Feed aggregator

Gully erosion prediction using weight of evidence and advanced machine learning models

Publication date: 15 January 2026

Source: Advances in Space Research, Volume 77, Issue 2

Author(s): Xiaohui Chen, Alireza Arabameri, M. Santosh, Hasan Raja Naqvi, Mohd Ramiz

CAU-Net: An attention-based feature enhancement model for ground-based cloud image segmentation applicable to <em>peri</em>-solar regions

Publication date: 15 January 2026

Source: Advances in Space Research, Volume 77, Issue 2

Author(s): Junye Zhu, Yutong Liu, Kefan Xu, Yangshu Lin, Keqi Wang, Zhiming Lin, Qiwen Jin, Chao Yang, Lijie Wang, Chenghang Zheng, Yongxin Zhang, Xuecheng Wu

How the ocean's hydrothermal systems made the first life on Earth possible

Phys.org: Earth science - Thu, 01/22/2026 - 17:40
Our planet is unique for its ability to sustain abundant life. From studies of the rock record, scientists believe life had already emerged on Earth at least 3.5 billion years ago and probably much earlier.

Increased soil salinity alters global inorganic carbon storage, finds study

Phys.org: Earth science - Thu, 01/22/2026 - 17:22
A new global study shows that increasing soil salinity is systematically reshaping the storage and distribution of soil inorganic carbon (SIC), a key but often-overlooked part of terrestrial ecosystems. The findings, published in the Proceedings of the National Academy of Sciences on January 20, provide the first comprehensive global assessment of how soil salinization influences inorganic carbon storage and highlight its implications for the global carbon cycle.

Beneath Antarctica's largest ice shelf, a hidden ocean is revealing its secrets

Phys.org: Earth science - Thu, 01/22/2026 - 17:06
Beneath Antarctica's Ross Ice Shelf lies one of the least measured oceans on Earth—a vast, dark cavity roughly twice the volume of the North Sea.

Rainfall–salinity link sustains prolonged La Niña events, study reveals

Phys.org: Earth science - Thu, 01/22/2026 - 16:23
La Niña—a climate phenomenon characterized by unusually cool sea surface temperatures in the central and eastern tropical Pacific Ocean—can persist for multiple years, exerting significant climate impacts worldwide. In recent decades, such prolonged La Niña events have grown more frequent. However, the mechanisms that sustain these multiyear cooling episodes have remained unclear.

Scientists may have solved 66 million-year-old mystery of how Earth's greenhouse age ended

Phys.org: Earth science - Thu, 01/22/2026 - 14:54
A 66 million-year-old mystery behind how our planet transformed from a tropical greenhouse to the ice-capped world of today has been unraveled by scientists. Their new study has revealed that Earth's massive drop in temperature after the dinosaurs went extinct could have been caused by a large decrease in calcium levels in the ocean.

Early news about the 22 January 2026 landslide at Mount Maunganui in New Zealand

EOS - Thu, 01/22/2026 - 08:10

Eight people have been killed or are missing in two landslides triggered by heavy rainfall in New Zealand

Substantial parts of New Zeealand have been suffering extreme rainfall – yet again – causing floods and landslides. The most serious event to date occurred at a camp site at Mount Maunganui on the Bay of Plenty in the North Island. Here, a landslide devastated a campsite close to the coast. Unfortunately, January is the main summer holiday period in New Zealand.

Stuff has a video of the landslide as it occurred. Meanwhile, The Guardian has a Youtube video with imagery of the aftermath:-

This still shows the basic components of the failure:-

The aftermath of the 22 January 2026 landslide at Mount Maunganui. Still from a video posted to Youtube.

The location is reported to be the Mount Maunganui Beachside Holiday Park. This makes the location [-37.63234, 176.17507]. This is Google Earth image of the site:-

Google Earth image of the site of the 22 January 2026 landslide at Mount Maunganui.

The image suggests a complex geology, with maybe a hint of previous landslides (this is very speculative). The geology of this area is primarily volcanic rocks, which may indicate a high landslide susceptibility. The images of the aftermath appear to suggest deeply weathered soils, and note the amount of water flowing through the debris.

News reports indicate that at least six people are missing, some of whom are children., The authorities are continuing to describe the operation at the site as a rescue.

Meanwhile, two other people were killed by an early morning landslide at Welcome Bay Road in Papamoa, also on the Bay of Plenty. This appears to have occurred at about [-37.7231, 176.20896]. One News has an image of the aftermath of the event that appears to show multiple shallow landslides on the same hillside.

