Turbulent fluxes are the atmosphere’s invisible hand on a glacier’s surface. Sensible heat flux, often written as H, describes the transfer of warmth from air to ice or vice versa, driven by the temperature difference between them. Latent heat flux, or LE, describes the energy carried by water vapor — when a glacier surface sublimates, ice turns directly to vapor and the glacier loses both mass and energy; when condensation or deposition occurs, the process reverses and the glacier gains. Together these two fluxes can account for roughly 14 to 35 percent of the energy budget of Tibetan Plateau glaciers, and they are a primary driver of sublimation-driven mass loss during the long cold season from October to May. Yet despite their importance, direct measurements over the plateau’s continental glaciers — the cold, dry, low-precipitation type that dominates the region — had never been systematically collected and analyzed until now.

The measurement site sat at 5,317 meters above sea level on the surface of the Dunde ice cap, which spans roughly 60 square kilometers between 4,400 and 5,500 meters elevation. The team deployed a Campbell IRGASON eddy covariance system that recorded three-dimensional wind components, virtual temperature, and water vapor concentration at 10 times per second from mid-May to mid-October 2023. Eddy covariance works by correlating rapid vertical wind fluctuations with simultaneous fluctuations in temperature and humidity, effectively weighing every parcel of air that crosses the sensor and tallying the heat and moisture it carries. After rigorous quality control — including coordinate rotation, density corrections following the Webb-Pearman-Leuning formulation, spectral corrections, and screening for low winds, extreme vertical velocities, and obstructed wind directions — 72 percent of the half-hourly flux data survived for analysis.

The observations painted a vivid picture of a glacier that is, on balance, losing energy to the atmosphere through turbulent exchange. The mean latent heat flux was −9.6 watts per square meter, indicating persistent sublimation throughout the study period, while the mean sensible heat flux was +6.9 watts per square meter, showing that the air generally warmed the ice. The latent heat term dominated in magnitude: daily mean LE ranged from −35.9 watts per square meter in late August to +10.4 watts per square meter in mid-July, and its absolute value was strongest in May and June at −15.1 watts per square meter on average. Because sublimation outweighed heat gain, the total turbulent flux averaged −2.9 watts per square meter — a net energy loss from the glacier surface. The team also documented a persistent near-surface temperature inversion, with air temperature exceeding surface temperature by an average of 1.4 degrees Celsius, a signature of the stable boundary layer that forms over cold ice.

With this observational benchmark in hand, the researchers put five widely used turbulent flux parameterizations to the test. The simplest, called C_log, applies a constant exchange coefficient driven by near-surface wind speed without any atmospheric stability correction — computationally cheap but blind to the nighttime surface inversions that characterize glacier climates. The C_kat method, derived from a simplified katabatic flow model, explicitly accounts for the glacier winds that drain cold air down the ice surface but lacks dynamic stability adjustments. Two methods, CRib1 and CRib2, use the bulk Richardson number — a dimensionless ratio of buoyancy to mechanical shear — to damp turbulent exchange under stable stratification. The most sophisticated, C_M–O, implements full Monin–Obukhov similarity theory with universal stability functions solved iteratively. For the observation-based schemes, the team derived a representative momentum roughness length of 1.2 × 10⁻⁴ meters from the eddy covariance data itself, and tested two alternative parameterizations of scalar roughness lengths for heat and vapor transfer, known as A87 and SvdB.

The verdict, across daily, seasonal, and diurnal timescales, favored the Richardson-number-based approaches. All five schemes captured the broad temporal evolution of the fluxes, but most systematically underestimated them, particularly during periods of abrupt change in early and mid-July and mid-August. For latent heat flux, root mean square errors ranged from 11.2 to 13 watts per square meter across most configurations, while sensible heat flux was simulated more accurately, with RMSE values of 3.5 to 4.1 watts per square meter for the Richardson-number methods. At the diurnal scale, CRib1 and CRib2 most faithfully reproduced the observed rhythm — latent heat peaking around 15:30 local time, sensible heat reaching its minimum near noon and maximum around 23:00. The SvdB scalar roughness parameterization consistently outperformed A87, especially during high-flux periods. The katabatic-wind-based C_kat method, despite having its empirical parameters recalibrated against the observations, performed worst overall.

