essentially told the elected officials that since building storm shelters isn’t required by law, it won’t do that.

The company responded to lawmakers Senator Elizabeth Warren (D-MA) and Representatives Alexandria Ocasio-Cortez (D-NY) and Cori Bush (D-MO), who sent a letter on December 15th, questioning the company’s lack of storm shelters or safe rooms at its warehouses. “Amazon’s apparent unwillingness to invest in a storm shelter or safe room at its Edwardsville facility is made even more concerning by the fact that installing one could be done by Amazon at relatively low cost,” the lawmakers wrote. “This cost is negligible for a company like Amazon, which brought in more than $500 billion in revenue over the 12-month period ending September 30, 2022 and clearly has the resources necessary to protect its workers should it have the will to do so.”

Company vice president of public policy Brian Huseman responded (via CNBC), “Amazon requires that its buildings follow all applicable laws and building codes. We have not identified any jurisdiction in the United States that requires storm shelters or safe rooms for these types of facilities.”

Lawrence Bryant / reuters

Huseman added that Amazon follows Occupational Safety and Health Administration (OSHA) and National Weather Service guidelines and will continue using a “severe weather assembly area” for sheltering in place instead of the requested storm shelters. The six employees and contractors who died at the warehouse tried to protect themselves in a bathroom; the surviving workers took refuge in an assembly area.

OSHA investigated the incident last April and ordered Amazon to review its severe weather policies, but it fell short of penalizing the company for its response. Additionally, Amazon hired a meteorologist, launched an internal center for monitoring severe weather and created emergency cards pointing out evacuation points and assembly areas.

Amazon reportedly began rebuilding the warehouse last June. The families of two of the employees killed there have sued the company for wrongful death.

Self-driving cars frequently have trouble with poor weather, but Waymo thinks it can overcome these limitations by using its autonomous taxis as weather gauges. The company has revealed that its latest car sensor arrays are creating real-time weather maps to improve ride hailing services in Phoenix and San Francisco. The vehicles measure the raindrops on windows to detect the intensity of conditions like fog or rain.

The technology gives Waymo a much finer-grained view of conditions than it gets from airport weather stations, radar and satellites. It can track the coastal fog as it rolls inland, or drizzle that radar would normally miss. While that's not as important in a dry locale like Phoenix, it can be vital in San Francisco and other cities where the weather can vary wildly between neighborhoods.

There are a number of practical advantages to gathering this data, as you might guess. Waymo is using the info to improve its Driver AI's ability to handle rough weather, including more realistic simulations. The company also believes it can better understand the limits of its cars and set higher requirements for new self-driving systems. The tech also helps Waymo One better serve ride hailing passengers at a given time and place, and gives Waymo Via trucking customers more accurate delivery updates.

The current weather maps have their limitations. They may help in a warm city like San Francisco, where condensation and puddles are usually the greatest problems, but they won't be as useful for navigating snowy climates where merely seeing the lanes can be a challenge. There's also the question of whether or not it's ideal to have cars measure the very conditions that hamper their driving. This isn't necessarily the safest approach.

This could still go a long way toward making Waymo's driverless service more practical, though. Right now, companies like Waymo and Cruise aren't allowed to operate in heavy rain or fog using their California permits — the weather monitoring could help these robotaxi firms serve customers looking for dry rides home.

Almost exactly five years after Hurricane Maria left Puerto Rico in the dark, the US territory is once again facing a power crisis. On Sunday, LUMA Energy, the company that operates the island’s electrical grid, announced that all of Puerto Rico had suffered a blackout due to Hurricane Fiona, reports Reuters.

With the storm nearing the island’s southwest coast, the National Hurricane Center warned of “catastrophic” flooding as Fiona began producing winds with recorded speeds of 85 miles per hour. Without even making landfall, the storm left a third of LUMA’s customers without power. On Twitter, Puerto Rico Governor Pedro Pierluisi said the government was working to restore power, but after the events of five years ago, there’s worry there won’t be an easy fix.