Return to The Landslide Blog homepage Text © 2026. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.

Evidence of 'lightning-fast' evolution found after Chicxulub impact

Phys.org: Earth science - Thu, 01/22/2026 - 06:51
The asteroid that struck the Earth 66 million years ago devastated life across the planet, wiping out the dinosaurs and other organisms in a hail of fire and catastrophic climate change. But new research shows that it also set the stage for life to rebound astonishingly quickly.

World on track to breach 1.5°C target by 2030

Phys.org: Earth science - Thu, 01/22/2026 - 04:10
Global average temperature increases could pass the 1.5 degrees Celsius threshold outlined in the Paris Agreement by the end of the decade, according to the EU's Copernicus Climate Change Service, putting the world at greater risk of never-seen-before extreme weather events.

Robust probabilistic estimation of statistical variations in earthquake records: application to induced seismicity in western Canada

Geophysical Journal International - Thu, 01/22/2026 - 00:00
SummaryAccurate characterization of the magnitude-frequency distribution of seismicity, and its associated uncertainties, is essential for seismic hazard assessment. This distribution is commonly described by the Gutenberg–Richter (GR) relation, parameterized by the b-value, which has been identified as a potential proxy for investigating many spatiotemporally varying Earth phenomena. Estimating the spatiotemporal variability of b-values often requires windowing, forcing a trade-off between resolution and statistical reliability. New probabilistic methods circumvent this by inferring both the number and locations of change points directly from earthquake catalogs. Nevertheless, accurately determining the b-value remains difficult because the GR relation only holds over a limited range of magnitudes. This research develops a general statistical model to address several methodological challenges in estimating the magnitude-frequency distribution of observed seismicity, including variations in space or time. The approach simultaneously solves for the b-value and magnitude-range limits. This avoids potential bias due to inaccurate manual truncation of earthquake catalogs. The model considers the entire observed catalog and parameterizes the decay of the distribution at both low and high magnitudes. Consequently, robust uncertainties in estimated b-values reflect uncertainty in the range of magnitudes over which the GR relation is observed to be valid. Importantly, spatiotemporal variations in the parameters that define the magnitude range are considered to be independent from the b-value, as we assume the physical factors that influence the GR relation are independent of the factors that limit the observed earthquake catalog. We demonstrate this methodology through application to simulated and observed earthquake catalogs. In particular, the value of our approach is highlighted through application to observed records of induced seismicity associated with fluid-injection operations in western Canada. Our results demonstrate accurate b-value estimates and associated uncertainties. Furthermore, the additional parameters that define the magnitude range serve as proxies for other factors including seismic network performance, recording duration, potential geometric limitations on earthquake size, and potential injection characteristics (in induced seismicity cases). Our approach also allows for the investigation of how these other factors may vary in space/time. Results from this work contribute to rigorous propagation of accurate b-value estimates, including uncertainties, into subsequent analyses such as seismic hazard models and regulatory protocols that are applied to industrial activity.

Machine learning for data-driven pattern recognition of seismic wind turbine emissions

Geophysical Journal International - Thu, 01/22/2026 - 00:00
SummarySeismic emissions from wind turbines (WTs) depend on the rotation of the WT blades and the wind direction-dependent movement of the WT. Mechanical coupling between the WT foundation and the subsurface generates complex seismic wavefields, making it challenging to manually separate the contributions of different signal sources, thus complicating data labelling. We address this challenge by applying unsupervised machine learning techniques that do not require labelled data. Our analysis focuses on seismic WT emissions recorded near Wind Farm Tegelberg in the eastern Swabian Alb, Southwest Germany. Specifically, we extract time-averaged wavelet features by temporal averaging the wavelet transformation of the continuous three-component seismic data and subsequently apply the clustering algorithm Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN). The resulting clusters not only capture the variations in the WT rotation rate but also reveal a clear dependency on wind direction, associated with the radiation pattern of different surface waves. Our results demonstrate the potential of HDBSCAN to uncover meaningful, source-related patterns in continuous seismic records.

Trump Administration to Speed Up Permitting for Deep Sea Mining, Even Beyond U.S. Boundaries

EOS - Wed, 01/21/2026 - 18:07
body {background-color: #D2D1D5;} Research & Developments is a blog for brief updates that provide context for the flurry of news regarding law and policy changes that impact science and scientists today.