The team then embedded three representative schemes — C_log, CRib2, and C_M–O with SvdB — into a coupled energy balance and snow-firn model to see how parameter choices ripple through to simulated glacier behavior. Using original published parameter values, the models produced surface temperature errors as large as 8.8 degrees Celsius RMSE and cumulative mass balance estimates that deviated markedly from an observation-constrained reference of −298.6 millimeters water equivalent. After recalibrating the turbulent flux parameters against the eddy covariance data, all three schemes improved: the CRib2 configuration showed the largest gains, cutting its surface temperature RMSE by 68.2 percent and its surface height error by 16.7 percent, while its cumulative mass balance error dropped 17.6 percent. The independently measured stake-based mass balance of −303 millimeters water equivalent closely matched the reference simulation, lending confidence to the evaluation framework.

Perhaps the most striking finding came from a four-day humid heatwave in early July 2023, when mean air temperature surged to 2.4 degrees Celsius — nearly 4 degrees above the non-heatwave July average — and relative humidity climbed to 94.3 percent. During this event, sensible heat flux more than doubled to a daily average of 17.5 watts per square meter, while the magnitude of the negative latent heat flux collapsed from −25.2 to −3.1 watts per square meter, and on July 11 it actually reversed direction, becoming positive at +5.1 watts per square meter as condensation or deposition replaced sublimation. The total turbulent flux flipped from a net loss of −13.8 watts per square meter during normal July conditions to a net gain of +14.4 watts per square meter during the heatwave — a complete sign reversal that accelerated surface melt, promoted snow metamorphism, and lowered surface albedo by exposing underlying ice. Such weakly negative or positive latent heat fluxes are normally seen over maritime glaciers or in the high-latitude Arctic, not atop a cold continental ice cap.

None of the five parameterizations fully reproduced this heatwave behavior. All schemes captured the surge in sensible heat flux — CRib1 and C_M–O with SvdB performing best at 16.7 and 16.4 watts per square meter respectively — but every method substantially overestimated the magnitude of latent heat flux during the event and failed to reproduce its occasional sign reversal. The authors attribute part of this gap to the half-hourly averaging of the driving meteorological data, which can obscure short-lived near-surface gradients, but the discrepancy also points to a genuine structural limitation: current bulk formulations struggle to represent latent heat exchange when warm, humid air sits over a melting surface whose temperature is capped at zero degrees Celsius. As the Tibetan Plateau continues to warm and moisten, such humid heatwaves are projected to become more frequent and intense, potentially pushing continental glaciers toward turbulent flux regimes more typical of maritime ice.

The study also contributes a valuable observation on surface roughness. The quality-controlled momentum roughness lengths derived from the eddy covariance data spanned five orders of magnitude, from 1.14 × 10⁻⁵ to 2.08 × 10⁻² meters, with a median of 1.2 × 10⁻⁴ meters. Comparing these values with published estimates, the authors note that continental glaciers on the Tibetan Plateau tend to exhibit broader ranges and generally lower characteristic roughness lengths than maritime glaciers, whose reported values cluster between 10⁻⁴ and 10⁻² meters. This systematic difference matters because roughness length directly controls the turbulent exchange coefficient, and using inappropriate values can bias flux estimates substantially. For a region often called the Asian Water Tower, where glacier meltwater sustains major river systems, closing this observational gap is not an academic exercise — it is a prerequisite for forecasting how the plateau’s frozen reservoirs will respond to a warming, and increasingly humid, future climate.

Subject of Research: Turbulent heat flux observations and parameterization evaluation over the Dunde Glacier on the Tibetan Plateau

Article Title: Evaluation of turbulent flux parameterizations over a continental glacier on the Tibetan Plateau

Article References: Xu, Y., Zhu, M., Gao, J., Lin, D., Hui, L., Zhu, F., Zhao, H., & Zhang, F. (2026). Evaluation of turbulent flux parameterizations over a continental glacier on the Tibetan Plateau. The Cryosphere, 20(10), 6021-6044. https://doi.org/10.5194/tc-20-6021-2026

Image Credits: AI Generated

DOI: 10.5194/tc-20-6021-2026

Keywords: Tibetan Plateau, Dunde Glacier, turbulent heat flux, eddy covariance, surface energy balance, glacier mass balance, Monin-Obukhov similarity theory, bulk Richardson number, aerodynamic roughness length, humid heatwave, sublimation, The Cryosphere

The central clue comes from an unexpected place: sodium gas. Water vapour had already been detected in the inner regions of PDS 70 as early as 2023, a puzzling observation in its own right, because the inner parts of a young planetary system are hot and irradiated, conditions in which water should not easily survive unless it is being replenished. The Lund researchers have now shown that variations in sodium gas observed in front of the star may be the signature of comets sweeping through the inner parts of the system. As a comet approaches its host star, the rising temperature causes the ice on its surface to transform directly into gas, a process known as sublimation. The escaping gas leaves a measurable imprint in the starlight, and it is precisely such an imprint that the team believes it has detected.

