- Lazaro Gamio
- Dec 7
Map of Southern California fires
Kepler 90 is the first known 8-planet system outside of our own. In this system, planets orbit closer to their star, and Kepler 90i orbits once every 14 days. Image: Google
A Google machine learning algorithm found two new planets in previously studied data from NASA's Kepler Space Telescope.
What's next: They hope to study Kepler's data from more than 150,000 stars to see if they can spot weak signals researchers missed.
On September 4, 2017, smoke from Western wildfires entered the gulfstream and spread across the country. Image: NASA Earth Observatory / Suomi NPP
2017 was a smoky year for the United States. In September, the haze was visible from space, appearing as a smear obscuring almost half of the country.
What's new: Climate change is increasing the length and severity of our fire seasons, and scientists are starting to quantify the health impacts of all that smoke. In a poster presented Wednesday at the fall meeting of the American Geophysical Union (AGU), Colorado State University atmospheric scientist Jeffrey Pierce estimated about 15% of the roughly 200,000 air quality related deaths in 2000 were caused by wildfire smoke. By 2100, he estimates it could reach 40%. That's just one model (some predict even more extreme fire seasons, some less) but it's clear that smoke poses a real and serious health risk.
Some good news: Although deaths from smoke increase in Pierce's model, he also shows deaths from other types of air particulates decreasing. If clean energy progresses as he has it doing in the model, the total number of air quality related deaths ultimately holds steady at 4%.
Yes, but: that 4% number is the average across the entire country. In areas with increased fire activity, says Pierce, exposure could increase — while areas that are traditionally polluted by fossil fuels, like the Ohio valley, may clean up.
The immediate impacts: It's well documented that on days with lots of air pollution, deaths increase, explains Pierce. Katelyn O'Dell, a research assistant at Colorado State University, looked at rates of hospitalization for respiratory illness and inhaler refills during the 2013 Oregon wildfire season. In results presented in a poster Monday at the AGU fall meeting, she reported that on a mildly smoky day (10 micrograms of the particle PM 2.5 per each cubed meter of air), there was a 7% increase in medication refills. But on a day with a PM 2.5 measurement of 100, there's over a 100% increase. O'Dell tells Axios that increase can last a week after the smoke exposure ends.
An underestimation: “The impact of smoke exposure is probably a lot greater than what we're saying," says O'Dell. The model doesn't catch people who get sick, but don't need to go to the hospital, for example. Her team is also starting to look at the impacts on heart health. “This is only the tip of the iceberg," she says.
Off the charts: In the 2017 fire season, regions of the US had PM2.5 levels of 500 or higher — some people were inhaling an entire years' worth of particulate matter in a single day.
The long-term impacts of breathing smoke are hard to parse, because it's hard to measure just how much smoke someone has been exposed to. But “there's a mechanism, and a pathology, so it makes sense it would have an impact," says Manvendra Dubey, a researcher at Los Alamos National Laboratory.
What's in the smoke: Plants uptake different types of salts and heavy metals (think: spinach is rich in iron) so they give off different types of particles when they burn. Dubey can even identify the region smoke came from by its chemical composition. He's creating models of what burns in different areas to help study the health impacts, since many of those particles are known to cause cancer. Dubey is particularly concerned about exposure to smoke from wildfires that destroy entire neighborhoods because there are many different types of fuel burning in a house.
The bottom line: “If people in California are being exposed to these smoke events more regularly and in higher concentrations, we would expect this to have an impact on the average lifespan of people in California," says Pierce. The same is true everywhere else fires increase.
Correction: This story initially stated that an estimated 25% of the roughly 200,000 air quality deaths in 2000 were due to smoke inhalation, and that number would increase to 75% in 2100. The correct number is roughly 15% (25,000 deaths) increasing to roughly 40% (75,000.)
Medicine is poised to be one place where AI makes a mark. In a study published this week, researchers report that a machine algorithm was as good — or better — than pathologists at detecting the spread of a type of breast cancer.
For all the talk about the promise of AI radically changing medicine, this is one of the first peer-reviewed studies to back claims that algorithms can detect abnormalities in pathology slides, says Eric Topol from the Scripps Research Institute.