In 2017, Hurricane Maria caused the largest blackout in US history when the Category 5 storm battered Puerto Rico, leaving 3.4 million people without power. The island had only recently begun rebuilding its weakened infrastructure, with blackouts a daily occurrence in some areas. Officials have tried to stress that Hurricane Fiona won’t bring a repeat of 2017. “This is not Maria, this hurricane will not be Maria,” Abner Gomez, the head of public safety and crisis management at LUMA Energy, told CNN before Sunday’s power outage.

Hurricane season is currently in full swing across the Gulf Coast and Eastern Seaboard. Following a disconcertingly quiet start in June, meteorologists still expect a busier-than-usual stretch before the windy weather (hopefully) winds down at the end of November. Meteorologists like Matthew Cappucci who, in his new book, Looking Up: The True Adventures of a Storm-Chasing Weather Nerd, recounts his career as a storm chaser — from childhood obsession to adulthood obsession as a means of gainful employment. In the excerpt below, Cappucci explains the inner workings of tropical storms.

Simon and Schuster

Excerpted from Looking Up: The True Adventures of a Storm-Chasing Weather Nerd by Matthew Cappucci. Published by Pegasus Books. Copyright © 2022 by Matthew Cappucci. All rights reserved.

Hurricanes are heat engines. They derive their fury from warm ocean waters in the tropics, where sea surface temperatures routinely hover in the mid- to upper-eighties between July and October. Hurricanes and tropical storms fall under the umbrella of tropical cyclones. They can be catastrophic, but they have a purpose—some scholars estimate they’re responsible for as much as 10 percent of the Earth’s annual equator-to-pole heat transport.

Hurricanes are different from mid-latitude systems. So-called extratropical, or nontropical, storms depend upon variations in air temperature and density to form, and feed off of changing winds. Hurricanes require a calm environment with gentle upper-level winds and a nearly uniform temperature field. Ironic as it may sound, the planet’s worst windstorms are born out of an abundance of tranquility.

The first ingredient is a tropical wave, or clump of thunderstorms. Early in hurricane season, tropical waves can spin up on the tail end of cold fronts surging off the East Coast. During the heart of hurricane season in August and September, they commonly materialize off the coast of Africa in the Atlantic’s Main Development Region. By October and November, sneaky homegrown threats can surreptitiously gel in the Gulf of Mexico or Caribbean.

Every individual thunderstorm cell within a tropical wave has an updraft and a downdraft. The downward rush of cool air collapsing out of one cell can suffocate a neighboring cell, spelling its demise. In order for thunderstorms to coexist in close proximity, they must organize. The most efficient way of doing so is through orienting themselves around a common center, with individual cells’ updrafts and downdrafts working in tandem.

When a center forms, a broken band of thunderstorms begins to materialize around it. Warm, moist air rises within those storms, most rapidly as one approaches the broader system’s low-level center. That causes atmospheric pressure to drop, since air is being evacuated and mass removed. From there, the system begins to breathe.

Air moves from high pressure to low pressure. That vacuums air inward toward the center. Because of the Coriolis force, a product of the Earth’s spin, parcels of air take a curved path into the fledgling cyclone’s center. That’s what causes the system to rotate.

Hurricanes spin counterclockwise in the Northern Hemisphere, and clockwise south of the equator. Though the hottest ocean waters in the world are found on the equator, a hurricane could never form there. That’s because the Coriolis force is zero on the equator; there’d be nothing to get a storm to twist.

As pockets of air from outside the nascent tropical cyclone spiral into the vortex, they expand as barometric pressure decreases. That releases heat into the atmosphere, causing clouds and rain. Ordinarily that would result in a drop in temperature of an air parcel, but because it’s in contact with toasty ocean waters, it maintains a constant temperature; it’s heated at the same rate that it’s losing temperature to its surroundings. As long as a storm is over the open water and sea surface temperatures are sufficiently mild, it can continue to extract oceanic heat content.