The observations analysed by the researchers date from 2018, and they reveal sodium gas moving at several kilometres per second relative to the star. More telling than the speed alone is the behaviour of the gas over time: it appears and then disappears over the course of several nights. That transient pattern is exactly what one would expect if comets were passing in front of the star, briefly releasing gas as they are heated and then moving on. A permanent reservoir of gas would not behave this way. The episodic appearance and vanishing of the absorption signal is the kind of fingerprint that astronomers have learned to associate with cometary activity, and it is the first time such a signal has been linked to a star of this type.

“This is the first time we have seen evidence of exocomets orbiting a star that is relatively cool, much like our Sun. Furthermore, this system is the youngest in which exocomet activity has been proposed,” says Aline Novais, an astronomy researcher at Lund University who led the analysis. The comparison with the Sun matters because most previous detections of exocomets have come from hotter stars, where the observational signatures are easier to isolate. Finding evidence of cometary activity around a cool, Sun-like star that is barely five million years old opens a window onto the earliest chapter of planetary development, the very phase during which the Earth and its neighbours were being assembled from dust, gas, and the volatile compounds delivered from farther out.

The implications reach well beyond a single system. Comets form in the cold outer regions of a planetary system, where water and other volatile substances exist as stable ices. In their distant homes, they are effectively frozen reservoirs of the raw materials of habitability. If nothing disturbs them, they remain there indefinitely. But young planetary systems are dynamically restless places, and the Lund team has used computer simulations to model how the comets around PDS 70 might be set in motion. The results show that objects far out in the system can be affected by the gravitational pull of the system’s gas giants and flung inward toward the star. The two massive planets still forming in PDS 70 are therefore not merely spectators; they may be the engines that redirect icy bodies into the inner system.

“Our study suggests that comets may be responsible for transporting water to the inner parts of the planetary system, where planets can form, in the same way as in the early Solar System,” says Novais. The mechanism is elegant in its simplicity. A comet that forms in the frigid outskirts carries ice from the moment of its birth. If a gravitational encounter with a giant planet alters its orbit and sends it plunging toward the star, sublimation releases that water, along with carbon-bearing molecules and other volatiles, into the inner regions where rocky planets are taking shape. In effect, the comets function as delivery vehicles, ferrying material across the vast temperature gradient that separates the outer system from the warm inner zone where terrestrial planets can hold liquid water on their surfaces.

That picture resonates strongly with one of the oldest questions in planetary science: where did Earth’s water come from? The young Earth formed in a region of the early Solar System that was almost certainly too hot for water ice to condense, which means the water in our oceans must have arrived from somewhere else. Water-rich asteroids and comets are the two leading candidate sources, and researchers have debated for decades which of the two played the dominant role. Isotopic measurements, particularly the ratio of deuterium to hydrogen in Earth’s water, have provided partial constraints, but the question remains open. What the PDS 70 observations offer is the chance to watch a comparable delivery process unfold in a system that is still in its infancy, rather than inferring it from the ancient, heavily altered record preserved in our own Solar System.

“It is reminiscent of a possible process in the early Solar System, in which comets may have helped to deliver water to the young Earth,” says Alexandra Stockwell Murphy, an astronomer at Lund University and a co-author of the study. The parallel is not exact, of course. PDS 70 is a specific system with its own architecture, its own giant planets, and its own history of dynamical evolution. But the physical processes at play, gravitational scattering by giant planets followed by sublimation of icy bodies in the inner system, are universal. If those processes are operating in PDS 70 today, they likely operated in the young Solar System as well, during the era when the giant planets were migrating and the population of small icy bodies was being reshuffled across the system.

The technical achievement behind the detection should not be understated. Sodium is a trace element in cometary material, but it is a valuable diagnostic because it produces strong, narrow absorption lines that stand out clearly against the spectrum of a star. Detecting the gas requires not only high-resolution spectroscopy but also careful discrimination between the signal from the passing comet and the much larger signal from the star itself and from the surrounding disc. The transient nature of the absorption, appearing and disappearing over several nights, combined with the measured velocity offset of several kilometres per second relative to the star, together provide a consistent picture of small, fast-moving bodies releasing gas as they pass through the inner system. The orbital simulations then close the loop, demonstrating that the gravitational influence of the system’s gas giants is sufficient to populate such orbits from the outer reservoir.