NOAA has finalized a rule that will expedite the permit and license application process for deep seabed mining and allow companies to mine beyond U.S. jurisdictional boundaries.

The changes were published in a 113-page regulation on 21 January.

The changes revise the Deep Seabed Hard Mineral Resources Act (DSHMRA) of 1980, which required individuals or corporations who wanted to explore and mine mineral-rich nodules in the deep sea to apply for an exploration license and a commercial recovery (large-scale extraction) permit separately. Now, applicants may apply for both the exploration license and commercial recovery permit at the same time. 

“By issuing the permit simultaneously, they’re committing to exploitation without the information that you would need to evaluate its impacts.”

“Deep seabed mining is key to unlocking a domestic source of critical minerals for the United States,” Neil Jacobs, NOAA administrator, said in a statement. “This consolidation modernizes the law and supports the America First agenda by enabling U.S. companies to access these resources more quickly, strengthening our nation’s economic resilience and advancing the discovery and use of critical seafloor minerals.” 

Critics are concerned that the move will loosen environmental oversight. “By issuing the permit simultaneously, they’re committing to exploitation without the information that you would need to evaluate its impacts,” Emily Jeffers, senior attorney at the Center for Biological Diversity, told Agence France-Presse.

Beyond U.S. Boundaries

The updated rule also states that DSHMRA gives NOAA the ability to issue exploration licenses and permits for the seabed beyond national jurisdiction. The International Seabed Authority (ISA), an autonomous international governing body, regulates deep sea mining in international waters for countries that are part of the 1982 Law of the Sea Convention. The United States has never been a party to that treaty but has mostly followed its guidelines.

Now, NOAA’s insistence that the United States can regulate U.S. companies’ deep sea mining beyond U.S. waters is expected to cause controversy among members of the ISA, which has for years been negotiating rules to govern mining in international waters. In December, the Trump administration announced it had received an application for mining exploration in international waters from the Metals Company.

The final rule follows an executive order issued last year calling for the rapid development of deep sea mining capabilities both domestically and beyond U.S. jurisdictional boundaries.

 
Related

In a response to that order, the ISA called it “surprising because for over 30 years the US has been a reliable observer and significant contributor to the negotiations of the International Seabed Authority.” In the statement, the ISA also said any unilateral action to mine the deep sea “sets a dangerous precedent that could destabilize the entire system of global ocean governance.”

The deep sea has never been commercially mined. Compared to other ecosystems, little is known about the ecology of the ocean floor or how these ecosystems support marine life. Disturbing these ecosystems could have wide-ranging consequences.

“Once nodules are removed by mining, all biodiversity and functions directly dependent on the minerals will be lost for millions of years at the mined location, as nodules need millions of years to re-form,” Sabine Gollner, a deep-sea marine biologist at the Royal Netherlands Institute for Sea Research told Eos in 2024. 

—Grace van Deelen (@gvd.bsky.social), Staff Writer

These updates are made possible through information from the scientific community. Do you have a story about how changes in law or policy are affecting scientists or research? Send us a tip at eos@agu.org. Text © 2026. AGU. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.

US forests store record carbon as natural and human factors combine

Phys.org: Earth science - Wed, 01/21/2026 - 17:33
U.S. forests have stored more carbon in the past two decades than at any time in the last century, an increase attributable to a mix of natural factors and human activity, finds a new study.

Grains of sand prove people—not glaciers—transported Stonehenge rocks

Phys.org: Earth science - Wed, 01/21/2026 - 17:10
Ask people how Stonehenge was built and you'll hear stories of sledges, ropes, boats and sheer human determination to haul stones from across Britain to Salisbury Plain, in south-west England. Others might mention giants, wizards, or alien assistance to explain the transport of Stonehenge's stones, which come from as far as Wales and Scotland.

Critical Atlantic Ocean currents kept going during last ice age, microfossils suggest

Phys.org: Earth science - Wed, 01/21/2026 - 16:00
During the last ice age, the Atlantic Ocean's powerful current system remained active and continued to transport warm, salty water from the tropics to the North Atlantic despite extensive ice cover across much of the Northern Hemisphere, finds new research led by UCL scientists.

Dredging sand and silt has consequences for the North Sea

Phys.org: Earth science - Wed, 01/21/2026 - 13:59
Through sand extraction and the disposal of dredged harbor silt, about 200 million tons of sediment are relocated every year in the coastal waters of the North Sea. The Wadden Sea is particularly strongly affected. This is the result of a new study by the Helmholtz Center Hereon, which for the first time evaluated comprehensive data on dredging activities along the North Sea coasts.

