Frost is one of the most punishing adversaries of tea cultivation. A single cold snap can scorch tender leaves, derail an entire season’s harvest, and degrade the delicate chemistry that gives tea its value. For growers and breeders, however, judging how badly a field has been damaged has long been a stubbornly subjective exercise: visual ratings lump continuous injury into crude categories, and laboratory assays destroy the very tissue they measure. A new study published in Plant Methods by Zhengtong He of Shihezi University, Zhiwei Chen of the Tea Research Institute of Shandong Academy of Agricultural Sciences, and colleagues offers a way out, pairing hyperspectral imaging with a purpose-built deep learning model to read frost damage directly from the way leaves reflect light.
The core problem the team confronted is that frost injury is not a binary event but a continuum. Leaves exposed to sub-zero temperatures suffer a cascade of overlapping stresses: photosynthetic pigments degrade, the machinery of photosynthesis falters, and cell membranes and walls break down. Any single measurement captures only a sliver of that cascade. To capture the whole picture, the researchers collected hyperspectral data together with eight key physicochemical indicators from tea plants growing across natural frost damage gradients in the field, rather than in the artificial conditions of a growth chamber. From these multidimensional physiological responses they constructed a composite measure, the tea cold stress response index, or TCSRI, which expresses frost damage severity as a continuous value instead of a coarse visual class.
Building TCSRI is a meaningful advance in itself. Traditional frost assessment often relies on ordinal scales, where an observer assigns a score of mild, moderate, or severe. Such scales throw away information and make it difficult to track recovery or compare cultivars with fine precision. By integrating several physiological dimensions into one index, the team created a target variable that reflects the plant’s actual internal state. The challenge then became predicting that index, and the individual chemical indicators behind it, from spectral data alone, without touching or destroying a single leaf.
Hyperspectral sensing is well suited to this task in principle. Across hundreds of narrow wavelength bands, plant leaves encode a wealth of information: chlorophyll absorption in the visible range, water and structural features in the near infrared. But the data are high-dimensional, with far more spectral channels than samples, a regime that invites overfitting, especially in field studies where collecting ground-truth measurements is laborious. Conventional chemometric approaches, from partial least squares regression to standard machine learning pipelines, often struggle to separate genuine physiological signals from noise, sensor drift, and the confounding effects of leaf surface properties and illumination variation.
The researchers’ answer is a model they call STAMPformer, short for spectral-prior-guided multi-view attention pooling. Its architecture reflects a deliberate design philosophy: rather than letting a neural network rediscover everything from scratch, the model is steered by what spectroscopists already know about how plants respond to stress. At its front end, a parameter-free attention convolution extraction module jointly models both the raw reflectance spectra and their first-order derivatives. Derivative spectra are a classic spectral prior: they accentuate subtle local features such as absorption peaks and shoulders that are easily drowned out in raw reflectance, while suppressing broad baseline drift and noise. By processing both views together, the module sharpens the local spectral signatures that carry physiological meaning.
On top of this extraction stage sits the second innovation, a multi-view attention pooling module. Deep networks processing spectral sequences produce long streams of features, but not all wavelengths matter equally for a given prediction. The pooling module adaptively aggregates sequence features drawn from different feature subspaces, learning to amplify the responses that are most diagnostic of cold stress and to down-weight the rest. This attention mechanism acts like a learned spotlight, and because it operates across multiple subspaces, it can capture complementary patterns that a single attention head might miss. The result is a compact representation tailored to the prediction task at hand, whether that is estimating a specific chemical indicator or the composite TCSRI.
The performance numbers are striking. Across all nine prediction tasks, spanning the eight physicochemical indicators and the TCSRI itself, STAMPformer achieved coefficients of determination above 0.87 and ratios of performance to deviation above 2.80, thresholds generally regarded as indicating reliable quantitative prediction. In other words, from a leaf’s reflectance fingerprint alone, the model could reconstruct the plant’s internal frost-stress state with an accuracy that approaches laboratory measurement, while remaining entirely non-destructive. The team did not stop at headline metrics. Ablation experiments systematically removed each proposed module to confirm that both the attention convolution extraction and the multi-view pooling contributed measurably to the gains, and paired statistical tests across multiple random seeds ensured the improvements were not artifacts of a lucky initialization.
Interpretability was a further priority, and here the study delivers something rarer than raw accuracy. Applying SHAP analysis, a technique that attributes each prediction to the contributions of individual input features, the researchers identified the wavelengths the model relied on most heavily. These key spectral regions correspond to physiologically meaningful processes: pigment degradation, photosynthetic impairment, and the disruption of cellular structure. That alignment between machine attention and plant biology matters on two fronts. It gives spectroscopists confidence that the model has learned genuine stress signatures rather than spurious correlations, and it hands plant physiologists a data-driven map of which parts of the spectrum track which facets of cold injury, potentially guiding future sensor design and breeding assays.
