Feed aggregator

The Atlantic Meridional Overturning Circulation (AMOC) is an enormous loop of ocean current in the Atlantic Ocean that carries warmer waters north and colder waters south, helping to regulate the climate in many regions. The collapse of this critical circulation system has the potential to cause drastic global and regional climate impacts, like droughts and colder winters, especially in Northwestern Europe.

Editors’ Highlights are summaries of recent papers by AGU’s journal editors. Source: Journal of Geophysical Research: Atmospheres

Atmospheric circulations are chaotic and unpredictable beyond a certain time limit. Quantifying predictability helps determine what forecast problems are potentially tractable. However, while predictability of weather close to the surface is a much-studied problem, showing a prediction limit of approximately 10 days, less is known about how predictable the atmosphere is at higher layers.

Garny [2025] applies a high-resolution global model to study atmospheric predictability from the surface to the mesosphere/lower thermosphere (MLT; 50-120 kilometers altitude), providing new understanding of coupling between atmospheric levels and fundamental behavior of the upper atmosphere. The author shows that the MLT is somewhat less predictable than lower atmospheric layers due to rapid growth of ubiquitous small-scale waves, with predictability horizons of about 5 days. However, atmospheric flows in the MLT on larger horizontal scales of a few thousand kilometers can remain predictable for up to 3 weeks.

This research highlights the importance of high-resolution, ‘whole atmosphere’ models to understand and predict circulations in the middle atmosphere and coupling from the surface to the edge of space.

Citation: Garny, H. (2025). Intrinsic predictability from the troposphere to the mesosphere/lower thermosphere (MLT). Journal of Geophysical Research: Atmospheres, 130, e2025JD043363. https://doi.org/10.1029/2025JD043363

—William Randel, Editor, JGR: Atmospheres

Text © 2025. 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.

Dust plays a major role in the formation of ice in the atmosphere. A new analysis of satellite data, published in Science, shows that dust can cause a cloud’s water droplets to freeze at warmer temperatures than they otherwise would. The finding brings what researchers had observed in the laboratory to the scale of the atmosphere and may help climate scientists better model future climate changes.

In 1804, French scientist Joseph Louis Gay-Lussac ascended to about 23,000 feet (7,000 meters) in a hydrogen balloon from Paris, without supplemental oxygen, to collect air samples. He noted that clouds with more dust particles tended to have more frozen droplets.

In the 20th century, scientists found that pure water can remain liquid even when cooled to −34.5°C. But once even tiny amounts of material, such as dust, are introduced, it freezes at much warmer temperatures.

“It’s like Schrödinger’s cat. Either there’s an ice crystal, or there’s a liquid droplet.”

In 2012, researchers in Germany were finally able to test this directly in a cloud chamber experiment. They re-created cloud conditions in the lab, introduced different types of desert dust, and gradually cooled the chamber to observe the temperatures at which droplets froze.

For Diego Villanueva, an atmospheric scientist at ETH Zürich in Switzerland and lead author of the new study, it was striking that scientists had uncovered these processes in the lab, yet no one had examined them in such detail in nature.

The challenges were obvious. To watch an ice crystal nucleate, researchers would need instruments on an aircraft or balloon to catch a micrometer-sized droplet in a cloud at just the right moment. “It’s like Schrödinger’s cat,” said Daniel Knopf, an atmospheric scientist at Stony Brook University who was not involved in the work.. “Either there’s an ice crystal, or there’s a liquid droplet.”

In the new study, Villanueva and his colleagues analyzed 35 years of satellite data on cloud tops across the Northern Hemisphere’s extratropics—a region spanning the U.S. Midwest, southern Canada, western Europe, and northern Asia. The researchers wanted to see whether dust influenced whether cloud tops were liquid or ice. They focused on cloud tops, rather than entire clouds, simply because the tops are visible in satellite imagery.

Desert Dust and Cold Clouds

Villanueva and his colleagues examined two satellite datasets covering 1982–2016, trying to infer microscopic details of cloud tops such as the number of ice crystals or droplet sizes. One dataset tracked whether cloud tops were liquid or ice, and the other measured how much dust was in the air at the same time. Although the team examined global patterns, they focused on the northern extratropical belt, where mixed-phase clouds are common and large amounts of dust from deserts like the Sahara and Gobi circulate.

But the “dataset quality was just so poor that everything that came out was basically just noise,” Villanueva added. In the end, the researchers focused on a simpler detail: the fraction of clouds with ice at their tops. “This took me nearly 3 years,” Villanueva said.

The analysis revealed that regions with more dust had more ice-topped clouds. The effect was strongest in summer, when desert winds lift the most dust.

A distinctive pattern emerged: A tenfold increase in dust roughly doubled the likelihood of cloud tops freezing. “You’d need 100 times more dust to see freezing become 4 times as frequent,” Villanueva explained.

“I think the study is quite elegant.”

The new work showed that the same processes researchers have observed at the microscale in laboratories occur at much larger scales in Earth’s atmosphere. Even after accounting for humidity and air movement, dust remained the key factor for ice nucleation in most instances, though there are exceptions. In some places, such as above the Sahara, few clouds form despite the presence of dust, perhaps, the authors suggest, because the movement of large swaths of hot air prevents freezing.