Why Are River Deltas Disappearing? They’re Sinking Faster Than Many People Realize

EOS - Wed, 01/21/2026 - 13:54

This story was originally published in the Louisiana Illuminator.

A new study says river deltas around the world aren’t just disappearing because of rising seas, but also because the land itself is sinking down into the waters, either as fast or faster than the rising oceans.

Researchers found some of the most rapid sinking is happening along the Mississippi River Delta in Louisiana. The study aims to better guide coastal restoration in disappearing river deltas around the globe, helping leaders, scientists and people living in coastal communities with hard decisions on what can— and should—be saved.

“Coastal areas account for less than 1% of the entire land area we have,” said Leonard Ohenhen, a professor at the University of California, Irvine and the lead author of the study. “But a whole significant population, more than 600 million people, live in those areas.”

“You have a sort of a hodgepodge of different reasons why deltas are sinking.”

The study, published this week in the academic journal Nature, found the contributions of  subsidence, or slowly submerging land, to disappearing coasts is often overlooked.

The fight to preserve rapidly sinking land has been a decades-long battle in the Mississippi River Delta, as well as a source of contention between scientific and political figures in the state. But deltas across the world are sinking, too, and fast.

“You have a sort of a hodgepodge of different reasons why deltas are sinking,” Ohenhen said.

He said river deltas naturally sink to some degree, with sediment carried downstream by rivers piling up and pushing down on the spongy, soft land already there. Humans can accelerate this natural process by engineering rivers such as the Mississippi and by extracting groundwater or oil.

“Relative sea level rise in the area is also really important. That’s the sea level rise plus subsidence,” said Alisha Renfro, a coastal scientist with the National Wildlife Federation. “It really helps us understand where we can make investments in restoration long-term that we might actually be able to hold on to.”

This map from the report shows which areas of the Mississippi River Delta are sinking. Areas in red and yellow are areas of land sinking more rapidly, while spots in blue and purple are building land upwards. Credit: University of California, Irvine

Lack of sediment is the main driver of subsidence in the Mississippi River Delta, Ohenhen said, creating hotspots of rapidly sinking pieces of land amid slightly more stable areas. Most of the deltas studied in the paper, around 70%, have subsidence problems primarily the result of groundwater withdrawal. But some, like the Amazon and Mississippi deltas, had subsidence issues driven by the disappearance of river-carried sediment to replenish the delta’s land.

Putting hard numbers to and pinpointing causes of subsidence—like human activity—is invaluable to restoring coastal land.

“I would say that really validates what, not just my organization, but what a lot of people have recognized for a long time—that this was a significant contributing factor in subsidence,” said James Karst with the nonprofit advocacy group Coalition to Restore Coastal Louisiana, referring specifically to the lack of sediment sent to the Mississippi River Delta.

Decisions about what pieces of the coast can be saved are even more urgent with the cancellation of two large-scale restoration projects in Louisiana.

Known as sediment diversions, the Mid-Barataria and Mid-Breton plans would have diverted freshwater from the Mississippi River into surrounding wetlands. They were scrapped by the state because of the prospective impact on fisheries for oysters, crabs and other marine species. Fish and oyster harvesters celebrated the projects’ demise, while scientists and coastal restoration advocates warned that time is running out to save the coast.

Boaters fish in the canals and wetlands just outside of New Orleans, Louisiana. Coastal restoration projects spearheaded by the state hope to preserve areas of subsiding land that are at risk of disappearing. Credit: Elise Plunk/Louisiana Illuminator

“In light of the cancellation of Mid-Barataria, I think what we, everybody, should be thinking of is, ‘What is the next best thing?’” Karst said. “Clearly it is not going to move forward, but we can’t do nothing.”

“People should be aware that we are in a part of the world that is changing and that is changing rapidly.”

“People should be aware that we are in a part of the world that is changing and that is changing rapidly,” he added. “If we want to position ourselves as individuals and as communities, we should be anticipating these changes and anticipating how they will affect us.”

While the average rate of subsidence for the Mississippi River Delta is around 3.3 millimeters per year, Ohenhen said, some areas of Louisiana are sinking at a rate of 3 centimeters per year, one of the fastest rates of all the deltas studied. That is paired with sea level rising by at least 7 millimeters per year along the Gulf Coast, he said, also one of the highest rates in the world. This puts some areas of Louisiana’s land at higher risk of loss than anywhere else.