Rainfall rates within tropical cyclones can exceed four inches per hour thanks to high precipitation efficiency. Because the entire atmospheric column is saturated, there’s little evaporation to eat away at a raindrop on the way down. As a result, inland freshwater flooding is the number one source of fatalities from tropical cyclones.

The strongest winds are found toward the middle of a tropical storm or hurricane in the eyewall. The greatest pressure gradient, or change of air pressure with distance, is located there. The sharper the gradient, the stronger the winds. That’s because air is rushing down the gradient. Think about skiing — you’ll ski faster if there’s a steeper slope.

When maximum sustained winds surpass 39 mph, the system is designated a tropical storm. Only once winds cross 74 mph is it designated a hurricane. Major hurricanes have winds of 111 mph or greater and correspond to Category 3 strength. A Category 5 contains extreme winds topping 157 mph.

Since the winds are derived from air rushing in to fill a void, or deficit of air, the fiercest hurricanes are usually those with the lowest air pressures. The most punishing hurricanes and typhoons may have a minimum central barometric pressure about 90 percent of ambient air pressure outside the storm. That means 10 percent of the atmosphere’s mass is missing.

Picture stirring your cup of coffee with a teaspoon. You know that dip in the middle of the whirlpool? The deeper the dip, or fluid deficit, the faster the fluid must be spinning. Hurricanes are the same. But what prevents that dip from filling in? Hurricane eyewalls are in cyclostrophic balance.

That means a perfect stasis of forces makes it virtually impossible to “fill in” a storm in steady state. Because of their narrow radius of curvature, parcels of air swirling around the eye experience an incredible outward-directed centrifugal force that exactly equals the inward tug of the pressure gradient force. That leaves them to trace continuous circles.

If you’ve ever experienced a change in altitude, such as flying on an airplane, or even traveling to the top of a skyscraper, you probably noticed your ears popping. That’s because they were adjusting to the drop in air pressure with height. Now imagine all the air below that height vanished. That’s the equivalent air pressure in the eye a major hurricane. The disparity in air pressure is why a hurricane is, in the words of Buddy the Elf, “sucky. Very sucky.”

Sometimes hurricanes undergo eyewall replacement cycles, which entail an eyewall shriveling and crumbling into the eye while a new eyewall forms around it and contracts, taking the place of its predecessor. This usually results in a dual wind maximum near the storm’s center as well as a brief plateau in intensification.

In addition to the scouring winds found inside the eyewall, tornadoes, tornado-scale vortices, mini swirls, and other poorly understood small-scale wind phenomena can whip around the eye and result in strips of extreme damage. A mini swirl may be only a couple yards wide, but a 70 mph whirlwind moving in a background wind of 100 mph can result in a narrow path of 170 mph demolition. Their existence was first hypothesized following the passage of Category 5 Hurricane Andrew through south Florida in 1992, and modern-day efforts to study hurricane eyewalls using mobile Doppler radar units have shed light on their existence. Within a hurricane’s eye, air sinks and warms, drying out and creating a dearth of cloud cover. It’s not uncommon to see clearing skies or even sunshine. The air is hot and still, an oasis of peace enveloped in a hoop of hell.

There’s such a discontinuity between the raucous winds of the eyewall and deathly stillness of the eye that the atmosphere struggles to transition. The eyes of hurricanes are often filled with mesovortices, or smaller eddies a few miles across, that help flux and dissipate angular momentum into the eye. Sometimes four or five mesovortices can cram into the eye, contorting the eyewall into a clover-like shape. That makes for a period of extraordinary whiplash on the inner edge of the eyewall as alternating clefts of calamitous wind and calm punctuate the eye’s arrival.

Last year, hurricanes hammered the Southern and Eastern US coasts at the cost of more than 160 lives and $70 billion in damages. Thanks to climate change, it's only going to get worse. In order to quickly and accurately predict these increasingly severe weather patterns, the National Oceanic and Atmospheric Administration (NOAA) announced Tuesday that it has effectively tripled its supercomputing (and therefore weather modelling) capacity with the addition of two high-performance computing (HPC) systems built by General Dynamics.