“I think the study is quite elegant,” Knopf said. He explained that taking 35 years of satellite data, finding a relationship between dust levels and frozen cloud top rates, and then showing that it lines up perfectly with lab experiments is basically “the nail in the coffin” for proving dust’s role in ice nucleation. Scientists now have robust satellite evidence of dust aerosols directly affecting cloud freezing, matching what laboratory experiments had predicted.

The finding has implications for climate modeling. To predict the effects of climate change more accurately, models must account for dust and the ways it affects cloud freezing and helps shape precipitation. Liquid-topped clouds reflect more sunlight and cool the planet, whereas ice-topped clouds let in more sunlight and trap heat.

However, Knopf noted that there is more work to be done to understand exactly what the new observations mean for scientists’ understanding of climate. “If you want to really know the precipitation or climate impacts [of dust], you really need to know the number of liquid droplets or the number of ice crystals,” he said.

Villanueva is motivated to keep looking at clouds and aerosols. In the next 10–20 years, the Earth may have drier surfaces because of climate change, which will likely produce more dust aerosols in the atmosphere. He added, “I want to know how clouds will respond in the scenario.”

—Saugat Bolakhe (@saugat_optimist), Science Writer

Citation: Bolakhe, S. (2025), Dust is the sky’s ice maker, Eos, 106, https://doi.org/10.1029/2025EO250328. Published on 5 September 2025. Text © 2025. 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.

This story was originally published in the Louisiana Illuminator.

Despite being called a “cruise,” the people on board The Pelican described the experience on the hypoxia monitoring expedition as very different from the elaborate dinners on a towering vacation ship or booze- and buffet-filled Caribbean itinerary.

Passengers describe waves up to 5 feet high in the Gulf of Mexico, swinging the 116-foot research vessel like a pendulum, plaguing anyone who didn’t have sturdy sea legs with bouts of seasickness. Daytime temperatures in late July soared ever higher as sweat dripped down the backs of hard-hat covered heads.

The guests on board The Pelican weren’t seeking pleasure or status. They were unpaid students and researchers who say they weathered the conditions in the name of science itself.

“It’s not glamorous, but it is very important.”

“It’s not glamorous, but it is very important,” said Cassandra Glaspie, assistant professor at Louisiana State University and the chief scientist for the National Oceanic and Atmospheric Administration’s annual hypoxia cruise.

The 11-day voyage provides vital information on the sea life and environmental conditions within the seasonal low-oxygen zone that develops off the coast of Louisiana. The data the cruise collects informs state and federal efforts to reduce the size of the “dead zone” and sheds light on impacts to those who rely on the water for their livelihoods, like shrimpers and fishermen.

Now, after its 40th year and 38th hypoxia cruise, The Pelican’s annually planned journey faces challenges to stay afloat, potentially undermining decades of research and future plans to get the dead zone under control.

A Decades Long Struggle

Biologists, undergraduate student researchers and crew alike celebrated the cruise’s 40th anniversary aboard The Pelican with a party that had an “old bird” theme, chosen to honor the boat, which has also been sailing for 40 years.

The Pelican and the hypoxia cruise’s 40th anniversaries party on the water. Credit: Yuanheng Xiong

More than just an excuse to eat cake (with rainbow sprinkles), the purpose of the cruise is to capture information snapshots of just how bad conditions get in the dead zone.

“We bring water up to the surface. We have a little chemistry lab…to figure out what the oxygen level is chemically, and then we can validate that against what our sensors are telling us,” Glaspie said.

The low-oxygen area appears annually as nutrients, primarily from agricultural fertilizers from the massive Mississippi River Basin, drain downriver and spur algae overgrowth.

Algae eat, defecate and die, using up the oxygen in the water when they decompose and sink to the bottom. Fish, shrimp and other marine life leave the low oxygen area when they can and suffocate when they can’t, putting pressure on the vital commercial Gulf fishery and the people who rely on it. Exposure to low-oxygen waters can also alter reproduction, growth rates and diet in fish species.

Glaspie took over the work of coastal scientist Nancy Rabalais, who launched the maiden cruise in 1985 and led it for decades after. Every summer begins with a forecast of the zone’s predicted size, estimated by various scientific models and measurements of nitrogen and phosphorus throughout the river basin taken throughout the year.

“A lot of times with pollution, you hear anecdotal evidence of how it might be increasing cancer rates or it might be causing fisheries to fail,” Glaspie said. “Here, we have an actual, measurable impact of nutrient pollution in the Mississippi River watershed.”

The Mississippi River/Gulf of America Hypoxia Task Force, an interagency federal, state and tribal effort to reduce the size of the dead zone, uses data from the cruise to determine whether it is meeting its goals.

In the past five years, the dead zone has been as large as 6,700 square miles, and even larger in previous years, reaching nearly the size of New Jersey.

While still more than two times the size that the Task Force wants, the Gulf dead zone was slightly smaller than forecasted this year, about the size of Connecticut at around 4,400 square miles.