“In the Mississippi River Delta, for example, that is one of the only deltas in the world where you have active relocation of people from the delta due to land loss,” Ohenhen said. “The time that we need to respond to these changes is now before the situation gets significantly worse.”

—Elise Plunk (@plunk.bsky.social), Louisiana Illuminator

This story is a product of the Mississippi River Basin Ag & Water Desk, an independent reporting network based at the University of Missouri in partnership with Report for America, with major funding from the Walton Family Foundation.

The underlying causes of the 8 February 2025 Junlian rock avalanche in Sichuan Province, China

EOS - Wed, 01/21/2026 - 07:27

A new paper (Jia et al. 2026) has found that the 8 February 2025 Junlian rock avalanche was caused by progressive weakening of the rock mass through wetting and drying cycles.

On 8 February 2025, the major Junlian rock avalanche landslide occurred at Jinping Village in Sichuan Province, China. A paper (Jia et al. 2026) has now been published in the journal Landslides that provides more details about the possible causes of this event. This link should provide access to the paper.

An earlier paper (Zhao et al. 2025), which I noted in June, has already described this landslide. This is a photograph of the aftermath of this event:

The aftermath of the 8 February 2025 Junlian rock avalanche in Sichuan, China. Image by Xinhua.

Unfortunately, the paper does not give a lat / long for this landslide, but I have previously noted that it is at [27.99885, 104.60801].

As a reminder, Zhao et al. (2025) determined that the initial failure was 370,000 m3, increasing to 600,000 m3 through entrainment. The landslide had a runout distance of 1,180 metres and a vertical elevation change of 440 m. In total, 29 people were killed.

The slightly odd thing about this failure is that the rainfall event that appears to have triggered it was unexceptional (c. 85 mm over the previous 30 days). I hypothesised that a progressive failure mechanism could have been in play.

Jia et al. (2026) have made some really interesting observations. First, this site was subject to previous landslides, most notably in February 2013. The paper notes that:

“all 173 people from 29 households under threat [from this earlier event] were included in the geohazard risk avoidance relocation subsidy program. Some farmers self-demolished their houses, but as some occasionally returned during the farming season, the Mu’ai Town Government, with support from the county government, organized mandatory demolition of unremoved houses in the area in 2018. ”

Further failures occurred in 2021 and 2022, whereupon all the households immediately below the unstable slope were relocated. However, homes located at a greater distance from the cliff were left in place – these were the people affected by the 2025 event.

Jia et al. (2026) suggest that initial movement of the landslide in the years before 2025 weakened the rock mass and opened pathways for the movement of water into the shear zone. Critically, their work suggests that successive wetting and drying cycles led to degradation of the the sandstones and mudstones forming the slope, moving the mass towards failure.

This weakening was sufficient to render the slope vulnerable to the effects of the rainfall in February 20925, triggering the Junlian rock avalanche.

We might take away to key messages from this work. The first is the need to understand the likely runout characteristics of a slope in determining the safety of the population. This is devilishly difficult. That there was an ongoing programme to relocate the most vulnerable people is (on the face of it) good, but it depends on this calculation.

Second is the need to understand the complexities of the processes occurring in a slope. In the case of the Junlian rock avalanche, it was the progressive weakening of the rock mass through wetting and drying cycles that meant that the slope could fail under the influence of unexceptional rainfall. As we drive climate change, similar processes will be occurring in many more slopes in China and elsewhere. That is going to pose a major challenge in terms of keeping people safe.

References

Jia, W., Wen, T., Chen, N. et al. 2026. Dry–wet cycle may trigger the catastrophic landslide in Junlian on February 8, 2025Landslides. https://doi.org/10.1007/s10346-026-02692-2

Zhao, B., Zhang, Q., Wang, L. et al. 2025. Preliminary analysis of failure characteristics of the 2025 Junlian rock avalanche, ChinaLandslideshttps://doi.org/10.1007/s10346-025-02556-1.

Return to The Landslide Blog homepage Text © 2026. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.

The way Earth's surface moves has a bigger impact on shifting the climate than we knew

Phys.org: Earth science - Wed, 01/21/2026 - 01:50
Our planet has experienced dramatic climate shifts throughout its history, oscillating between freezing "icehouse" periods and warm "greenhouse" states.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer