SPACE/COSMOS
UNM researchers analyze moon record, challenging Earth’s water origins
Moon’s 4-billion-year impact record suggests meteorites didn’t supply Earth’s water
University of New Mexico
image:
Tony Gargano, Ph.D., in the University of New Mexico's Center for Stable Isotopes.
view moreCredit: University of New Mexico
A long-standing idea in planetary science is that water-rich meteorites arriving late in Earth’s history could have delivered a major share of Earth’s water. A new study argues that the Moon’s surface record sets a hard limit on that possibility: even under generous assumptions, late meteorite delivery since about 4 billion years ago could only have supplied a small fraction of Earth’s water.
In a paper published in the Proceedings to the National Academy of Sciences titled Constraints on the impactor flux to the Earth–Moon system from oxygen isotopes of the lunar regolith, researchers led by Tony Gargano, Ph.D., at the Lunar and Planetary Institute and The University of New Mexico analyzed a large suite of Apollo lunar regolith samples using high-precision triple oxygen isotopes. Earth has erased most of its early bombardment record through tectonics, and constant crustal recycling. The Moon, by contrast, preserves a continuously accessible archive: lunar regolith, the loose layer of debris produced and reworked by impacts over billions of years.
Ever since the Apollo missions, scientists have tried to read that archive using elements that concentrate in impactors - especially ‘metal-loving’ siderophile elements, which are abundant in meteorites but scarce in the Moon’s silicate crust. But regolith is an especially challenging mixture: impacts can melt, vaporize, and rework material repeatedly, and post-impact geological processes can separate metal from silicate, complicating attempts to reconstruct the type and amount of impactor material.
“The lunar regolith, which is a collection of loose ‘soil’ and broken rock at the surface, acts like a long-term mixing layer,” said Gargano. “It captures impact debris, stirs it in, and preserves those additions for immense spans of time. That is why it is such a powerful archive. It lets us study a time-averaged record of what was hitting the Earth–Moon system.”
The new study takes a different approach. Instead of relying on metal-loving tracers, it uses oxygen - the dominant element by mass in rocks - and its triple-isotope “fingerprint” to separate two competing signals that normally get tangled in lunar regolith: (1) the addition of meteorite material and (2) isotopic effects from impact-driven vaporization. From measuring offsets in the oxygen isotope composition of regolith, the team finds that at least ~1% by mass of the regolith reservoir consists of impactor-derived material that are best explained from carbon-rich meteorites that were partially vaporized upon impact.
“Triple oxygen isotopes give us a more direct and quantitative way to approach the problem. Oxygen is the dominant element in most rocks, and the triple-isotope framework helps us distinguish true mixing between different reservoirs from the isotopic effects of impact-driven vaporization,” said Gargano. “In practice, that lets us isolate an impactor fingerprint from a regolith that has a complicated history, with fewer assumptions and a clearer chain from measurement to interpretation.”
The team translated these impactor fractions into water-delivery bounds for the Moon and Earth, expressed in Earth-ocean equivalents for scale. For the Moon, the implied delivery since ~ 4 billion years ago is tiny on an Earth-ocean scale. But tiny compared to Earth’s oceans does not mean unimportant for the Moon. Instead, the Moon’s accessible water inventory is concentrated in small, cold-trapped reservoirs, and water is the kind of resource that matters immediately for sustained human presence for important things like life support, radiation shielding, and fuel. In other words, the long-term trickle of impactor-derived water can be negligible for Earth yet still be a meaningful contributor to the Moon’s available water budget.
The researchers then extended the same accounting to Earth, using a commonly applied scaling in which Earth receives substantially more impactor material than the Moon. Even if Earth experienced roughly 20× the impactor flux and even adopting the extreme megaregolith end-member, the cumulative water delivers only a few percent of an Earth Ocean at most. That makes it difficult to reconcile the late-delivery of water-rich meteorites as the dominant source of Earth’s water, given that independent estimates yield several ocean-mass equivalents of water in the Earth in total.
