The new system, called the GNSS-based Profiling Tropospheric Humidity Network, or GPTHNet, was developed by Zhou Shoudong, Wang Degao, Zheng Qingqing and Jiang Peng and published in Frontiers in Astronomy and Space Sciences. The core idea is deceptively simple: GNSS signals traveling from satellites to a ground receiver are delayed and bent as they pass through the neutral atmosphere, and the size of that delay reveals how much water vapor lies in the column of air above the station. Meteorologists have exploited this effect for decades to calculate precipitable water vapor, a single number representing the total moisture in the column. The problem is that this column total says nothing about how the moisture is layered — whether the vapor is concentrated near the ground, pooled beneath a temperature inversion, or distributed through the middle troposphere where storms draw their fuel.

Traditionally, that vertical detail has come from radiosondes, the instrumented balloons launched twice daily from stations that may be hundreds of kilometers apart. Radiosondes are accurate, but the atmosphere between two launches, twelve hours apart, can transform entirely — a fact that matters enormously for forecasting flash floods, severe convection and monsoon rainfall. Alternatives such as microwave radiometers and Raman lidars can profile moisture continuously, but they are expensive, and their performance degrades in heavy precipitation precisely when humidity data are most valuable. GNSS tomography, which divides the troposphere into three-dimensional voxels sampled by many slant signal paths, offers a continuous alternative but demands a dense network of receivers covering the region of interest, along with carefully tuned mathematical constraints.

The Chinese team’s approach sidesteps that infrastructure requirement entirely. Instead of physically inverting signal paths, their multilayer perceptron learns the statistical relationship between what a single station can measure and what the humidity profile looks like. The network’s inputs fall into two categories. The first part describes the current atmospheric state at the profiling moment: surface air pressure, temperature and absolute humidity measured at two meters, the GNSS-derived precipitable water vapor, and the decimal day of year, which encodes seasonality. The second part provides an initialization snapshot from the most recent radiosonde launch before the profiling epoch: ground-level meteorology, the radiosonde-derived column water vapor, the full measured humidity profile, and the eastward and northward wind components resampled onto standard heights. From these inputs, the network outputs absolute humidity at 40 height levels from 250 meters to 10,000 meters, at 250-meter intervals.

Training data came from the Atmospheric Radiation Measurement User Facility’s Tropical Western Pacific site in Darwin, Australia, a location chosen for its extreme humidity variability and its monsoon cycle. The researchers assembled 5,872 input samples from radiosonde and surface measurements spanning 2002 to 2011, correcting known dry-bias and time-lag artifacts in the Vaisala RS-80 and RS-92 sondes using the ARM Sonde Adjust value-added product. The dataset was split 60:20:20 into training, validation and test sets. Rather than guessing at an architecture, the team grew their networks neuron by neuron, training each configuration a thousand times, adding hidden layers until the validation loss converged. The final initialization model stacks an input layer of 130 neurons, two hidden layers of 91 and 81 ReLU units, and a linear output layer of 40 neurons, trained with a quadratic cost function and the Adam optimizer at a learning rate of 0.01, with early stopping guarding against overfitting.

A second model, ANN_NoInit, was built for situations where no radiosonde sounding is available to seed the retrieval. It relies only on the surface and GNSS inputs, using six input neurons and three hidden layers of 100, 82 and 56 units. The trade-off is clear in the results: both models track radiosonde profiles well, but the initialized version captures fine structure — including humidity inversion layers in the middle and upper atmosphere — that the uninitialized model smooths away. On the withheld test data, the initialized model’s bias stayed within ±0.14 grams per cubic meter at all standard levels, with a root mean square error peaking at 1.72 grams per cubic meter at 500 meters and falling to just 0.07 grams per cubic meter at the top level.

The crucial test came with data the networks had never seen in any form. The researchers constructed GNSS–radiosonde-combined inputs, in which the real-time part of the input was drawn from the DARW station of the International GNSS Service rather than from balloons, converting the station’s zenith total delay into precipitable water vapor using a hydrostatic delay model and their own time-varying global temperature model for the wet-delay conversion. After a 45-meter height adjustment between the GNSS and radiosonde sites, the GNSS-derived water vapor agreed with the balloon values with a correlation of 0.98. Feeding 419 such samples from 2012 into the trained networks, the initialized model achieved a maximum root mean square error of 1.90 grams per cubic meter at 500 meters, degrading gracefully to below 0.1 grams per cubic meter at the highest levels — accuracy the authors report as comparable to full GNSS tomography, but achievable with a single receiver.