Crucially, the model also proved itself on data it had never seen. Independent external validation, using samples not involved in training, demonstrated that STAMPformer’s predictive ability generalizes beyond its original dataset, addressing one of the most common failure modes of machine learning in agricultural sensing. For a field where models are often tuned and tested on the same plots, this external check is an important credibility marker, suggesting the framework could be deployed across different farms, seasons, and frost events rather than remaining a laboratory curiosity.
The practical implications reach well beyond tea. An accurate, non-destructive, continuous measure of cold damage opens the door to rapid post-frost field surveys, letting growers quantify losses within hours instead of waiting for symptoms to develop or sending samples to a lab. For breeders screening for cold-tolerant germplasm, a spectral readout of TCSRI could accelerate the evaluation of hundreds of candidate lines without destructive sampling. And in the emerging vision of the smart tea garden, where drones and handheld spectrometers roam the rows, a model like STAMPformer could feed real-time frost maps into irrigation, shading, and recovery-management decisions. The authors frame their work as supporting frost monitoring, cold-tolerant germplasm evaluation, and intelligent garden management, and the architecture’s reliance on general spectral priors rather than tea-specific tricks suggests the approach could be adapted to other crops facing abiotic stress. As climate variability makes late-season frosts increasingly erratic, tools that turn light reflected from a leaf into a precise physiological diagnosis may become as essential to the tea garden as the thermometer.
Subject of Research: Hyperspectral deep learning prediction of continuous frost damage severity in tea plants
Article Title: STAMPformer: spectral-prior-guided multi-view attention learning for hyperspectral prediction of continuous frost damage severity in field-grown tea plants
Article References: He, Z., Chen, Z., Ding, Z., Wang, Z., Zhang, F., Zhao, Y., & Dong, C. (2026). STAMPformer: spectral-prior-guided multi-view attention learning for hyperspectral prediction of continuous frost damage severity in field-grown tea plants. Plant Methods. https://doi.org/10.1186/s13007-026-01597-4
Image Credits: AI Generated
DOI: 10.1186/s13007-026-01597-4
Keywords: tea plant, frost damage, hyperspectral imaging, deep learning, attention mechanism, cold stress index, non-destructive prediction, SHAP analysis, precision agriculture, spectral priors, plant physiology, smart farming
New research led by University of the Witwatersrand and University of Southern California paleoanthropologists sheds more light on one of the most debated transitions in human evolution: when our ancestors stopped regularly climbing trees and committed fully to walking on two legs.
The post Fossil Evidence Suggests Early Homo Had Abandoned Tree-Climbing by 1.8 Million Years Ago appeared first on Sci.News: Breaking Science News.
© Yonhap News Agency, via Reuters
When two black holes merge, the resulting object rings like a struck bell, radiating gravitational waves that fade in a distinctive pattern called ringdown.
The post Physicists Devise New Way to Listen for ‘Hairy’ Black Holes appeared first on Sci.News: Breaking Science News.
It's some striking military technology.Prehistoric art hidden in caves throughout southwestern Ireland for thousands of years could push back the timeline on the earliest human presence on the Emerald Isle by as much as 5,000 years, archaeologists say.
The remarkable discoveries may reveal a previously unknown chapter in the ancient history of human migrations across the British Isles during the Ice Age, and how they once traversed a portion of the landscape that is now sunken beneath the Irish Sea.
Reported by a team of researchers with the international Lost Land Bridge Project, the discoveries involve prehistoric markings located in eight caves in Ireland that could potentially date as far back as 15,000 years ago.
Some are engravings into the cave’s stone walls, while others are markings made with ochre and charcoal, and include lines traced along cave surfaces that appear to depict ancient animals that include deer, and extinct species such as mammoths.
If confirmed, Ireland’s long-hidden ancient artwork could dramatically shift past thinking about the earliest evidence of human arrivals on the island, which was previously believed to have occurred around 10,500 years ago.
If humans arrived in Ireland this early, it also raises an intriguing possibility: that they might have done so by traversing terrain that would only have been exposed amid the lower sea levels that prevailed toward the end of the last Ice Age.
This effect was due to the amount of water that was held within the vast glaciers near Earth’s poles, which, in addition to lowering sea levels, helped to reshape the geography of northwestern Europe during the Pleistocene.
Because of this, certain portions of the ancient landscape that are now submerged beneath ocean waters would have been exposed. Among these, one that stretched from modern Pembrokeshire in Wales and southeastern Ireland near Rosslare might have been used by ancient travelers who could have made their way to Ireland, leaving a record of their Ice Age journey on the walls of caves they used as shelter, according to Dr. George Nash of the University of Liverpool, who coordinated the project.
