news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

On July 24 (Wednesday), the main contents of the well-known foreign scientific website are as follows:

Nature website (www.nature.com)

1. Google AI tool can predict long-term climate trends and weather in minutes

Google's machine learning model NeuralGCM is a computer model that combines traditional weather forecasting techniques and machine learning. It outperforms other artificial intelligence (AI)-based tools in predicting weather scenarios and long-term climate trends.

The tool, published in a recent issue of the journal Nature, is the first machine learning model capable of generating accurate comprehensive weather forecasts - for a range of weather scenarios. Its development opens a new chapter in weather prediction, with predictions that are faster, less energy-intensive, and richer in detail than models based entirely on artificial intelligence.

Current weather forecasting systems typically rely on general circulation models (GCMs), which simulate Earth's ocean and atmospheric processes through the laws of physics and predict how these processes may affect weather and climate. But GCMs require a lot of computing resources, and advances in machine learning are beginning to provide more efficient alternatives.

Stephan Hoyer, an AI researcher at Google Research, and his team developed and trained NeuralGCM, a model that combines a traditional physics-based atmospheric solver with some AI components. They used the model to make short- and long-term weather and climate forecasts. To evaluate the accuracy of NeuralGCM, the researchers compared its predictions with real-world data and the outputs of other models, including GCM and models based purely on machine learning.

Like current machine learning models, NeuralGCM can produce accurate short-term, deterministic weather forecasts one to three days in advance, using a fraction of the energy required by existing tools. But when making long-term forecasts beyond seven days, its error rate is much lower than other machine learning models. In fact, NeuralGCM's long-term forecasts are similar to those of the European Centre for Medium-Range Weather Forecasts' ensemble model ECMWF-ENS, which is widely considered the gold standard for weather forecasting.

2. After a major technological breakthrough, scientists are expected to create the largestHeavy elements

Researchers have demonstrated a new way to create superheavy elements, providing a way to make element 120, the heaviest element in the universe yet.

Scientists at the Lawrence Berkeley National Laboratory (LBNL) in the United States announced that they have successfully created a known superheavy element, element 116, for the first time using a beam of titanium. After upgrading the laboratory's equipment, the team plans to use similar techniques to try to create element 120. The heaviest element created so far is Og, element 118, which was first synthesized in 2002.

The LBNL team presented their findings at the Nuclear Structure 2024 conference in Lemont, Illinois, USA, and posted a preprint on the arXiv server.

Superheavy elements don't occur naturally on Earth, but scientists think they may appear in stars. They are highly radioactive and break down quickly through nuclear fission, which has few immediate practical applications. But by creating new elements, scientists deepen their understanding of how the universe works and fill in theoretical models about the behavior of the atomic nucleus and its limits - such as how many protons and neutrons it can hold.

To create the new element, the researchers usedparticleAccelerators collide beams of ions with atoms in solid targets in the hope of sparking nuclear reactions that would fuse the nuclei and create elements with more protons and neutrons. But the raw materials available are running out of steam. A recently discovered group of superheavy elements, numbered 114 through 118, were all created by bombarding targets made of actinides with beams of calcium-48. This isotope of calcium is particularly stable, which makes it well suited to fuel the necessary nuclear fusion reactions.

However, calcium only allows scientists to get deep into the periphery of the periodic table. Scientists have tried to create superheavy elements using beams of particles heavier than calcium-48, including isotopes of titanium and chromium. To determine whether a titanium-50 beam could be used to create superheavy elements, the LBNL team created livermorium-290. The team used Berkeley Lab's 88-inch cyclotron facility to accelerate the titanium beam and fire it at a target made of plutonium.

Science Daily website (www.sciencedaily.com)

1. Nano-imaging technology helps understand protein and tissue preservation in ancient bones

Nanoscale 3D imaging of ancient bones not only provides further insight into the changes that soft tissues undergo during fossilization, but also serves as a rapid and practical way to determine which specimens might be suitable for preserving ancient DNA and protein sequences, according to a preliminary study from North Carolina State University.

Comparing modern bones with those from the Ice Age using nanoimaging methods provides a better understanding of the changes that collagen and blood vessels undergo during fossilization.

The researchers compared small samples of leg bones from modern cattle, crocodiles and ostriches with leg bones from mammoths, steppe bison, reindeer and horses from the Pleistocene epoch, which were extracted from melting ancient permafrost in Canada's Yukon Territory.