The next chapter of the story may arrive soon. “And when the Extremely Large Telescope, which is currently being built in Chile, becomes operational in the coming years, we will be able to find out whether there are any further planets in the system and thus gain an even clearer picture of how water and other building blocks of planets are transported,” concludes Jens Hoeijmakers, an astronomy researcher at Lund University. The Extremely Large Telescope, with its enormous collecting area and advanced instrumentation, will be capable of imaging and characterising planets around nearby young stars with unprecedented sensitivity, and PDS 70 is certain to be among its prime targets. If additional planets are found, their gravitational influence could be incorporated into more refined models of how icy bodies are scattered inward. For now, the sodium signal flickering in front of a five-million-year-old star offers a rare and tantalising glimpse of a process that may have made Earth habitable, playing out again, 370 light-years away, in a system still writing its opening pages.

Subject of Research: Evidence of water-delivering exocomets orbiting the young star PDS 70

Article Title: Comets may have transported water to a young planetary system

Article References: Comets may have transported water to a young planetary system. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: exocomets, PDS 70, water delivery, planetary formation, protoplanetary disk, sodium gas, sublimation, Lund University, Nature Communications, Extremely Large Telescope, young star, gas giants

The study, led by William Rudisill of Lawrence Berkeley National Laboratory together with Dan Feldman, Adrienne Marshall, and Arielle Koshkin, draws on an unusually rich set of observations from three co-located field campaigns in the East River Watershed near Crested Butte, Colorado. The Surface Atmosphere Integrated Field Laboratory, or SAIL, deployed by the U.S. Department of Energy’s Atmospheric Radiation Measurement program, provided cloud radar, lidar, radiation, and balloon-sonde measurements. The Sublimation of Snow campaign contributed a 20-meter instrumented tower with eddy-covariance sensors, and the SPLASH campaign added down-looking radiometers used to retrieve the temperature of the snow surface itself. Together, these campaigns captured simultaneous measurements of the atmospheric energy budget, snow thermodynamics, wind circulations, turbulence, and manual observations of surface hoar during the winters of 2022 and 2023.

The physics of surface hoar formation is a delicate balancing act. On clear nights, the snow surface radiates heat to space, cooling far below the temperature of the air above it. The team confirmed that surface hoar is favored when the snow surface is roughly 10 degrees Celsius colder than the near-surface air, creating a humidity gradient that drives water vapor from the atmosphere onto the snowpack. Because the snow surface is, by definition, saturated with respect to ice, its specific humidity falls rapidly as it cools, while the air above remains relatively moist. Molecular diffusion alone is too slow to explain the observed accumulation of hoar mass, so turbulent exchange must supply the vapor, yet windy nights are precisely the conditions that qualitatively preclude hoar growth. Resolving this paradox is central to the new work.

Clouds turn out to be a decisive factor. Using two winters of data, the researchers composited fully clear and fully cloudy 24-hour periods and found that nocturnal clouds exert a radiative forcing of 30 to 40 watts per square meter, preventing the snowpack from cooling and effectively shutting down the humidity gradient that drives deposition. On clear nights, the snow surface temperature plummeted to below minus 24 degrees Celsius shortly before sunrise, and the humidity difference between air and snow reached values of minus 0.5 to minus 1.0 grams per kilogram, consistently favoring deposition. Balloon soundings revealed that these gradients were not merely a near-surface phenomenon: the air was drier in terms of specific humidity than the snow surface well above the 800-meter depth of the surrounding valley, meaning the deposition-favoring profile extended through the entire planetary boundary layer and into the troposphere.

The team also confronted the turbulence paradox directly. Clear nights in the valley were dominated by shallow katabatic winds, gentle drainage flows that peaked at around 5 meters above the ground and ran down-valley. These flows generated just enough turbulence to deliver vapor to the surface without warming it. Critically, the researchers found that surface hoar events occurred at bulk Richardson numbers above 0.2, a stability regime in which the standard stability correction used in many land models predicts essentially zero turbulent exchange. Models that truncate turbulence at this threshold simply cannot reproduce observed hoar growth. By adopting a long-tailed stability correction, the team enabled the physics-based SUMMA model to produce deposition rates sufficient to explain the measured surface hoar mass, a finding with broad implications for how snow models handle highly stable boundary layers.

Ground truth came from a labor-intensive technique known as Stössel box measurements. On nine evenings in February 2023, researchers weighed snow samples placed in boxes, left them exposed overnight, and reweighed them the following morning. Six events showed appreciable deposition, with redundant measurements agreeing within 40 percent and values ranging from 40 to 180 grams per square meter. The SUMMA model, validated against these measurements and against eddy-covariance fluxes, reproduced snow surface temperature and humidity with correlations of 0.99, though it underestimated hoar mass on some nights. Notably, even the largest deposition events represent a small addition to the snowpack’s total mass, but they exert an outsized influence on its mechanical structure and on the surface energy balance, where a single 100-gram-per-square-meter night translates to about 6.5 watts per square meter of heating.