“The lunar regolith is one of the rare places we can still interpret a time-integrated record of what was hitting Earth’s neighborhood for billions of years,” said Gargano. “The oxygen-isotope fingerprint lets us pull an impactor signal out of a mixture that’s been melted, vaporized, and reworked countless times. The main takeaway from our study is that Earth’s water budget is hard, if not impossible, to explain if we only consider a single, late delivery pathway from water-rich impactors from the outer solar system. Even though some meteorite types carry a lot of water, their broader chemical and isotopic fingerprints are quite exotic relative to Earth. Habitability models have to satisfy such empirical constraints, and our study adds a constraint that future theories will need to reproduce.”
“Our results don’t say meteorites delivered no water,” added Simon. “They say the Moon’s long-term record makes it very hard for late meteorite delivery to be the dominant source of Earth’s oceans.”
Gargano framed the work as part of a scientific lineage that began with Apollo. “I’m part of the next generation of Apollo scientists - people who didn’t fly the missions, but who were trained on the samples and the questions Apollo made possible,” Gargano said. “The value of the Moon is that it gives us ground truth: real material we can measure in the lab and use to anchor what we infer from meteorites and telescopes.
“Apollo samples are the reference point for comparing the Moon to the broader Solar System,” Gargano added. “When we put lunar soils and meteorites on the same oxygen-isotope scale, we’re testing ideas about what kinds of bodies were supplying water to the inner Solar System. That’s ultimately a question about why Earth became habitable, and how the ingredients for life were assembled here in the first place.”
Apollo samples also matter because the Moon preserves that impact story across deep time in a way Earth does not. The Moon does not just tell us about the Moon. It preserves an accessible record of the impact environment of the inner solar system, which helped set the boundary conditions under which Earth became habitable. There is still real wonder in that. Scientists have rocks collected decades ago, from another world, and they are still capable of changing how we think about the origin of Earth’s water and the conditions that made life possible.
“What modern techniques add to this amazing legacy of scientific exploration is precision and interpretive power. We can now resolve subtle isotopic signals that allow quantitative tests of formation and habitability models,” said Gargano. “That is why Apollo science keeps evolving. The samples are the same, but our ability to interrogate them, and the questions we can ask of them, are fundamentally better.”
In addition to his research findings, Gargano is equally proud of what scientists are doing in terms of training and outreach because it captures that same arc: taking something that feels distant and making it tangible and impactful to our lives.
“At UNM, I have been training Albuquerque high schoolers in planetary science and geochemistry, including senior Brooklyn Bird and junior Violet Delu from the Bosque School,” said Gargano. “These students are getting hands-on training in geochemistry using UNM’s unique collection of Astromaterials, and they are learning the physical craft of laboratory science: how to prepare and handle samples, how to make high-quality measurements, and how to think clearly about uncertainty and reproducibility.
“But the deeper lesson is the transformation that happens when a student realizes they can hold a piece of another world, make a measurement, and pull meaning out of it. They learn how a chemical signal becomes a geologic story, and how that story scales up into an explanation for how a planetary body evolved to become the way it is. Experiences like that change what students think is possible for themselves. They build confidence, technical ability, and a sense of belonging in a field that can otherwise feel out of reach.”
Bird and Delu will both be presenting their independent research projects at the 57th Lunar and Planetary Science Conference this spring and will also be educators to their peers and younger students through Bosque School outreach events. This is a model Gargano is excited to carry forward to other places in the country, so that more underserved students can gain access to world-class research experiences and obtain skill sets in geochemistry that open doors for them internationally.
Image from Apollo 17 mission.
Credit
NASA
UNM's Institute of Meteoritics Meteorite Museum.
Credit
University of New Mexico
Journal
Proceedings of the National Academy of Sciences
Method of Research
Data/statistical analysis
Subject of Research
Not applicable
Article Title
Constraints on the impactor flux to the Earth–Moon system from oxygen isotopes of the lunar regolith
Article Publication Date
20-Jan-2026
The hidden microbial communities that
UCD Research & Innovation
Microorganisms live in biofilms - the equivalent of microbial “cities”- everywhere on Earth. These city-like structures protect and house microbial communities and play essential roles in enabling human and plant health on our planet. Now, a new Perspective article published in npj Biofilms and Microbiomes sets out a path to uncover the role of biofilms in health during long-duration spaceflight, and how spaceflight research can reshape our understanding of these microbial communities on Earth.