Independent validation against Raman lidar, which profiles moisture every 10 minutes at 75-meter vertical resolution, exposed both the method’s potential and its limits. During two weeks of the monsoon season, the initialized model’s relative error grew from roughly 12 percent at the lowest level to about 75 percent aloft, while in the dry season the error peaked near 2,250 meters, where absolute humidity drops sharply. The uninitialized model fared notably worse in dry conditions, because when surface moisture is decoupled from the upper column, the network without an initialization profile can only fall back on the climatological average it learned during training. Part of the systematic underestimation against the lidar likely reflects instrument differences rather than retrieval failure: the lidar operates on an entirely different measurement principle and is known to suffer calibration uncertainties in the lower troposphere, and the neural network’s quadratic loss function naturally regresses predictions toward the mean, smoothing sharp humidity peaks.

One practical concern was the initialization lag. Because balloons launch every twelve hours, the training samples always paired a current snapshot with a twelve-hour-old profile. Testing 435 samples with shorter lags showed a reassuring result: the initialization profile itself degrades rapidly as it ages, with some errors exceeding 3.5 grams per cubic meter, while the network’s output barely changes — the fitted slope of its error against lag time was a negligible −0.0087 grams per cubic meter per hour. The network has effectively learned to correct stale soundings, which is exactly what an operational system needs.

The authors are candid about the approach’s weaknesses. Neural networks hide their physics inside opaque weights, making errors hard to diagnose, and assimilating any new input variable would require designing and retraining an entirely new architecture. But they also point to a practical path forward: because the input and output design is site-independent, the trained networks can be adapted to new locations through transfer learning, fine-tuning with just one to two years of local radiosonde profiles rather than a decade of training data. Given that thousands of GNSS meteorological stations already stream data continuously around the globe, a technique that turns any one of them into a standalone humidity profiler — feeding severe-weather warnings, nowcasting systems and numerical weather models at refresh rates balloons can never match — could reshape how the lowest layer of the atmosphere is monitored, one receiver at a time.

Subject of Research: Retrieval of tropospheric absolute humidity profiles from a single ground-based GNSS meteorological station using artificial neural networks

Article Title: Remotely profiling the tropospheric absolute humidity using a single ground-based GNSS meteorological station based on an artificial neural network

Article References: Shoudong, Z., Degao, W., Qingqing, Z., & Peng, J. (2026). Remotely profiling the tropospheric absolute humidity using a single ground-based GNSS meteorological station based on an artificial neural network. Frontiers in Astronomy and Space Sciences, 13, Article 1984982. https://doi.org/10.3389/fspas.2026.1984982

Image Credits: AI Generated

DOI: 10.3389/fspas.2026.1984982

Keywords: GNSS meteorology, water vapor profiling, artificial neural network, precipitable water vapor, radiosonde, Raman lidar, troposphere, multilayer perceptron, weather forecasting, machine learning, absolute humidity, atmospheric remote sensing

The central finding of the study is sobering for forecasters: none of the numerical weather prediction models available at the time could anticipate the severity of the rainfall intensity. Numerical weather prediction, or NWP, relies on physics-based simulations of the atmosphere run on supercomputers, and while these models have become remarkably skilled at forecasting synoptic-scale systems such as monsoon depressions and cyclones, they routinely struggle with localized, short-lived convective extremes. The Kolkata event fell squarely into that blind spot. The rain was not produced by a large, well-organized weather system that models could track for days in advance; instead, it emerged from intense convection confined to a small area, persisting in nearly the same location for hours and stacking up rainfall totals that overwhelmed urban drainage.

To understand the event, the authors assembled an unusually comprehensive set of observations. They drew on radiosonde and radio wind soundings, known as RSRW, which are balloon-borne instruments launched from meteorological stations that profile temperature, humidity and wind from the surface up through the depth of the troposphere. They complemented these in-situ measurements with ERA5, the fifth-generation reanalysis dataset produced by the European Centre for Medium-Range Weather Forecasts, which blends historical observations with model physics to create a physically consistent picture of the atmosphere. On top of that, the researchers examined radar and satellite observations, which captured the storm’s structure and evolution in real time. This multi-instrument approach allowed them to reconstruct both the thermodynamic environment, meaning the temperature and moisture conditions that fuelled the storm, and the dynamic behaviour of the convection itself.

What emerged from the RSRW and ERA5 analysis was a set of meteorological parameters with significantly elevated values over the affected area, and, importantly, those elevated values coincided closely with the spatial and temporal distribution of the rainfall. In other words, the atmosphere over Kolkata and its surroundings displayed a distinctive thermodynamic fingerprint before and during the deluge: extreme instability, abundant moisture through a deep column of air, and the kind of profile that strongly favours vigorous, sustained convection. The authors argue that these parameters can serve as precursors for extreme rainfall conditions of short duration, offering a way to anticipate such events even when conventional NWP rainfall forecasts fail to capture their localized, high-impact nature.