“This is a big day for NOAA and the state of weather forecasting,” Ken Graham, director of NOAA’s National Weather Service, said in a press statement. “Researchers are developing new ensemble-based forecast models at record speed, and now we have the computing power needed to implement many of these substantial advancements to improve weather and climate prediction.”

General Dynamics was awarded the $505 million contract back in 2020 and delivered the two computers, dubbed Dogwood and Cactus, to their respective locations in Manassas, Virginia, and Phoenix, Arizona. They'll replace a pair of older Cray and IBM systems in Reston, Virginia, and Orlando, Florida.

Each HPC operates at 12.1 petaflops or, "a quadrillion calculations per second with 26 petabytes of storage," Dave Michaud, Director, National Weather Service Office of Central Processing, said during a press call Tuesday morning. That's "three times the computing capacity and double the storage capacity compared to our previous systems... These systems are amongst the fastest in the world today, currently ranked at number 49 and 50." Combined with its other supercomputers in West Virginia, Tennessee, Mississippi and Colorado, the NOAA wields a full 42 petaflops of capacity. 

With this extra computational horsepower, the NOAA will be able to create higher-resolution models with more realistic physics — and generate more of them with a higher degree of model certainty, Brian Gross, Director, NOAA’s Environmental Modeling Center, explained during the call. This should result in more accurate forecasts and longer lead times for storm warnings.

"The new supercomputers will also allow significant upgrades to specific modeling systems in the coming years," Gross said. "This includes a new hurricane forecast model named the Hurricane Analysis and Forecast System, which is slated to be in operation at the start of the 2023 hurricane season," and will replace the existing H4 hurricane weather research and forecasting model.

While the NOAA hasn't yet confirmed in absolute terms how much of an improvement the new supercomputers will grant to the agency's weather modelling efforts, Ken Graham, the Director of National Weather Service, is convinced of their value. 

"To translate what these new supercomputers will mean for for the average American," he said during the press call, "we are currently developing models that will be able to provide additional lead time in the outbreak of severe weather events and more accurately track the intensity forecasts for hurricanes, both in the ocean and that are expected to hit landfall, and we want to have longer lead times [before they do]."

When the United Arab Emirates launched the Arab world’s first-ever mission to Mars in the summer of 2020, its desire was that its Hope probe would help provide scientists with a better understanding of the Red Planet’s weather systems. And it’s now done exactly that. According to The National, the probe recently spent two weeks tracking a massive dust storm across the surface of Mars.

Hope began following the weather event on December 29th. The probe entered the orbit of Mars equipped with a high-resolution camera and an infrared spectrometer. It used those tools to track the geographic distribution of dust, water vapor and carbon dioxide ice clouds displaced by the raging storm. Its orbital position allowed Hope to observe any variance in those elements in timescales measured in minutes and days, a feat previous missions to Mars didn’t have the ability to do. 

What it saw was how quickly a storm can spread across the red planet. In the span of a single week, the storm it was tracking grew to stretch across more than 1,550 miles of Martian surface. In the process, it completely obscured geographic landmarks like the Hellas impact crater and sent dust haze as far as 2,485 miles away from the origin point of the storm. In addition to providing a play-by-play of a Martian storm, scientists hope the data Hope collected will allow them to gain a better understanding of how those storms can help water escape the planet's atmosphere.

Tornadoes ripped through six states on Friday, killing dozens. Among the dead were six workers at an Amazon warehouse in Edwardsville, IL, which collapsed while they sheltered inside. The incident is now the subject of an OSHA investigation.

The mass casualty event is likely Kentucky's "deadliest tornado system in state history," according to ABC. The twisters also touched down in December, well outside the normal tornado season. While this may have been an unusually extreme weather event for many reasons, Amazon's decision to schedule its workers during potentially deadly conditions isn't. Reportedly, at the time the cyclone touched down in warehouse's parking lot — producing winds estimated at 155 miles per hour — the facility was not only operating, but undergoing a shift change. 