The bottom line: Radiologists and pathologists are likely to be the first in medicine affected by AI. But researchers working on the technologies don't see them replacing doctors, and instead aiding them. And even that role will require more data about the impact on the medical profession and whether AIs are accurate enough to diagnose patients.
“It is the early days,” Aidoc CEO Elad Walach says. “There’s not enough research at this point. Deep learning has been commoditized generally but it hasn’t been commoditized for the medical domain. The algorithms out there aren’t good enough as is. We need a lot of R&D to make AI work in this space. It is not just plug and play.”
What’s new: Babak Ehteshami Bejnordi and his colleagues from Radboud University Medical Center in the Netherlands evaluated algorithms submitted in a competition to analyze tissue samples from the lymph nodes of breast cancer patients. (Cancer cells are most likely to spread to these nearby areas first so they're involved in determining a patient's prognosis.) They then compared the accuracy of AI diagnoses with those of pathologists in two different situations where the researchers had a gold-standard test to check both:
The result: The top seven algorithms — all deep learning methods, which have lately seen progress in image and pattern recognition — performed better than the pathologists in identifying the metastases, but were on par with the pathologist whose time wasn’t restricted.
Keep in mind: The time constraints put — or not — on the pathologists in the study aren't the reality in which they practice. And, the AIs detected just one type of breast cancer. "We need to see it borne out across lots of other pathologies not just lymph nodes for breast cancers," says Topol. "This is the most impressive paper yet. But there are limitations. This is done in silico and is not a real world validation."
PathAI: Andy Beck, whose team won the AI competition in the new study and who is now CEO of PathAI, says AI’s arrival to pathology will be a transformation rather than a disruption.
Seeing it as the latter “betrays a lack of understanding of how these fields operate. There are so many things physicians do. Typically an AI does one specific thing very well. We aren’t even close to doing the whole breadth of what a physician does.”
Aidoc: This Israeli startup is developing technology that can detect visual abnormalities — whether it be a cancer, stroke, bleeding or an edema — in head and neck CT scans. Their focus right now, says Walach, is on optimizing radiologists’ workflow from the current practice of reviewing cases in order received to getting AI to flag urgent ones first.
They’re currently testing their technology in 5 U.S. sites. Early, unpublished results at one hospital found that the AI could spot an abnormal scan with 98% sensitivity compared to what clinicians call the "ground truth" (in this study, the diagnosis by three radiologists working without AI), says Walach. They will look to publish their findings soon.
Walach says, “There is a need for peer reviewed publications about the outcomes not just the accuracy of these systems, and leading companies should invest time and resources in publishing clinical evidence.”
A big challenge: Like other cancer tests, there is a risk of detecting — and then treating — a cancer that isn't there. That isn't unique to machines but "algorithms are tuned to perform at maximum sensitivity, meaning there may be false positives," says Stanford University's Daniel Rubin, who develops imaging tools for radiology. “As we introduce these technologies, if people don’t improve accuracy and there are more false positives, it will increase the cost of health care.”
Go deeper: We asked four medical experts whether AI might help their profession
We used to define fire by seasons: they varied from place to place, but there was a period of time that fires were not reliably seen before, and a date they probably wouldn’t be seen after. That is no longer the case, as the destructive fires burning this December in Southern California make clear.Data: CAL FIRE. Data was compiled with help from Jill Hubley. Get the data. Chart: Lazaro Gamio / Axios
The above chart shows all fires that burned over 300 acres each year from 2000 to 2017 in California, including this months’ blazes.
The future, right now: We tend to talk about extreme weather and fire events as a "glimpse into our future under climate change." But these previously-rare events are increasingly common. "If ‘unexpected’ becomes the norm, because we only talk about extreme weather, how do we change the conversation?" Jeff Rosenfeld, the editor in chief of the bulletin of the American Meteorlogical Society, asked at the American Geophysical Union's annual fall meeting on Wednesday.
The 2017 fire season was one of the worst ever seen across North America:
It's only getting worse: Wildfire seasons are expected to last longer and burn hotter as climate change makes many areas warmer and dryer. "Warmer spring causes earlier snowmelt, [and] warmer summer temperatures can dry out the wood," says Jeffrey Pierce of Colorado State University.