Led by researchers at the University of Glasgow in Scotland and Maynooth University and University College Dublin (UCD) in Ireland working within the GeneLab Microbes Analysis Working Group around the NASA Open Science Data Repository, the new study Biofilms: from the cradle of life to life support – 22nd Jan 2026 explores biofilms as a major frontier for both human and crop health in space and on Earth.
Biofilms are organised microbial communities structured within a matrix of microbial polymers that defines how microbes interact with hosts. On Earth, these host-biofilm interactions underlie essential functions across human and plant tissues, including nutrient uptake and use, stress tolerance and pathogen control. In space, evidence suggests these ancient interactions can be compromised, and require coordinated, mechanistic study.
“Biofilms are often considered from an infection viewpoint and treated as a problem to eliminate, but in reality they are the prevailing microbial lifestyle that supports healthy biological systems,” said Dr Katherine J. Baxter from the University of Glasgow, first-author and Co-ordinator of the UK Space Life and Biomedical Sciences Association (UK Space LABS). “Spaceflight offers a distinctive and invaluable testbed for biofilm organisation and function, and, importantly, evidence so far makes it clear that biofilms need to be better understood, managed, and likely engineered to safeguard health during spaceflight.”
Spaceflight and even spaceflight simulations on earth can alter biofilm architecture, gene regulation, signalling, and stress tolerance, with effects varying across microbial species and experimental platforms. The team outlines a roadmap for applying advanced genetics and biochemical approaches, or “multiomics”, that can uncover biofilm structure and functions across interkingdom multispecies microbial communities interacting within high complexity biological systems.
“Plants will sit at the centre of long-duration spaceflight missions, and plant performance depends on biofilm interactions in and around plant root systems,” said Dr Eszter Sas, co-author and metabolomics specialist at Maynooth University. “By combining multispecies genetics and biochemistry, modern multiomics has the exciting capability to reveal new biofilm mechanisms from spaceflight responses, and is starting to fill in major gaps in our understanding of signalling and metabolism at the interface of biofilms and plant roots.”
The research was coordinated through the ecosystem of open access data, tools, platforms, and Analysis Working Groups around the NASA Open Science Data Repository, which was an expansion of NASA GeneLab. Experimentation in space is incredibly challenging and costly, so ‘Open Science’ approaches and communities are needed for shared standards, methodology and transparent analysis to ensure what is learnt from each spaceflight mission is maximised and so that discoveries are translated to applications on Earth.
“This work reflects collaboration spanning the globe, with a strong Open Science community for shared thinking and shared discovery,” said Prof Nicholas J. B. Brereton, senior author at University College Dublin. “The translation of value runs both ways, spaceflight can reveal new biology under unfamiliar stress, and those insights can tell us a lot about how life might survive in space but also inform approaches for health and agriculture on Earth.”
The research includes a call for action on coordinated open biofilm research that moves beyond narrow model systems to support analogue and cross-mission experimentation that accelerates the path from observation to useful intervention.
Read the Perspective article in npj Biofilms and Microbiomes: https://www.nature.com/articles/s41522-025-00875-8. DOI: 10.1038/s41522-025-00875-8.
Journal
npj Biofilms and Microbiomes
Method of Research
Commentary/editorial
Subject of Research
Not applicable
Article Title
Biofilms: from the cradle of life to life support
Article Publication Date
22-Jan-2026
Rethinking where life could exist beyond earth
The Hebrew University of Jerusalem
image:
The traditional habitable zone is shown by the diagonal orange stripe. Its distance from the host star (horizontal axis, in Astronomical Units) increases with luminosity, which increases with the mass of the star and the stellar type (M, K, G, etc.) shown on the vertical axis. The ellipses represent the extensions of the habitable zone presented in the research.
view moreCredit: Amri Wandel
Astronomers have long searched for life within a rather narrow ring around a star, the “habitable zone,” where a planet should be neither too hot nor too cold for liquid water. A new study argues that this ring is too strict: on tidally locked worlds that keep one face in daylight and the other in permanent night, heat may still circulate enough for liquid water to persist on the dark side, even when the planet orbits closer to cool M- and K-dwarf stars than conservative climate models allow. The study also points outward: liquid water could exist far beyond the classical outer edge, hidden beneath ice as subglacial or intraglacial lakes, meaning the number of worlds worth checking for water, and potentially life-friendly conditions, may be much larger than the traditional map suggests.