Federal and state officials lauded the limited success of the zone’s smaller size in a July 31 press conference held to discuss the results of the hypoxia cruise’s 2025 findings. They also called for continued work.

“It requires strong collaboration between states, tribes, federal partners and stakeholders,” said Brian Frazer, the EPA’s Office of Wetlands, Oceans and Watersheds director.

Mike Naig, Iowa’s agriculture secretary, said states should be “scaling up” initiatives to reduce nutrient pollution.

Whether or not this will actually happen is uncertain.

Funding Cuts

Since the Trump administration took office, funding for nutrient reduction efforts upriver as well as money to operate the cruise itself have been scaled back or cut entirely.

The Environmental Protection Agency’s 319 and 106 funding programs under the Clean Water Act are the main funding mechanisms for states to reduce nutrient pollution throughout the Basin. Those grants aren’t funded in President Trump’s proposed FY 2026 budget, said Frazer.

The 106 programs have historically doled out $18.5 million annually, according to the EPA, with additional money sometimes allocated from Congress. The 319 program provided $174.3 million in FY 2025.

The cuts to these programs are not yet final. Congress can decide to add in additional funding, and has in past years.

States rely on both funds to reduce and monitor nutrient runoff in their waters, said Matt Rota, senior policy director for Healthy Gulf, a nonprofit research group. Rota has monitored policy changes surrounding the Gulf dead zone for more than 20 years, and he questions whether current reduction strategies can be maintained, let alone efforts redoubled.

“It’s always good to see a dead zone that’s smaller than what was predicted,” Rota said. “I am not confident that this trend will continue.”

“It’s a relatively inexpensive program. … This is baseline stuff that our government should be doing.”

Aside from cuts to reduction efforts, money for The Pelican’s annual cruise is also slipping away. Glaspie said that, ideally, the cruise has 11 days of funding. It costs about $13,000 a day to operate the vessel, she said.

“It’s a relatively inexpensive program” with big payoffs for seafood industry workers who rely on the water for their livelihoods, Rota said. “This is baseline stuff that our government should be doing.”

Funding for the hypoxia cruise has been part of the National Oceanic and Atmospheric Administration’s annual operational budget, making it a more reliable source than grant funding. But with the Trump administration taking a hatchet to government-backed research, there is increasing uncertainty over whether The Pelican and its crew will embark upon future missions.

This year, Glaspie said, NOAA defunded a day of the cruise. The Gulf of America Alliance, a partnership group to support the Gulf’s economic and environmental health amongst the five bordering states, stepped in to make up the difference. Glaspie said having that additional day was a saving grace for the research.

“This is a fine-tuned machine, and the consequences for cutting it short are really predictable and well-known,” she said. “If I’m asked to create an estimate of the surface area of hypoxia, and we’re not able to cap off the end in Texas waters, I’m not really going to be able to give a reliable estimate.”

Even without additional cuts, Glaspie said she already conducts the hypoxia cruise “on a shoestring budget.” Researchers on board don’t get paid, and every person who supports its mission—besides the crew that runs the boat—are volunteers.

“It’s tough for me to not pay people. I mean, they’re working solid 12-hour shifts. It is not easy, and they are seasick for a lot of this, and they can’t call home,” Glaspie said. “It doesn’t sit well with me to not pay people for all this work, but this is what we’ve had to do because we don’t have the money to pay them.”

Students Jorddy Gonzalez and Lily Tubbs retrieve the CTD sensor package after measuring dissolved oxygen at a regular stop on the annual hypoxia cruise while students watch. Credit: Cassandra Glaspie A Rapidly Changing Gulf

Defunding research as climate change intensifies—creating extreme heat in the Gulf—could further undermine hypoxia containment efforts and the consistency of decades worth of data collection.

“I think the rising temperatures is a big question,” Rota said.

“We have 40 years of data, which is almost a gold standard,” Glaspie said. “We’ve just reached that threshold where we can really start to ask some more detailed questions about the impacts of hypoxia, and maybe the future of hypoxia.”

Despite this year’s smaller zone surface area, low oxygen levels went deeper into the water than Glaspie had ever seen before.

“The temperature drops [as the water gets deeper], the salinity increases, and the oxygen just goes basically to zero,” she said.

In some areas, Glaspie’s measurements showed negative oxygen levels.

“Oxygen doesn’t go in the negative. It was just so low that the sensor was having trouble with it,” she said. “It’s the first time I’ve seen it like this.”

She also noticed unusually large amounts of algae on the surface of the water “like ectoplasm in Ghostbusters.”

The smaller-than-forecasted size of the dead zone surprised researchers on The Pelican who saw just how deep the low oxygen levels went.

“None of us really thought until the estimate came out that it was below average size because we’re able to see the three-dimensionality of it. That’s not really incorporated into that estimate,” Glaspie said.

She also noticed unusually large amounts of algae on the surface of the water “like ectoplasm in Ghostbusters.” Toxic algae blooms can kill fish and other sea life as well as poison humans.