Nash believes this ancient crossing could have comprised a 46-mile-long area, where ancient nomadic people following natural subsistence patterns would have encountered vast expanses of grassland and vegetation, along with ample resources that included horses, bison, reindeer, mammoths, and other animals that these ancient hunter-gatherers were skilled at hunting.
Evidence from throughout Europe and other parts of the ancient world already shows that ancient hunter-gatherers were well-accustomed to following migrating animals and seasonal resources, and that the landscape could have provided a natural route westward.
Nash recently said that until now, archaeologists had relied mainly on accidental discoveries that suggested a human presence in Ireland by around 10,500 years ago. However, a growing number of archaeologists had begun to doubt that this was the period when humans initially arrived.
“We just weren’t looking hard enough, and not in the right places,” Nash said.
To determine locations where reliable evidence of an earlier human presence might be found, the researchers examined several features associated with prehistoric caves in southern Wales, where documented discoveries linked them to an early human presence.
Based on these characteristics, the team then relied on computer modeling to help them identify Irish locations that had similar features.
The team was assisted by Shane Diffily, an expert in caving, and digital mapping consultant Darren Lee, which allowed them to quickly narrow their search to just under 30 potentially promising caves throughout southwestern Ireland, according to Heritage Daily.
Of those locations, eight were found to contain what appeared to be various kinds of prehistoric rock art. The exact locations of those caves have not been released by the team, to protect the sites from vandalism, or potential looting—a major issue that has long complicated archaeological investigations.
Along with direct engravings into limestone surfaces and markings made with ochre and charcoal, researchers also discovered a few instances of what appear to be “finger flutings,” where soft surfaces such as clay or sediment were marked by ancient artists who dragged their fingers along such surfaces.
One of the most significant discoveries uncovered by Nash and the team appears to be a portion of a deer, and an image within one of the caves that appears to represent a mammoth. Examples like these could potentially help provide connections to similar examples known from prehistoric rock art that has been documented in Wales.
From those sites, some of the same prehistoric artists may have made their way across the ancient land bridge that was once available, leaving their markings at similar sites that once offered shelter as they arrived in Ireland.
The initial discoveries at the newly identified eight cave locations may represent only a small portion of the long-overlooked archaeological evidence throughout Ireland, which may document its earliest arrivals. Future investigations will likely focus on sites located near Wexford, Waterford, Cork, and other areas along routes that people may have followed as they made their way to southwestern Ireland.
As investigations continue, Nash and the research team say that reliable dating of the art will be an important next step toward determining how early these prehistoric artists might have arrived.
Once that testing is completed, it will help determine whether the timescales on the arrival of Ireland’s earliest human inhabitants will indeed be pushed back by thousands of years. If so, the newly uncovered cave art could offer an unprecedented look at the lives of those ancient hunter-gatherers who once traversed a landscape before it sank beneath rising ocean waters as Ice Age glaciers retreated.
Micah Hanks is the Editor-in-Chief and Co-Founder of The Debrief. A longtime reporter on science, defense, and technology with a focus on space and astronomy, he can be reached at micah@thedebrief.org. Follow him on X @MicahHanks, and at micahhanks.com.
© Jason Gulley for The New York Times
A recent analysis of bones found in two separate underwater cave systems in South Australia has provided researchers with a new framework for investigating how animal remains, including those of extinct megafauna, came to rest in these submerged environments.
Well-preserved bones have often been found in submerged caves, but interpreting how they accumulated and what happened to them afterward has been difficult. Researchers have had limited tools for determining the environmental processes that alter remains over time.
The new study, led by Griffith University PhD candidate Meg Walker and Professor Julien Louys of the Australian Research Centre for Human Evolution, identifies physical, chemical, and molecular markers that can provide clues about the environmental conditions each specimen experienced. The findings were published in PLOS One.
With help from specialist cave divers from the Cave Divers Association of Australia, Walker’s team examined animal bones recovered from Green Waterhole and Gouldens Sinkhole, two underwater cave systems near Mount Gambier, South Australia. Gouldens Sinkhole reaches 63 meters deep through a nearly 30-meter-wide shaft.
The researchers used radiocarbon dating methods on 41 specimens and examined them at various scales, starting with a visual evaluation of spatial distribution and surface wear before progressing to portable X-ray fluorescence, scanning electron microscopy, and protein analysis. Using these techniques, the researchers identified chemical residue, signs of bacterial damage, and traces of ancient DNA. The group also identified changes associated with different preservation environments by comparing submerged bones with those recovered from nearby dry caves.