For years, astronomers have relied on a simple rule of thumb when searching for life beyond Earth: look for planets in the “habitable zone,” the narrow range around a star where liquid water can exist on a planet’s surface. In our own solar system, that zone lies roughly between the orbits of Earth and Mars.
But many of the planets now being discovered do not fit neatly into this framework, orbiting stars quite different than our sun, at distances closer than the inner edge of the habitable zone or further out.
In a new study published in The Astrophysical Journal astrophysicist Prof. Amri Wandel from the Hebrew University asks what happens when scientists break the assumptions built into traditional habitability models.
The focus is on tidally locked exoplanets, worlds that always face their star with the same hemisphere. These planets experience permanent daylight on one side and permanent night on the other, a configuration often considered to challenge surface liquid water and life.
Wandel’s analysis suggests otherwise.
Using an analytical climate model that tracks temperature across the surface of such planets, the study shows that worlds orbiting M- and K-dwarf stars could sustain liquid water on their night side, even when they orbit significantly closer to their star than classical habitable-zone models allow. Heat transported from the day side can keep parts of the night side above freezing, expanding the range of environments where water may persist.
This extended definition of the habitable zone may help explain recent observations by the James Webb Space Telescope, which detected water vapor and other volatile gases in the atmospheres of warm, close-in Super-Earths orbiting M-dwarf stars—planets previously thought to lie outside the safe range for surface water.
The study also looks in the opposite direction, beyond the outer edge of the habitable zone. Even on cold planets far from their stars, liquid water could exist beneath thick ice layers, in the form of intraglacial lakes or subglacial melting, further widening the habitable zone and enhancing the number of worlds that may support water-based environments by a large factor.
By revisiting the assumptions behind the habitable zone and recalculating its boundaries, this research reframes where astronomers might look for conditions suitable for life, suggesting that potential habitats may exist on planets once ruled out.
Journal
The Astrophysical Journal
Method of Research
Computational simulation/modeling
Subject of Research
Not applicable
Article Title
Exoplanets beyond the Conservative Habitable Zone. I. Habitability
Article Publication Date
22-Jan-2026
Fast AI for satellite learns to quantify uncertainty
Journal of Remote Sensing
image:
Validation of the probabilistic machine learning framework using the SpT model against the OCO-2 Level 2 data product. (A) Scatter density plots comparing XCO2 values and associated uncertainties from the SpT model and OCO-2 for each year from 2021 to 2024. (B) Scatterplot illustrating the agreement between uncertainty estimates from the SpT model and OCO-2 across the full time series (2021–2024). (C) Case studies for Beijing, Tokyo, and Seoul, with the light blue shaded area indicating the 3 uncertainty bounds from the SpT model. C1-C2 shows Beijing, 2021 October 11 (C1) and 2024 January 28 (C2). C3-C4 shows Tokyo, 2022 September 26 (C3) and 2024 February 13 (C4). C5-C6 shows Seoul 2021 October 01 (C5) and 2024 February 12 (C6).
view moreCredit: Journal of Remote Sensing
Accurately tracking atmospheric greenhouse gases requires not only fast predictions but also reliable estimates of uncertainty. Researchers have developed a lightweight machine learning framework that enables satellite-based trace gas retrievals to quantify their own uncertainty while maintaining high computational efficiency. By combining probabilistic modeling with ensemble learning, the approach delivers rapid and trustworthy CO₂ measurements, helping bridge the gap between data-driven speed and scientific reliability.
Satellite missions generate massive volumes of atmospheric data that are essential for monitoring greenhouse gases and informing climate policy. Traditional physics-based retrieval algorithms provide reliable uncertainty estimates but are computationally intensive and struggle to scale with growing data streams. In contrast, machine learning methods offer dramatic speed improvements but typically produce only single-value predictions, lacking uncertainty information needed for scientific interpretation and decision-making. Existing probabilistic machine learning approaches often require heavy computation, complex tuning, or labeled uncertainty data, limiting their operational use. Based on these challenges, there is a strong need to develop scalable retrieval methods that combine machine learning efficiency with rigorous uncertainty quantification.