Go deeper: Inside Climate News looks at our future with fire in a warming world.
Ginkgo Bioworks is in the business of designing and printing microbes, but it recently raised $275 million — bringing its valuation to over $1 billion — because it sees itself more akin to a software company.
Why it matters: The Boston-based company says that, by providing specialized services for designing, developing, and printing DNA, it can open possibilities for customers in industries outside of health care.
For example, a fragrance company can pay Ginkgo Bioworks for a project that would otherwise be too expensive to do in-house, according to co-founder and CEO Jason Kelly. A pharma company would have to hire scientists to do the development itself.
Deal details: Investors in the round include Viking Global, Y Combinator's Continuity Fund, Cascade Investment (Bill Gates' investment vehicle), and General Atlantic. Though Kelly declined to specify the new valuation, a Delaware filing first spotted by Recode pegs it at $1.3 billion.
Part of the proceeds will be used to open a third lab facility.
Women living within half a mile from hydraulic fracturing sites are 25% more likely to have babies with low birth weight than mothers who lived more than two miles beyond the sites, according to a new study released Wednesday.
Why it matters: The findings by researchers at the University of Chicago, Princeton University and the University of California, Los Angeles suggests that hydraulic fracturing — a technique used to force out oil and natural gas from the earth — imposes negative health impacts on locals despite the enormous economic benefits it generates. While most drilling operations are in remote areas, some sites in places that are heavily populated.
However, the researchers told The Washington Post that their intention was not to condemn fracking, adding that “There’s a big effect within one kilometer of sites, which the oil and gas industry dislikes, but the impact on the population beyond that may not be massive, which opponents of fracking won’t like.”
How it was done: Researchers examined the weights of more than 1.1 million infants born to mothers living at different distances from active sites in Pennsylvania between 2004 and 2013, when hydraulic fracking transformed the state into a major producer of natural gas.
At least three instances of extreme weather would not have happened without climate change, according to the American Meterological Society’s annual report on extreme weather and climate change. Past reports found certain weather events were ‘influenced’ or made more frequent by climate change, but the tools researchers used weren’t powerful enough to measure just how much climate change played a role. This is the first time the report has definitively pointed the finger at global warming.
Why it matters: These weather anomalies are becoming more common, say the report authors, and they can have massive health and economic impacts. If the role climate change played in causing them can be pinpointed, researchers may be able to better predict how climate change might impact our future. For example, by understanding how marine heat waves change weather, scientists were able to predict the 2016/2017 Somalian drought, and mitigate some of the loss of life.
What they did: Researchers create models of the world’s weather with and without the influence of human-caused climate change. Then, they quantify the likelihood of the differences between the two occurring by chance. Models like this have been used to effectively predict the impacts of climate events like El Nino and La Nina.
They found three major weather events “could not have happened without climate change:
The impact: The global heat record and heatwave in Asia caused deaths, fires and crop loss, and the Blob caused fish stocks to crash, harmed seabird populations and led to harmful algal blooms that closed fisheries along the length West Coast.
“We’ve known for a long time that climate change can alter the risk of some of these extremes,” says Stephanie Herring, an author on the report and scientist with NOAA, but “it always fell in the realm of possibility that they could have happened without climate change.”
There were also several weather events identified in the report that were made worse by climate change. Herring highlighted:
Yes, but: Not all extreme weather was related to climate change — snow storm Jonas, for example, did not appear to be linked to global warming. Over 130 papers have been published since the report was first created, and over half failed to find an association between climate change and the event they were examining.
The report does not set out to prove that climate change is influencing extreme weather. Instead, it aims to hold such claims to high standards of scientific rigor, and help improve the methodology used to pinpoint the specific impacts of climate change.
Looking forward: We’re entering a new era of how we talk about extreme weather, according to Jeff Rosenfeld, editor and chief of the bulletin of the American Meteorolgical Society. “We can no longer be shy talking about the connection between human-caused climate change and extreme weather.”
Chris Funk, a researcher with the USGS who was also involved in the report, adds that now “culturally and scientifically, we need to expect the unexpected” and learn how to deal with our extreme weather as the new normal.
The bottom line: Scientists are now saying that they’re ‘virtually certain’ specific extreme weather events would not happen without climate change. “I’ve never seen that sort of language until now,” says Rosenfeld. “’Virtually certain’ is almost unheard of.”
Anti-aging company Calico has remained fairly quiet since being formed in 2013 via around a $1 billion investment from Google. But that changed a bit today when senior executive Daphne Koller was interviewed at a San Francisco conference hosted by CB Insights:
Aging is a universal societal problem.
— Daphne Koller, Calico's chief computing officer and former CEO of Coursera
Koller says that Calico's primary research involves 750 mice, which are broken into five groups based on different regimes of caloric intake. The idea is to get a better sense of the aging trajectory of mammalian organisms and how caloric intake — the "one intervention shown to extend life among multiple species" — can also be affected by genetics and the environment.
Calico also is exploring the issue of cellular aging since, as Koller says, "as cells age, a lot of stuff begins to go wrong." It's using yeast cells for this research, and its engineers have developed a "yeast tracker" so they can determine the age of a yeast cell without having to constantly stare into a microscope or review countless hours of video to identify cell divisions.
"I don’t think we’re secretive so much as we don’t like to talk about our work until it’s complete," Koller said.
Rainfall from Hurricane Harvey, which struck Houston over the summer, was at least 15% heavier due to human caused climate change, according to two independent studies by researchers at the Lawrence Berkeley National Laboratory and Royal Netherlands Meteorological Institute. Hurricanes like Harvey are also three times more likely today than in 1900, researchers reported.
Why it matters: This isn't the first time scientists have attributed violent weather events to a warming planet, The Washington Post reports. Scientists have also warned of the increased likelihood of droughts such as the one in Texas in 2010 and floods similar to Colorado's in 2013. These findings suggest cities and communities may need to reassess their risk and find new ways to prepare for harsher weather as climate change continues.
How it works: "Climate change, caused by increasing greenhouse gases in the atmosphere from the burning of fossil fuels, is raising temperatures globally. Warmer air can carry more moisture, which can lead to more extreme rainfall events, and warmer ocean surface temperatures are known to intensify the most powerful hurricanes," according to the press release from the American Geophysical Union fall meeting taking place this week.
NEW ORLEANS — The Poás volcano in Costa Rica erupted suddenly and violently in late April, hurling chunks of rock into the air, destroying a nearby observation platform and wrecking a single piece of monitoring equipment: a sensor recording gas concentrations in the bottom of the crater. Since it’s not a good idea to walk into the heart of an erupting volcano, Maarten J de Moor, a volcanologist at OVISCORI in Costa Rica, sent in a drone.
Why it matters: de Moor’s drone (and his destroyed sensor) recorded changes in gas concentrations before the eruption, which he presented at the fall meeting of the American Geophysical Union. Such changes can signal if, and when, a volcano might erupt, and may even indicate how large the eruption will be. “If we can measure the gas compositions during eruptions, we can learn about those eruption processes,” says de Moor.
There are several different ways to monitor a volcano, including:
The old-fashioned way, where you dress head-to-toe in protective gear and gather fumes from a vent with a test-tube.
With a permanent sensor, like the MultiGAS sensor de Moor used.
With a sensor (in de Moor's case, a miniature MultiGAS) attached to a drone.
Yes, but gas sensors aren’t the only thing you can attach to a drone, and inactive volcanoes are also worth monitoring. That's why Einat Lev trekked for days to a remote Chilean volcano (accompanied by an entourage of horses and grad students), hauling batteries, an electric generator and a drone.
What she did: Lev, a volcanologist at Columbia University, uses the unmanned vehicles to scan the topography of cooled lava flows. Although you can get some idea of lava flow structure by looking at satellite images, the level of detail provided by drone can't be matched. By measuring the rocky dips and ridges, she can infer how fast the lava was moving, how hot it was, how liquid or solid it was. Lev also presented her findings at the meeting.
Why she did it: If we learn how a volcano erupted in the past, we might better understand how the same — or a different — volcano might erupt under similar circumstances.
A bonus: Drones might seem expensive, but it costs way more to do this sort of research from a helicopter.