“If I had to say what would be important for us to monitor in the future, it would be these algal blooms, and making sure that we’ve got a good handle on which ones have harmful species,” she said.

This is why Glaspie, donned in her sun-protective clothes and work boots, braves the waves, the heat and the journey across the Gulf every year.

“This is our finger on the pulse of our nutrient pollution problem that Louisiana is inheriting from the entire country,” Glaspie said. “We cannot take our finger off that pulse. It is unfair to Louisiana. We have this pollution problem. We need to understand it.”

—Elise Plunk (@elise_plunk), 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.

Author(s): Michael Updike, Nicholas Bohlsen, Hong Qin, and Nathaniel J. Fisch

A primary technical challenge for harnessing fusion energy is to control and extract energy from a nonthermal distribution of charged particles. The fact that phase space evolves by symplectomorphisms fundamentally limits how a distribution may be manipulated. While the constraint of phase-space vol…

[Phys. Rev. E 112, 035202] Published Fri Sep 05, 2025

Author(s): F.-X. Zhou, C.-W. Lian, R. Yan, Y. Ji, J. Li, Q. Jia, and J. Zheng

Absolute growth of stimulated Raman side scattering (SRSS) in inertial confinement fusion appears to be absent in experiments. Based on simulations the authors find that absolute growth of SRSS occurs only in the limit of an infinite laser beam width. This finding may have implications for the design of experiments.

#AdvancingField #OpenDebate

[Phys. Rev. E 112, L033201] Published Fri Sep 05, 2025

SummaryThis study focuses on the hydrophone component of a dense ocean bottom cable dataset from the North Sea. This data had already been used in the past to illustrate the high resolution power of full waveform inversion based strategies. We have developed a highly scalable implementation of a visco-acoustic full waveform inversion engine making it possible to double the frequency content of the inverted data compared to previous studies, using simultaneously up to almost 50,000 CPU. This results in a multi-parameter reconstruction of the subsurface, where the P-wave velocity, the density and the quality factor are reliably reconstructed down to 2 km depth, with a resolution of about 10 meters.

SummaryHarmaliére is an active landslide located in a mountainous region of southern France where the presence of a thick layer of clays provides favorable conditions for the development of slowly moving landslides. However, at Harmaliére, the alternation of sudden reactivation and quiet episodes suggests that specific structural and geomechanical properties control its kinematics. In order to shed light on its subsurface properties, we deployed for one month a dense network of seismic nodes within the landslide and recorded active-source and ambient noise seismic data. These datasets have been first independently processed with dedicated interferometry-based processing and inversion workflows to reconstruct P-wave (active sources) and S-wave (seismic noise) velocity models. Full wave-equation tomography is then performed to improve the reliability and resolution of the obtained elastic model by iteratively fitting the virtual gathers obtained by cross-correlation of the ambient-noise recordings. As opposed to conventional ambient-noise tomography, the approach fully accounts for topography and 3D elastic heterogeneities. The obtained high-resolution 3D models are then qualitatively interpreted in terms of landslide properties and geological lithologies, that can influence landslide kinematics.

SummaryDebris flows pose a significant threat to the sustainable development of mountainous regions. As an effective real-time sensing technique, microseismic monitoring plays a critical role in the detection and analysis of debris flow activity. However, current microseismic monitoring technologies face challenges in distinguishing mixed signals originating from different sources, limiting our understanding of the full dynamic evolution of debris flow events. To address this issue, we propose an unsupervised deep clustering-based signal classification framework, which focuses on analyzing the signal characteristics at various stages of debris flow events. A two-dimensional spectrogram dataset was constructed, encompassing signals from debris flows, rockfalls, earthquakes, and environmental noise. A deep autoencoder was employed to compress spectral features into a 16-dimensional latent space, followed by clustering using deep embedded clustering and Gaussian Mixture Models. Experimental results demonstrate that, after optimizing the feature space and data partitioning strategy, the proposed method achieves an average classification accuracy of 96.81 per cent across the four signal types. Power spectral density distribution analysis further confirms that this method not only accurately identifies debris flow signals but also effectively captures their energy distribution and dynamic evolution at different stages. Interpretability analysis reveals strong correlations between the extracted latent features and conventional seismological parameters, particularly the peak count of the time-domain autocorrelation function and the first quartile of the central frequency. Based on this method, a complete segmentation of debris flow events was successfully achieved, revealing the typical signal characteristics and temporal evolution of each stage. Cross-station validation indicates that the proposed framework demonstrates strong robustness and generalization across different monitoring locations. In addition, preliminary exploration of its integration with supervised learning suggests its potential applicability in real-time monitoring scenarios, offering a novel approach for debris flow early warning. This study presents an efficient and intelligent method for debris flow signal recognition and dynamic monitoring.

SummaryBayesian formulations of inverse problems are attractive due to their ability to incorporate prior knowledge, account for various sources of uncertainties, and update probabilistic models as new information becomes available. Markov chain Monte Carlo (MCMC) methods sample posterior probability density functions (pdfs) provided accurate representations of prior information and many evaluations of likelihood functions. Dimensionality-reduction techniques such as principal component analysis (PCA) can assist in defining the prior pdf and the input bases can be used to train surrogate models. Surrogate models offer efficient approximations of likelihood functions that can replace traditional and costly forward solvers in MCMC inversions. Many problem classes in geophysics involve intricate input/output relationships that conventional surrogate models, constructed using samples drawn from the prior pdf fail to capture, leading to biased inversion results and poor uncertainty quantification. Incorporating samples from regions of high posterior probability in the training may increase accuracy, but identifying these regions is challenging. In the context of full waveform inversion, we identify and explore high-probability posterior regions using a series of successively-trained surrogate models covering progressively expanding wave bandwidths. The initial surrogate model is used to invert low-frequency data only as the input/output relationship of high-frequency data are too complex to be described across the full prior pdf with a single surrogate model. After a first MCMC inversion, we retrain the surrogate model on samples from the resulting posterior pdf and repeat the process. By focusing on progressively narrower input domain regions, it is possible to progressively increase the frequency bandwidth of the data to be modeled while also decreasing model errors. Through this iterative scheme, we eventually obtain a surrogate model that is of high accuracy for model realizations exhibiting significant posterior probabilities across the full bandwidth of interest. This surrogate model is then used to perform an MCMC inversion yielding the final estimation of the posterior pdf. Numerical results from 2D synthetic crosshole Ground Penetrating Radar (GPR) examples demonstrate that our method outperforms ray-based approaches, as well as results obtained when only training the surrogate model using samples from the prior pdf. Our methodology reduces the overall computational cost by approximately two orders of magnitude compared to using a classical finite-difference time-domain forward scheme.

Scientists are using artificial intelligence to understand escalating unrest in Italy's Campi Flegrei, a volcanic area that is home to hundreds of thousands of people.

The seas have long sustained human life, but a new UC Santa Barbara study shows that rising climate and human pressures are pushing the oceans toward a dangerous threshold.

Dr. Seunghak Lee, Jaeshik Chung, and Sang Hyun Kim of the Water Resources Cycle Research Center at the Korea Institute of Science and Technology (KIST) observed how oil and water interact in porous media under various conditions using a microfluidic system that allows precise observation of microscopic fluid flows.

The explosion of animal life in Earth's oceans half a billion years ago during and after the Cambrian Period is commonly attributed to a substantial and sustained rise of free oxygen (O2) in seawater. Some researchers even argue for near-modern levels of ocean oxygenation at this time.

Abiotic organic synthesis during geological processes has long drawn scientific interest, as it is believed to have laid both the material and energetic groundwork for the emergence of early life on Earth.

El Niño-Southern Oscillation (ENSO) is the strongest interannual variability signal in Earth's climate system. The shifts between its warm and cold phases profoundly impact global extreme weather, ecosystems, and economic development. However, current climate models show large discrepancies in their future projections of ENSO sea surface temperature (SST) variability.

Editors’ Highlights are summaries of recent papers by AGU’s journal editors. Source: AGU Advances

Permafrost is considered a critical global component of the cryosphere given its climate-sensitive nature and its key geomorphological and ecosystem role. Permafrost is also affected by wildfires which may cause the crossing of a tipping point in cryospheric systems. In fact, wildfires may reduce vegetation, destroy organic layers, modify surface albedo, leading to active layer thickening and ground subsidence. Permafrost itself is subjected to long term deformation after wildfires, but this deformation is currently poorly understood.

Cao and Furuya [2025] use remote sensing to explore ground surface deformation signals across multiple fire scars in the past five decades in North Yukon. The authors find that post-permafrost evolution follows three distinct phases characterized by land subsidence soon after the event and final recovery of the permafrost over a 50-year timescale, which implies soil uplift. Such an uplift phase is rarely reported and is related to vegetation regeneration and soil greening after the fire. These provide thermal protection, suggesting a critical mechanism of permafrost recovery. These findings highlight the resilience of boreal-permafrost systems against wildfires, but continuous monitoring is needed as wildfire and climate change intensify.

Citation: Cao, Z., & Furuya, M. (2025). Decades-long evolution of post-fire permafrost deformation detected by InSAR: Insights from chronosequence in North Yukon. AGU Advances, 6, e2025AV001849. https://doi.org/10.1029/2025AV001849

—Alberto Montanari, Editor-in-Chief, AGU Advances

Text © 2025. 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.

Undergraduate research experiences (UREs) in science, technology, engineering, and mathematics (STEM) offer students hands-on research experience and essential professional skills and connections to prepare them to succeed in the workforce. They also cultivate students’ sense of belonging, confidence, and identity—and promote retention—in STEM fields [National Academies of Sciences, Engineering, and Medicine, 2017; Rodenbusch et al., 2016].

To be effective, UREs should be thoughtfully designed to meet the needs of students who may otherwise miss out on career opportunities tied to networking and community-building through such programs. Existing URE programs have followed a range of approaches, but traditionally, many have been centered around short-duration, time-intensive, individual, mentor-directed experiences, such as full-time summer internships in field or laboratory settings. However, these traits can inadvertently exclude some student populations, a concern that is leading many programs to modify their structure and design to engage broader groups.

To lower barriers to participation in UREs, we developed the Authentic Research through Collaborative Learning (ARC-Learn) program at Oregon State University (OSU). ARC-Learn, which ran from 2021 to 2024 and comprised two overlapping student cohorts, offered a long-term, low-intensity program focused on Arctic science and inclusive mentorship. It was designed to help students engage in a science community, foster identities as STEM professionals, and develop critical scientific and data literacy skills and 21st century competencies such as teamwork and communication.

Table 1. Design Features of ARC-LearnFeatureDescriptionDuration18 months (2 academic years)Intensity2–4 hours per weekLocationOn campus or remoteMentorshipMultiple mentors working in teams with multiple studentsTopic selectionStudent drivenStudent supportMentors, peers, program administrators, academic advisorsMentorship developmentInclusive mentorship training, facilitated peer learning communityResearch tasksDevelop research question, find data and analyze data, draw conclusions, and present findingsStudent developmentDiscover own strengths as researchers, work with a team, supplemental training in missing skills

ARC-Learn incorporated alternative design features (Table 1) to meet the needs of students who do not typically have access to time-intensive field or lab-based UREs, such as transfer students, remote students, and those with dependent care, military service, and other work commitments [Scott et al., 2019] or who have nontraditional enrollment patterns (e.g., dual enrollment in both university and community college, varying enrollment from term to term).

The program was framed in the context of Arctic science because of the region’s outsize effects on climate, ecosystems, and communities globally and to engage students with long-term research investments in polar regions [Marrongelle and Easterling, 2019]. The Arctic also offers a dynamic and interdisciplinary context for a URE program, enabling students to follow their interests in investigating complex science questions. In addition, numerous long-term Arctic monitoring programs offer rich datasets useful in all kinds of STEM careers.

Despite encountering challenges, the ARC-Learn model proved successful at engaging and motivating students and also adaptive as program organizers made adjustments from one cohort to the next in response to participant feedback.

The ARC-Learn Model

With support from mentors and peers, students experienced the whole research arc and gradually took ownership of their work.

Each ARC-Learn cohort lasted 2 academic years and included a dozen or more students. Participants received a stipend to offset costs associated with participation, such as childcare and missed work time, and had the option of obtaining a course credit each term to meet experiential learning requirements. With support from mentors and peers, they experienced the whole research arc and gradually took ownership of their work through three key phases of the program.

Early year 1: Build research teams. Some URE mentorship models involve a mentor primarily driving selection of a research topic and the student completing the work. In ARC-Learn, students learned from multiple mentors and peers, while mentors supported each other and received feedback from students (Figure 1). The students self-selected into research teams focused on a broad topic (e.g., marine heat waves or primary productivity), then developed individual research questions based on their strengths and interests.

Fig. 1. Some models of undergraduate research experiences have involved a mostly one-way transfer of knowledge from a single mentor to a single student, with the mentor deciding the research topic and the student completing the work. In ARC-Learn, students learned from multiple mentors and peers as part of small-group research teams, while mentors supported each other and received feedback from students.

Mentor-student teams met every other week—and students met one-on-one with mentors as needed—to support individual projects. The entire cohort also met twice a month to discuss topics including the fundamentals of Arctic science and the scientific process and to report out on progress toward milestones.

Late year 1 to middle of year 2: Develop research questions and find and analyze data. With no field or lab component to the program, ARC-Learn students worked exclusively with existing data. These data came from NASA and NOAA satellite-based sources such as the Moderate Resolution Imaging Spectroradiometer (MODIS), Advanced Very High Resolution Radiometer, and Soil Moisture Active Passive (SMAP) instruments; shipboard sources such as NOAA’s Distributed Biological Observatory, the Alaska Ocean Observing System, and the University-National Oceanographic Laboratory System’s Rolling Deck to Repository; and the National Science Foundation’s (NSF) Arctic Data Center and NOAA’s National Centers for Environmental Information.

Students often revised their research questions or the datasets they used multiple times to produce meaningful findings (Figure 2). Notably, access to these datasets proved critical to the educational experience of ARC-Learn students, highlighting the importance of maintaining them in public archives for future URE activities.

Fig. 2. ARC-Learn students developed their own research questions and worked exclusively with existing data to answer them. Students often revised their research questions or datasets multiple times to produce meaningful findings.

Many students struggled with finding, cleaning, analyzing, and interpreting data, often because of limited experience with tools such as geographic information system software and programming languages such as Python and R. At times, the required expertise was beyond even their mentors’ knowledge. Hands-on skill development workshops during cohort meetings connected students with additional mentors proficient in specific platforms and tools to help fill knowledge gaps and help students overcome obstacles.

Although the students encountered occasional setbacks, they reported that achievements such as settling on a final research question and creating rich data visualizations proved deeply rewarding and motivated further progress.

Late year 2: Share the results. Over several months, students created research posters with feedback and support along the way from their teammates, mentors, and the entire cohort. The program concluded with a grand finale, featuring on-campus gatherings for remote and in-person students, a dress rehearsal poster session, a celebratory dinner, and final presentations at a university-wide undergraduate research symposium.