Using time-of-flight secondary ion mass spectrometry (TOF-SIMS) to scan and image the surfaces of the structures, the researchers identified chemical signatures present in the structures and helped further confirm that they were collagen and blood vessels.

The idea behind this initial study is that this nanoscale approach could be used on bones from all of the fossil record to better understand the chemical and structural changes that occur in organic tissues during fossilization. The technique could also potentially be used to screen ancient bone specimens for potential preservation of DNA and protein sequences.

2. Astrophysicists discover supermassive objectsBlack HoleandDark matterContact helps solve the "final parsec problem"

Researchers have discovered a link between supermassive black holes and dark matter particles, which are some of the largest and smallest entities in the universe, respectively.

Their new calculations show that a pair of supermassive black holes (SMBHs) can merge into a larger black hole because of the behavior of previously overlooked dark matter particles, providing a solution to a long-standing "final parsec problem" in astronomy. The research was published this month in Physical Review Letters.

In 2023, astrophysicists announced the detection of a “hum” of gravitational waves that permeated the universe. They hypothesized that this background signal was emitted by millions of merging pairs of supermassive black holes, each with billions of times the mass of our sun.

However, theoretical simulations suggest that as these massive objects spiral in pairs, their approach would stall at a distance of about 1 parsec (about 3 light-years), preventing them from merging.

This "last parsec problem" conflicts not only with theories that merging massive black holes are the source of the gravitational wave background, but also with theories that massive black holes formed through the merger of smaller black holes.

"We show that adding previously overlooked effects of dark matter can help supermassive black holes overcome the final separation and merger gap," said co-author of the paper. "Our calculations explain how this happens, contrary to previous views."

Scitech Daily website (https://scitechdaily.com)

1. The expiration date of drugs may affect future Mars exploration missions

A new study led by Duke Health shows that more than half of the medications stored on astronauts' spacecraft will expire before they return to Earth after a three-year journey to Mars, including major drugs such as painkillers, antibiotics, allergy medicines and sleeping aids.

Astronauts could end up relying on ineffective or even harmful drugs, according to a study published in the latest issue of the Nature Publishing Group journal npj Microgravity.

Expired medications may lose some or most of their potency. The actual stability and potency of medications in space compared to Earth remains largely unknown. The harsh space environment, including radiation, could reduce the effectiveness of medications.

As space agencies plan long-duration missions to Mars and elsewhere, expired medicines could pose a challenge to those missions, the researchers noted.

Using the International Drug Expiration Database, the researchers determined that 54 of the 91 drugs had a shelf life of 36 months or less.

The most optimistic estimate is that about 60% of these drugs will lose their effectiveness before the end of a Mars mission. Under more conservative assumptions, that number jumps to 98%.

Rather than assuming that drugs would degrade faster, the study focused on the inability to resupply them for a Mars mission. This lack of supply would affect not only drugs, but other important supplies, such as food.

Increasing the amount of medicine on board a spacecraft could also help make up for the reduced effectiveness of expired drugs, the authors said.

2. Chinese astronomers have developedgalaxyNew techniques for extracting information from surveys

Scientists from the National Astronomical Observatories of the Chinese Academy of Sciences (NAOC), in collaboration with international partners, have recently designed an innovative technique to efficiently extract information from galaxy surveys, paving the way for future cosmic exploration and surveys.

Their research results were published in the latest online issue of Communications Physics.

In this era of precision cosmology, large-scale galaxy redshift surveys are powerful tools for probing the universe. By observing a large number of spectra from distant galaxies, astronomers are able to create density fields of galaxies at different times of the universe. These density fields carry important information about the clustering of galaxies, which can be quantified using two-point and N-point (N>2) correlation functions.

However, it is difficult to use n-point functions in practice due to various complexities involving the measurement and modeling of these quantities.

After working on this challenging task for several years, the NAOC research team and partners developed a whole set of new methods to extract multi-point correlation functions from the two-point correlation functions of galaxies.

The researchers said: "This opens a new window for the effective use of high-order information in galaxy surveys, and has important cosmological implications for upcoming galaxy surveys including the Dark Energy Spectroscopic Instrument (DESI), the Fixed Focus Spectrograph (PFS), and the China Survey Space Telescope (CSST). (Liu Chun)