The sensitivity experiments revealed a subtle but important role for snow density. Because low-density snow conducts heat poorly, it insulates the surface from the warmer snow below, allowing the surface to cool more strongly and favoring deposition. When the team prescribed high thermal conductivity representative of dense snow, the overnight deposition dropped by 54 percent relative to a low-density scenario. This suggests that surface hoar should form preferentially on clear nights following fresh, low-density snowfall, a hypothesis the authors note has yet to be confirmed by direct field observation. Both observations and the model also converged on clear meteorological thresholds: deposition gives way to sublimation when overnight air temperatures rise above roughly minus 8 to minus 10 degrees Celsius or when wind speeds exceed 2 to 3 meters per second.

Armed with a validated model configuration, the researchers turned to the future. They forced SUMMA with an ensemble of nine dynamically downscaled global climate models from the WUS-D3 dataset under the SSP3-7.0 emissions scenario, running simulations from 1980 to 2100. The results are striking. The total wintertime nocturnal water vapor flux onto the snowpack declines at a rate of 6.1 grams per square meter per degree of warming, an 81 percent decrease by the end of the century. The frequency of surface hoar events falls by 14 percent per winter, at a rate of about 2.5 fewer nights per degree of warming. Counterintuitively, this decline occurs even as atmospheric humidity increases, because the snow surface warms faster than the air, weakening the very gradients that drive crystal growth. Meanwhile, sublimation events grow larger and more frequent, so the snowpack increasingly sheds vapor rather than gaining it.

The implications extend well beyond avalanche forecasting. Surface hoar influences snow albedo, surface roughness, emissivity, radar-scattering properties, and specific surface area, all quantities that remote sensing missions rely on to retrieve snow water equivalent and other properties from orbit. A systematic decline in surface hoar could therefore bias satellite retrievals and alter how snowpacks absorb solar energy, potentially feeding back into melt timing. The authors caution that their study site has a cold continental climate, so the projected declines may represent a low-end estimate for warmer snow regions, where the frequency of sufficiently cold nights could collapse nonlinearly. They also acknowledge open questions, including discrepancies between eddy-covariance fluxes and manual hoar measurements, the role of katabatic winds in non-standard turbulent transport, and the possibility that some hoar mass originates from vapor moving upward within the snowpack itself.

What emerges from this work is a portrait of a fundamental snow metamorphism process that is exquisitely sensitive to the details of a warming atmosphere: to clouds, to humidity, to wind, and to the density of the snow beneath. The feathery crystals that avalanche forecasters dread may become scarcer in a warmer world, but understanding exactly how and where requires reconciling measurements that currently disagree by factors of several. As the authors conclude, additional work on turbulent exchange under extreme stability and on vapor transport in complex terrain is needed before surface hoar can be reliably predicted in the land and snow models that underpin hazard forecasting and climate projection alike. For now, the study provides both a warning and a roadmap: the coldest, clearest nights that build the snowpack’s most treacherous weak layers are precisely the nights that a warming climate is steadily erasing.

Subject of Research: Climate sensitivity of surface hoar formation on snowpacks in the Colorado Rockies

Article Title: Reduced surface hoar in a warming world

Article References: Reduced surface hoar in a warming world. (n.d.). https://doi.org/10.5194/tc-20-5585-2026

Image Credits: AI Generated

DOI: 10.5194/tc-20-5585-2026

Keywords: surface hoar, snowpack, avalanche, The Cryosphere, climate change, Colorado Rockies, snow metamorphism, turbulent flux, SUMMA model, SAIL campaign, sublimation, longwave radiation

The team worked at a site called S10, an automatic weather station perched at 1,850 metres elevation in the lower accumulation zone, roughly 140 kilometres from the ice sheet’s western margin. S10 sits along the Kangerlussuaq transect, a well-instrumented line of stations running inland from the west coast, and it experiences blowing snow on about 16 percent of days. That frequency, combined with a rare treasure trove of direct measurements, made it the natural laboratory for the question. In September and October 2012, researchers from the Institute for Marine and Atmospheric Research Utrecht deployed a snow particle counter, a laser-based instrument that counts and sizes individual snow grains hurtling past at about one metre above the surface, alongside an 8-metre tower carrying a sonic anemometer to measure the turbulent stress that lifts the snow in the first place.

The physics of blowing snow is deceptively intricate. Once the friction velocity of the wind exceeds a threshold set by the density and texture of the surface snow, grains begin to saltate, bouncing along the ground in a shallow layer a few centimetres thick. Above that, turbulence suspends smaller particles into the boundary layer, where they can drift hundreds of metres upward and, in Antarctica, hundreds of kilometres horizontally. Suspended grains are superb sublimators: the swirling air around each crystal strips away water vapour, cooling the air and humidifying it at the same time. This sublimation is, alongside wind erosion, the only process that removes mass from the interior of the Greenland Ice Sheet, which makes it a first-order term in the surface mass balance that climate models must get right.

What had been missing for Greenland was the direct radiative fingerprint of all that airborne ice. Earlier work in Antarctica had hinted at it. Observations at Japan’s Mizuho Station in the 1980s showed that drifting snow reduced net longwave cooling at the surface, and later studies in the Canadian High Arctic measured downwelling longwave fluxes increasing by up to 36 watts per square metre during blowing snow events over rough terrain. Satellite analyses even detected the signature of blowing snow in outgoing longwave radiation at the top of the atmosphere over East Antarctica in winter. Models such as the Modèle Atmosphérique Régional had begun to include these effects for the southern continent, but the regional climate model used most widely for Greenland, RACMO2.4p1, still treated blowing snow as radiatively invisible.

Tax and his colleagues closed that gap with an elegant piece of computational plumbing. The model’s blowing snow routine, based on the double-moment bulk scheme PIEKTUK-D, already calculates the mixing ratio of suspended snow, the size of the particles, and how far the layer extends above the surface. The model’s radiation scheme, ecRad-1.4.1, describes how ice clouds absorb and scatter light. The researchers simply represented the blowing snow layer as a low-level ice cloud, feeding the snow mixing ratio, an effective particle radius, and a cloud fraction into the radiation code. When the blowing snow mixing ratio exceeded a threshold of one millionth of a kilogram per kilogram of air, the layer was treated as fully cloud-covered. They then ran three experiments: a baseline with no radiative effect of blowing snow, one that included it only above the lowest model level about 10 metres up, and one that folded the entire near-surface snow layer into the radiation calculation.

The results were unambiguous. During blowing snow events at S10, representing the airborne snow as an ice cloud increased downwelling longwave radiation at the surface by an average of 5.8 watts per square metre, with peaks reaching 50 watts per square metre under clear skies. At the same time, downwelling shortwave radiation fell by an average of 1.2 watts per square metre as the suspended grains scattered sunlight. The longwave effect dominated, so the net radiation at the surface increased. In the case study of a particularly strong event in October 2012, when the vertically integrated horizontal snow transport reached 0.4 kilograms per metre per second, the blowing snow layer extended to roughly 175 metres altitude and drove shortwave heating rates in the near-surface air up to 35 kelvin per day, while suppressing longwave cooling at rates approaching 150 kelvin per day, values far exceeding those of ordinary clouds because the snow is concentrated in such a shallow layer.

Crucially, the improvements were not merely theoretical. When the researchers compared the coupled simulations against five and a half years of radiation measurements from the automatic weather stations at S10 and its neighbour KAN_U, operated by Utrecht University and the Danish PROMICE programme, the agreement with observed downwelling longwave radiation improved markedly, particularly during clear-sky blowing snow events when the background atmospheric emissivity is at its lowest. The root mean square error of the simulated surface radiation balance during blowing snow events dropped by 3.4 watts per square metre, and the bias shrank by 4.6 watts per square metre, because the gain in longwave accuracy outweighed a slight deterioration in the already well-simulated shortwave fluxes. For a model whose longwave radiation has long been a known weak point over ice sheets, this is a meaningful correction.

The study also exposed honest limitations. The blowing snow routine captured the timing of events well but underestimated peak horizontal transport fluxes by 69 to 79 percent during the short summer observational window, a deficiency also seen over Antarctica and likely tied to how the model handles the availability of loose surface snow, its compaction in summer, and the interaction between falling snow and wind-blown snow. The researchers also had to make judgement calls with the laser data, excluding two events contaminated by concurrent snowfall and discarding an anomalous peak in the largest particle size class that they attributed to precipitation rather than transport. And the ice optics parameterisation borrowed from cirrus cloud physics was designed for high-altitude clouds, not near-surface snow, an assumption the authors flag for further testing.

Why should anyone beyond the modelling community care? Because the surface energy balance is the ledger that determines melt. At the cold, high elevation of S10, an extra few watts per square metre will not trigger melting, but closer to the warm, low-lying margins of the ice sheet, where most runoff is generated, the same radiative warming could tip the balance during summer. The authors expect the effect to be larger near the margins and note that in coastal Greenland, more frequent cloud cover will partly mask the direct signal, while secondary interactions between blowing snow and low-level clouds could amplify it. Over Antarctica, where blowing snow layers are thicker and more extensive, the radiative impact is likely to be greater still. Getting these processes into models is therefore a prerequisite for trustworthy projections of both ice sheets.

The broader lesson is a familiar one in polar science: the boundary layer, that thin skin of air hugging the ice, keeps surprising us with processes that continental-scale models have long ignored. A cloud of snow you can barely see through, lasting a few hours, turns out to bend the radiation budget in measurable ways. Tax and his colleagues recommend that climate models couple their blowing snow and radiation schemes directly, and their offline code and datasets are publicly available for other groups to build on. As the Arctic warms and the Greenland Ice Sheet sheds hundreds of gigatonnes of ice each year, accounting for every blanket and every mirror in the system is no longer optional. It is the difference between a forecast and a guess.

Subject of Research: Radiative effects of blowing snow on the surface energy balance of the Greenland Ice Sheet

Article Title: Impact of blowing snow on the surface radiation balance near the western margin of the Greenland Ice Sheet

Article References: Tax, S. M., van Tiggelen, M., Feenstra, T. N., Smeets, P. C. J. P., Gadde, S. N., van Dalum, C. T., van de Berg, W. J., & van den Broeke, M. R. (2026). Impact of blowing snow on the surface radiation balance near the western margin of the Greenland Ice Sheet. The Cryosphere, 20(10), 5609-5627. https://doi.org/10.5194/tc-20-5609-2026

Image Credits: AI Generated

DOI: 10.5194/tc-20-5609-2026

Keywords: Greenland Ice Sheet, blowing snow, surface radiation balance, surface mass balance, RACMO2.4p1, regional climate modelling, sublimation, longwave radiation, shortwave radiation, boundary layer, sea level rise, The Cryosphere

The research, led by Kaifeng Ma of the Northwest Institute of Eco-Environment and Resources of the Chinese Academy of Sciences together with Junfeng Liu, Chuntan Han and Rensheng Chen, systematically evaluated six typical gridded evapotranspiration products against ground observations from ten sites spread across the high-latitude cold regions of the Northern Hemisphere. The products tested represent the main families of approaches currently in use: EB, a surface energy balance method; CAMELE, a collocation-analyzed multi-source ensemble; ETMonitor, a model driven by multi-source satellite observations; GLDAS, a land data assimilation system; GLEAM, a global land evaporation and soil moisture dataset; and ML, a machine learning based product. Each of these datasets embodies a different philosophy for estimating a flux that cannot be directly measured from space, and each has been validated extensively in temperate and tropical environments. What the team wanted to know was whether that validation record carries over to landscapes where subfreezing temperatures and persistent snow cover dominate for much of the year.

The answer, in short, is no. During the warm season, when soils thaw, vegetation is active and the physics of evaporation resembles the conditions under which most of these products were developed, the picture is mixed but tolerable. The machine learning product and CAMELE performed relatively well, achieving mean correlation coefficients of 0.72 and 0.60 respectively with the ground-based observations. Yet even in the warm season the study found significant spatial heterogeneity among sites, meaning that a product that tracks evapotranspiration skillfully at one location may drift badly at another only a few hundred kilometers away. That kind of spatial inconsistency matters enormously for applications such as basin-scale water balance calculations, where errors at individual grid cells propagate into regional water resource assessments.

The cold season is where the datasets truly fall apart. When snow blankets the ground, soils freeze and the sun sits low on the horizon, the reliability of all six products decreases substantially, and at most sites the correlation coefficient approaches zero. In practical terms, a correlation of zero means the product’s estimates of evapotranspiration carry essentially no information about what is actually happening on the ground. This is a striking result, because winter is precisely the period when the high-latitude water budget is subtle and easily misrepresented. Snow sublimation, the direct conversion of snow and ice into vapor, can remove a meaningful share of the winter snowpack, and blowing snow events can export moisture from a basin entirely. If gridded products cannot reproduce even the timing of these fluxes, then any attempt to close the regional water balance using them alone is built on sand.