Amazon operates a staggering number of fulfillment, sortation and delivery centers across the country, and as a result, some of them are bound to be taken by surprise by the forces of nature. Excessive snow on the roof of one warehouse in Pennsylvania resulted in an evacuation when workers noticed the it buckling. Two contractors were killed by a collapsing wall when a tornado touched down without warning in Baltimore. 

But the National Weather Service had been warning of possible tornadoes 36 hours ahead of the deaths in Edwardsville; the morning before the storms it cautioned of the "likely threat" of "damaging winds in excess of 60 mph." Edwardsville is in what FEMA categorizes as Wind Zone IV, the part of the country at the greatest risk of tornadoes. 

Amazon is perhaps better known in media coverage for its punishing productivitygoals. But its operating standards have produced a pattern of incidents where workers were expected to clock in during extreme weather events. Warehouses stayed open during tropical depression Ida in September, the torrential rains of which caused widespread flooding and led to 14 deaths in New York. Some of Amazon's drivers told me they were delivering packages through the floodwaters of hurricane Irma back in 2017.

The Camp Fire of 2018 was the deadliest and costliest wildfire in California's history. Smoke from the destruction also briefly made Sacramento the most polluted city on earth. Despite air quality warnings being issued for the city on November 8th, an Amazon warehouse there did not send its workers home until the afternoon of the 10th. 

By far, however, the most pervasive issue across Amazon's warehouses has been extreme heat. Workers in the Pacific Northwest were expected to report for duty during a historic heatwave this past summer which was eventually deemed a mass casualty event. Specifically, a worker complained that some areas of a warehouse in Kent lacked fans, and estimated temperature inside hit 90 degrees. New York warehouse workers also reported fainting and excessive heat around the same time. In May of this year, excessive heat led to a death inside the company's Bessemer, Alabama warehouse. 

These are only some of the most recent examples. Workers have been lodging similar complaints for at least a decade about dangerous temperatures inside Amazon's facilities in Chicago, Portland and Pennsylvania's Lehigh Valley, among others. Even when immediate symptoms like fainting, vomiting or heat stroke are not present, long term heat exposure can exacerbate existing health problems such as heart conditions and asthma. 

None of this speaks to criticisms of Amazon's safety measures related to COVID-19, or its objectively sky-high injury rate compared to other warehousing operations.

What's concerning is that, according to the overwhelming majority of the scientific community, severe winds, rain and heat are likely to get worse due to man-made climate change. Amazon, however, has not offered a satisfactory explanation for why it continues to schedule shifts during potentially deadly weather, nor would it provide Engadget with any details of the extreme weather plan in effect at the Edwardsville facility. 

“We’re deeply saddened by the news that members of our Amazon family passed away as a result of the storm in Edwardsville, IL," an Amazon spokesperson told Engadget. "Our thoughts and prayers are with the victims, their loved ones, and everyone impacted by the tornado. We also want to thank all the first responders for their ongoing efforts on scene. We’re continuing to provide support to our employees and partners in the area.”

Once upon a time, we would run home from the bus stop to watch Gargoyles and Brady Bunch reruns on the family TV, a late-1970s console Magnavox number that sat on the floor and was about 50% more cabinet than CRT. The old TV, a streamlined white Zenith at least ten years older, had been relegated to the man cave in the basement. It looked so mod compared to the “new” TV, but that’s not the aesthetic my folks were after. They wanted their electronics to double as furniture.

This little TV is a happy medium between the two styles, and for us, it’s all about those feet. But instead of cartoons, it switches between showing the current weather and the top news headlines. Inside that classy oak cabinet is an LCD, an ESP32, and an SD card module. The TV uses OpenWeatherMap and pulls the corresponding weather image from the SD card based on time of day — light images for day, and dark images for night.

We love that it shows the SMPTE color bars, aka the standard American TV test pattern as it switches between weather and news. After showing the top headlines, it automatically switches back to the weather channel. Be sure to check out the short demo video after the break.

Do you like your tiny televisions in strange places? Here’s one you can use to trim your tree this year.