Light played an important role in some of the most noticeable changes to the bones. Near the cave entrances, where sunlight was available, algae and aquatic plants grew on the bone surfaces and left black biological stains, while cyanobacteria produced circular etchings and tunnels on the outer surfaces. Deeper within the caves, in areas where sunlight did not reach, the bones mostly stayed the same and retained their original characteristics.
This distinction is important because it shows that staining alone does not indicate that a bone was buried in sediment or fully submerged, as previously assumed. Instead, it shows that the bone was simply exposed to light.
Bones from the dry sections of the cave systems showed different patterns. In the absence of water and algae, land-based bacteria penetrated the bone from within, while plant roots created long grooves on the surface as they grew in search of nutrients. These features are distinct from those observed in submerged specimens.
The bones in the team’s study were not from ancient megafauna, belonging instead to cattle, kangaroos, emus, sheep, pigs, dingoes, rabbits, possums, quolls, and swamp rats. European settlers likely deposited the specimens in the caves after they arrived in the Mount Gambier region in the 1840s. DNA tests showed that one bone came from a Merino sheep.
The ultimate goal is to apply this framework to older material. Caves in the Mount Gambier region, such as Tank Cave and Englebrechts Cave, have yielded remains of Australia’s Pleistocene megafauna, including Diprotodon, a large marsupial similar in size to a rhinoceros, Thylacoleo carnifex, the largest known marsupial predator, and short-faced kangaroos that reached about 2 meters in height. Some of these fossils have remained undisturbed for more than 60,000 years.
Using Walker’s framework when looking at these specimens might help researchers understand how Australia’s megafauna fossils accumulated and could also provide a method for investigating factors such as human activity and environmental change that caused their extinction.
“This study has delivered the first framework for interpreting how megafauna fossils formed, survived, and changed in underwater caves,” Walker said. “It will provide archaeologists and paleontologists worldwide with a powerful new tool for reconstructing past environments and histories in these challenging conditions.”
Austin Burgess is a writer and researcher with a background in sales, marketing, and data analytics. He holds an MBA, a Bachelor of Science in Business Administration, and a data analytics certification. His work focuses on breaking scientific developments, with an emphasis on emerging biology, cognitive neuroscience, and archaeological discoveries.
Some 1.3 billion years ago, when multicellular organisms were just emerging on Earth and life was still confined to the ocean, two behemoth black holes smashed together in a far-flung galaxy. The collision was so powerful that it jolted the universe, creating a ripple in the fabric of space-time that traveled across the cosmos atContinue reading "How gravitational waves transformed our universe"
The post How gravitational waves transformed our universe appeared first on Astronomy Magazine.
Scientists have discovered China’s first known freshwater cave snail, an eyeless species so rare that only three living adults have been found. Deep inside a mountain cave in southwestern China, scientists found an animal that appears built for a world where sunlight never reaches. The tiny freshwater snail has no eyes, almost no color, and [...]The Met Office has confirmed that its June temperature 'record' of 37.7°C came from a junk site. But rather than dropping it, it has upped it to 38°C! This not a serious scientific organisation, says Paul Homewood.
The post The Met Office is No Longer a Serious Scientific Organisation appeared first on The Daily Sceptic.
Se você gosta de acompanhar o mundo da aviação, então vale a pena conferir o app Overflight, o qual oferece uma visualização no estilo radar de todo o céu visível, com o diferencial de exibir com destaque a aeronave que está passando acima de você. É possível deslizar a tela para focar em qualquer outra aeronave e, quando o céu está tranquilo, o app permanece em estado de espera até a próxima ocorrência.
O aplicativo, é claro, também exibe detalhes como informações sobre a companhia aérea, o número do voo e o tipo de aeronave, bem como altitude, velocidade vertical, velocidade em relação ao solo, trajetória, rumo e distância.
Ele ainda pode mostrar, com destaque, o tráfego militar e de interesse especial assim que eles entram no céu. Vale notar, porém, que nem todas as aeronaves transmitem sua posição (alguns voos militares operam sem transmissão).
Na versão Pro, é possível conferir rotas em tempo real (da origem ao destino), ouvir anúncios de tráfego por voz, receber alertas de aproximação de aeronaves e ativar um modo “sempre ligado” para exibição ambiente contínua — ideal para a Apple TV, cuja versão foi recém-lançada.
O Overflight utiliza dados ADS-B gratuitos fornecidos por usuários e não requer a criação de conta ou cadastro. Além disso, sua localização é usada apenas para identificar as aeronaves acima de você.
O app está disponível gratuitamente na App Store para iPhone, iPad e Apple TV. A versão Pro, por sua vez, custa R$25/mês, R$130/ano ou a partir de uma licença vitalícia por R$500.
via 9to5Mac
© Allison Joyce/Agence France-Presse — Getty Images
Login