Zoe’s Story

After a successful 7-year military career, Zoe enrolled at OSU to study the Arctic through her participation in ARC-Learn. As a student in cohort 2, she experienced several challenges along the research arc before finding success, and her experience helps illustrate the program’s model.

Zoe joined fellow students and mentors in the Marine Heatwaves research team and then narrowed her focus by exploring scientific literature and talking with her primary mentor to understand physical and chemical factors associated with marine heat waves as well as their effects on ocean ecosystems. She developed several research questions focused on how factors such as atmospheric pressure and temperature have affected the development and extent of marine heat waves off Alaska since 2010.

As Zoe and her mentor considered available datasets and relevant literature further, they realized that her questions were still too broad given the number of variables affecting ocean-atmosphere interactions. At one of the full-cohort meetings, she shared her difficulties and frustrations, prompting another mentor to offer their help. This mentor worked with Zoe to understand a key meteorological feature—the Aleutian Low—in the area she was studying, as well as relevant data available through the European Union’s Copernicus Climate Change Service [Hersbach et al., 2023] and the appropriate analysis platform.

“We jumped in and learned it together. She helped me find the right data, which in turn, allowed me to finalize my research question,” Zoe said.

Nuanced and iterative feedback from mentors and peers guided ARC-Learn participants, including Zoe, to design posters that balanced visual presentations of data alongside descriptive text to explain research findings. Credit: Ryan Brown

From that point, Zoe quickly landed on a focused question that she could address: Does a disruption in the Aleutian Low lead to marine heat waves over the North Pacific region? The final step was to develop a visually striking poster to invite attention, questions, and ideas during the research symposium.

“Seeing other people interested in my research…was validating of me as a scientist.”

Zoe’s experience at the poster session captured what we heard from many other students in the program. Even after her 2 years of being immersed in her project and working with mentors and peers, she said she felt imposter syndrome as a student trying to become a scientist and thought no one would care about her research.

“But people were really interested,” she said. “Seeing other people interested in my research, able to read and understand it on a poster, [and] ask me questions and suggest ideas was validating of me as a scientist.”

A Responsive Approach to URE Design

Through ARC-Learn, program leads sought to expand knowledge about the benefits and challenges of a long-duration, lower-intensity, team-based URE model. Because it was a design-based research program, mentor, student, and coordinator feedback was collected and continually used to make program adjustments [Preston et al., 2022, 2024].

Feedback was collected through pre-, mid-, and end-of-program surveys, as well as pre- and end-of-program interviews, and analyzed by a research and evaluation team. Findings were reported to the program leads, who also met regularly with external expert advisers to get additional recommendations for adjustments. By running two overlapping cohorts (the second started when the first was halfway completed), organizers could address issues that arose for the first cohort to improve the experience of the second one.

Lessons from ARC-Learn are documented in a practitioner guidebook, which discusses practical considerations for others interested in implementing alternative URE models [Brown et al., 2024]. In the guidebook, we examine each design component of ARC-Learn and offer recommendations for designing UREs that meet enrolled students’ specific learning needs and develop their science skills to meet relevant workforce demands.

Novel elements of the Authentic Research through Collaborative Learning (ARC-Learn) program were important in influencing participants’ persistence and success.

A few valuable lessons learned include the following.

Attrition. Expect high attrition rates in UREs designed for nontraditional students, and do not react by making drastic program changes that risk sacrificing otherwise successful program elements. We observed a 45% attrition rate in each cohort, which is indeed high but perhaps not surprising considering the population involved in the program—largely transfer students and those with caregiving or work responsibilities.

Most participants who left did so because of life crises or obligations that paused their research and educational goals. This observation embodies the complexity of students’ lives and reinforces the need for continually creative, flexible, inclusive program structures. For those who completed ARC-Learn, novel elements of the program (e.g., working in teams) were important in influencing their persistence and success.

Remote research applications. The first cohort started in 2021 entirely via remote instruction during the COVID-19 pandemic, before eventually transitioning to a hybrid approach as in-person instruction resumed. All ARC-Learn students in cohort 1 returned to campus, except one Ecampus student, who remained online. The program team and mentors struggled to balance the needs of the remote student, who eventually became somewhat detached from their research team.

As teamwork, camaraderie, and inclusivity are important qualities of the program, we decided for cohort 2 to recruit enough Ecampus students (plus two dedicated mentors) to form a research team of their own. The remote team was engaged and productive—meeting deadlines and producing high-quality work—highlighting the potential of all-remote URE models for students who might otherwise lack access to meaningful research opportunities.

Student-driven research. ARC-Learn empowered students to pursue their own research questions, fostering their autonomy and ownership of their work. However, the open-endedness of selecting their own research paths and the lack of guardrails proved challenging for participants.

We thus hired a program coordinator to provide one-on-one logistical support; establish clear expectations, timelines, and scaffolded assignments; and arrange workshops to teach programming and data analysis skills. This approach, as reported by students who worked with the coordinator, helped many program participants stay on track and ultimately complete their research project.