A research team from Shanghai Jiao Tong University reported this advance in Journal of Remote Sensing, published (DOI: 10.34133/remotesensing.0881) on December 26, 2025. The study introduces a probabilistic machine learning framework designed for satellite-based trace gas retrievals, with a focus on carbon dioxide monitoring. The work addresses a critical bottleneck in current satellite data processing: how to rapidly analyze vast datasets while still providing uncertainty estimates required for climate science, data assimilation, and policy-relevant applications. The framework was validated using long-term observations from NASA’s Orbiting Carbon Observatory-2 (OCO-2) mission.
The study demonstrates that uncertainty-aware machine learning can be achieved without sacrificing speed. By modifying neural networks to predict both expected values and associated uncertainty, the framework transforms standard deterministic models into probabilistic ones. When applied to satellite CO₂ retrievals, the method produced highly accurate concentration estimates while simultaneously quantifying uncertainty. Validation against OCO-2 operational products showed strong temporal and spatial consistency, with over 99% of reference values falling within the predicted uncertainty bounds. Importantly, the probabilistic model matched the accuracy of physics-based methods while operating thousands of times faster. This combination of speed, accuracy, and reliability marks a significant advance over existing machine learning approaches in atmospheric remote sensing.
The framework integrates two key innovations: likelihood-based learning and snapshot ensemble modeling. Instead of predicting a single output, the neural network simultaneously estimates the mean and variance of each retrieval, enabling direct uncertainty learning from existing satellite products. A Gaussian negative log-likelihood loss function ensures that both overconfident and underconfident predictions are penalized, encouraging well-calibrated uncertainty estimates.
To capture model-related uncertainty efficiently, the researchers employed snapshot ensembles, which extract multiple model instances from a single training run using cyclical learning rate scheduling. This avoids the heavy computational cost of training many independent models.
When tested on OCO-2 data from 2017 to 2024, the probabilistic model achieved retrieval speeds of milliseconds per satellite sounding, compared with minutes for traditional algorithms. Case studies over major cities showed that predicted uncertainty patterns closely followed those from physics-based retrievals, while remaining slightly more conservative—reflecting both measurement noise and model uncertainty.
“Uncertainty is not a luxury—it’s essential for trustworthy climate data,” said a member of the research team. “Our goal was to keep the speed advantages of machine learning while restoring the uncertainty information that scientists and policymakers rely on.This framework shows that we don’t need complex or computationally expensive solutions to achieve reliable probabilistic predictions at scale.”
The researchers adapted an existing Transformer-based neural network used for CO₂ retrievals by adding outputs for uncertainty estimation. Training data consisted of OCO-2 spectral measurements and corresponding retrieval products. The model was trained using a Gaussian likelihood loss and cyclical learning rate scheduling, with multiple snapshots collected to form an ensemble. Performance was evaluated using independent multi-year satellite observations, statistical correlation analysis, and regional case studies over East Asia.
Beyond carbon dioxide, the framework can be applied to other satellite-based retrieval tasks, including methane monitoring and aerosol profiling. Its lightweight design makes it particularly well suited for next-generation Earth observation missions, where data volumes will continue to grow rapidly. By making uncertainty-aware machine learning practical at scale, the approach could improve climate monitoring, air quality assessment, and data-driven environmental decision-making worldwide.
###
References
DOI
Original Source URL
https://spj.science.org/doi/10.34133/remotesensing.0881
Funding Information
This work is supported by the National Natural Science Foundation of China (grants nos. 52276077 and 52120105009).
About Journal of Remote Sensing
The Journal of Remote Sensing, an online-only Open Access journal published in association with AIR-CAS, promotes the theory, science, and technology of remote sensing, as well as interdisciplinary research within earth and information science.
Journal
Journal of Remote Sensing
Subject of Research
Not applicable
Article Title
From Deterministic to Probabilistic: A Lightweight Framework for Probabilistic Machine Learning in Trace Gas Remote Sensing
No comments:
Post a Comment