To go beyond simple correlation-based evaluation, the team applied a statistical technique known as the three-cornered hat method, or TCH. Originally developed in metrology to estimate the individual instabilities of clocks when no absolute reference is available, the three-cornered hat approach exploits the differences between multiple imperfect datasets to estimate the relative uncertainty of each one without requiring a perfect ground truth. This is an elegant solution to a chronic problem in flux validation: eddy-covariance towers, the standard instruments for measuring evapotranspiration on the ground, are themselves imperfect, particularly in winter, when energy balance closure problems and instrument icing are well documented. By treating the products and the observations as a set of mutually correlated estimates, the TCH analysis allowed the researchers to quantify how much uncertainty each product contributes independently of the others.

The TCH results reinforced the seasonal pattern. Relative uncertainty was lower during the warm season than during the cold season across the product suite, confirming that the winter degradation is not an artifact of one flawed dataset but a systemic weakness shared by energy balance models, satellite-driven products, reanalysis systems, ensembles and machine learning approaches alike. The convergence of evidence from two independent evaluation frameworks, direct comparison with site observations and the model-free uncertainty decomposition, makes the conclusion difficult to dismiss. Whatever the technical lineage of a given product, the combined effects of low temperatures and snow cover push it beyond the regime in which its underlying assumptions were calibrated.

Why should the cold season be so hostile to evapotranspiration estimation? The technical reasons are rooted in process parameterization. Most evapotranspiration algorithms were built around the physics of liquid water: stomatal conductance of leaves, aerodynamic transfer from moist soil, and radiative driving by abundant sunlight. Under snow cover, the relevant processes shift to sublimation from a snow surface, vapor transport through a turbulent boundary layer over ice, and energy exchanges complicated by the high albedo and low thermal conductivity of snow. Field studies have long shown that closing the surface energy balance over homogeneous midwinter snowpack is notoriously difficult even with careful instrumentation, so it is unsurprising that global products struggle. In addition, the machine learning products, however skillful in the training domain, inherit the distributional blind spots of their training data, which are dominated by flux tower sites in warmer, snow-free conditions. When asked to extrapolate to a frozen tundra or taiga landscape in January, they are effectively guessing.

The stakes are rising because the high latitudes are warming faster than almost anywhere else on Earth. The Arctic has warmed nearly four times faster than the globe since 1979, and this amplified warming is reshaping snow cover duration, permafrost extent, vegetation composition and the partitioning of energy at the surface. Boreal peatlands are contributing an increasing share of regional evapotranspiration as the climate warms, and changes in winter sublimation feed directly into snow mass balance and spring runoff, which in turn supply major river systems. Climate models and hydrological projections rely on gridded evapotranspiration datasets both as inputs and as benchmarks. If those datasets are unreliable in the very regions undergoing the most rapid change, then model evaluations, drought indices and carbon budget assessments for the circumpolar North carry hidden errors that no amount of averaging will remove.

The authors are explicit about the practical implication: current mainstream evapotranspiration products should be used with caution in high-latitude cold regions, especially during the cold season. This is not a call to abandon the datasets, which remain valuable tools across most of the planet, but a warning against uncritical application at the climatic margins. The study points toward a clear research agenda: improving process parameterization under low-temperature and snow-covered conditions, expanding the network of winter flux observations in cold regions so that training and validation data reflect the frozen world, and developing products that explicitly represent sublimation and snow-atmosphere exchange rather than treating winter as a low-flux afterthought. Emerging techniques such as structure-from-motion photogrammetry for measuring snow surface sublimation offer promising routes to the ground truth data that winter validation has lacked.

For the broader Earth science community, the study is a reminder that global datasets are only as good as the environments in which they have been tested. The elegant statistics of ensemble products and deep learning can conceal, rather than cure, physical blind spots. As the planet’s frozen margins thaw and shift, the demand for trustworthy evapotranspiration information in the high-latitude Northern Hemisphere will only grow, and closing the gap between what the grids say and what the snow actually does has become one of the more urgent tasks in land surface science.

Subject of Research: Uncertainty in gridded evapotranspiration products across high-latitude cold regions of the Northern Hemisphere

Article Title: Substantial uncertainties in gridded evapotranspiration products in high-latitude cold regions of the Northern Hemisphere

Article References: Ma, K., Liu, J., Han, C., & Chen, R. (2026). Substantial uncertainties in gridded evapotranspiration products in high-latitude cold regions of the Northern Hemisphere. Theoretical and Applied Climatology, 157(10), Article 642. https://doi.org/10.1007/s00704-026-06549-5

Image Credits: AI Generated

DOI: 10.1007/s00704-026-06549-5

Keywords: evapotranspiration, high-latitude cold regions, snow cover, three-cornered hat method, gridded datasets, remote sensing, machine learning, sublimation, climate change, Northern Hemisphere, land-atmosphere exchange, Theoretical and Applied Climatology