Mentor coordination. Enabling student success also meant supporting mentors. Organizers provided inclusive mentorship trainings and facilitated a peer learning community. They also made programmatic adjustments in response to experiences in the first cohort.

The student-driven nature of the research sometimes resulted in mismatches between student interests and mentor expertise in cohort 1. So in cohort 2, we engaged mentors earlier in the planning process to define thematic areas for the research teams, creating topics broad enough for students to find an area of interest but narrow enough for mentors to provide guidance. In addition, many mentors had field schedules typical of polar scientists, often resulting in weeks to months at sea. We purposefully paired mentors and asked about planned absences so we could fill any gaps with additional support.

Overall, students in cohort 2 reported feeling highly supported and valued by their mentors and that mentors created welcoming environments to ask questions and solve problems together.

A Foundation to Build On

Participants gained a deep understanding of the complexities and challenges of modern science as well as knowledge and skills needed in scientific education and careers.

From students’ feedback—and the research they did—it’s clear that participants who completed the ARC-Learn program gained a deep understanding of the complexities and challenges of modern science as well as knowledge and skills needed in scientific education and careers. The program thus highlights paths and lessons for others looking to develop successful alternatives to traditional UREs.

Many former ARC-Learn students are continuing to develop research skills, particularly in polar science, through internships and employment in field and lab research efforts. Zoe is working toward a bachelor’s degree in environmental sciences and exploring interests in environmental hazards, conservation, and restoration. For her, the program served as a foundation from which she is building a career and establishing confidence in herself as a scientist.

“I thought I’d have to play catch-up the whole time as an older, nontraditional student,” she said. But through the experience, “I realized I could start anywhere.”

Acknowledgments

ARC-Learn was a collaboration between OSU’s College of Earth, Ocean and Atmospheric Sciences and STEM Research Center. This work is supported by the U.S. NSF (award 2110854). Opinions, findings, conclusions, and recommendations in these materials are those of the authors and do not necessarily reflect the views of NSF.

References

Brown, R., et al. (2024), ARC-Learn Practitioner Guidebook: Practical considerations for implementing an alternative model of undergraduate research experience, Ore. State Univ., Corvallis, https://doi.org/10.5399/osu/1177.

Hersbach, H., et al. (2023), ERA5 monthly averaged data on single levels from 1940 to present, Copernicus Clim. Change Serv. Clim. Data Store, https://doi.org/10.24381/cds.f17050d7.

Marrongelle, K., and W. E. Easterling (2019), Support for engaging students and the public in polar research, Dear Colleague Letter prepared for the U.S. National Science Foundation, Alexandria, Va., www.nsf.gov/funding/opportunities/dcl-support-engaging-students-public-polar-research/nsf19-086.

National Academies of Sciences, Engineering, and Medicine (2017), Undergraduate Research Experiences for STEM Students: Successes, Challenges, and Opportunities, 278 pp., Natl. Acad. Press, Washington, D.C., https://doi.org/10.17226/24622.

Preston, K., J. Risien, and K. B. O’Connell (2022), Authentic Research through Collaborative Learning (ARC-Learn): Undergraduate research experiences in data rich Arctic science formative evaluation report, STEM Res. Cent., Ore. State Univ., Corvallis, https://doi.org/10.5399/osu/1156.

Preston, K., J. Risien, and N. Staus (2024), Authentic Research through Collaborative Learning (ARC-Learn): Undergraduate research experiences in data rich science summative evaluation report, STEM Res. Cent., Ore. State Univ., Corvallis, https://doi.org/10.5399/osu/1178.

Rodenbusch, S. E., et al. (2016), Early engagement in course-based research increases graduation rates and completion of science, engineering, and mathematics degrees, CBE Life Sci. Educ., 15(2), ar20, https://doi.org/10.1187/cbe.16-03-0117.

Scott, G. W., S. Humphries, and D. C. Henri (2019), Expectation, motivation, engagement and ownership: Using student reflections in the conative and affective domains to enhance residential field courses, J. Geogr. Higher Educ., 43(3), 280–298, https://doi.org/10.1080/03098265.2019.1608516.

Author Information

Ryan Brown (ryan.brown@oregonstate.edu), Laurie Juranek, and Miguel Goñi, College of Earth, Ocean and Atmospheric Sciences, Oregon State University, Corvallis; and Julie Risien and Kimberley Preston, STEM Research Center, Oregon State University, Corvallis

Citation: Brown, R., L. Juranek, M. Goñi, J. Risien, and K. Preston (2025), An accessible alternative for undergraduate research experiences, Eos, 106, https://doi.org/10.1029/2025EO250326. Published on 4 September 2025. Text © 2025. 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.

Cranfield University experts have developed a new method to precisely identify soil erosion hotspots along waterways, allowing for preemptive mitigation measures to be put in place that protect land and water systems.

A study by researchers at the University of Oxford, University of Leeds, and University College London has identified a new constraint on the chemistry of Earth's core, by showing how it was able to crystallize millions of years ago. The study is published in Nature Communications.