This precursor-based approach matters because thermodynamic indices have a long history in severe weather forecasting. Classic indices such as the K index, the Total Totals index and various convective available potential energy measures, developed by researchers including George, Miller, and Litynska and colleagues in the mid-twentieth century, condense complex sounding data into single numbers that indicate whether the atmosphere is primed for thunderstorms. Bondyopadhyay has previously worked on determining suitable thermodynamic indices for predicting thunderstorm events over eastern India, and the new study extends that line of thinking from ordinary thunderstorms to the extreme tail of short-period rainfall. Rather than asking simply whether storms will form, the precursor framework asks whether the environment is loaded with enough moisture and instability to produce rainfall of exceptional intensity if convection does fire.

The radar and satellite component of the analysis added a second, equally important layer of insight. The observations revealed strong convection prevailing over the Kolkata area in an almost stagnant state for several hours. This persistence is the key to understanding why the event was so destructive. Convective storms often move with the prevailing winds, spreading their rainfall across a swath of territory; when a storm instead remains anchored over the same small area, its rainfall accumulates on one spot, and the hydrological consequences multiply. The nearly stationary character of the Kolkata convection meant that a few hours of intense rain fell on a single urban basin, producing flooding that the city’s infrastructure had no capacity to absorb. The authors emphasize that this kind of near-stagnant convective signature, visible in real-time radar and satellite data, provides highly useful information for nowcasting, the short-range forecasting discipline concerned with conditions over the next few hours.

To place the Kolkata event in context and to generalize their methodology, the researchers also analysed two secondary cases in which high amounts of rainfall were reported within very short periods. This comparative strategy is essential for distinguishing a genuine precursor signal from a one-off coincidence. If the same constellation of elevated thermodynamic parameters appears across multiple short-duration extreme rainfall events, then the parameters can plausibly be used as operational warning indicators rather than merely as diagnostic curiosities examined after the fact. The study’s framing also connects to a broader scientific debate about terminology: as recent work on cloudburst disambiguation has pointed out, not every extreme short-term rainfall event constitutes a cloudburst in the strict sense, and careful case-by-case analysis of the kind performed here helps refine how meteorologists classify and communicate these hazards.

The broader climatic backdrop gives the study additional urgency. Research over the past decade has documented a striking rise in extreme precipitation across India, including a threefold increase in widespread extreme rain events over central India and a strengthening of summer monsoon precipitation extremes linked to long-term climatic and anthropogenic factors. Dense cities are particularly exposed to this trend because short-duration rainfall extremes translate almost directly into urban flooding: impervious surfaces prevent infiltration, drainage systems are designed for historical rainfall statistics, and millions of residents and assets are packed into low-lying areas. An event like the September 2025 Kolkata deluge, unprecedented in intensity for the city, illustrates how quickly the tail of the rainfall distribution can outpace the assumptions embedded in urban planning and emergency response.

It is notable that much of the existing literature on short-duration extreme rainfall in India focuses on the Himalayan region, where cloudbursts and the flash floods and debris flows they trigger have been studied extensively in places such as Kedarnath, Uttarkashi and Nainital, often using reanalysis datasets like ERA5 and the India Meteorological Department’s high-resolution reanalysis. The Kolkata event demonstrates that the same analytical toolkit, sounding-based thermodynamic diagnostics combined with reanalysis and remote sensing, applies equally to a sprawling tropical delta city far from the mountains. The authors acknowledge the India Meteorological Department for providing the data and support that made the research possible, and their institutional position inside the Regional Meteorological Centre Kolkata underscores the operational motivation behind the work: finding practical, computationally light indicators that duty forecasters can monitor alongside model guidance.

The practical takeaway is a two-track strategy for anticipating and managing short-duration rainfall extremes. Before an event, elevated values of key thermodynamic parameters derived from soundings and reanalysis can flag environments capable of producing extreme localized rainfall, compensating for the known inability of NWP models to resolve the precise intensity and placement of such convection. During an event, radar and satellite monitoring of convective motion becomes decisive: if strong convection is detected and it is nearly stationary, forecasters can warn of accumulating rainfall and escalating flood risk in real time. Neither track alone would have been sufficient for Kolkata in September 2025, but together they form a coherent nowcasting and early-warning framework. As extreme short-period rainfall grows more frequent in a warming climate, the identification of reliable meteorological precursors of the kind documented in this study may prove to be one of the most valuable tools available for protecting densely populated cities from the next unprecedented night of rain.

Subject of Research: Meteorological precursors of an extreme short-duration rainfall and flooding event over Kolkata, India

Article Title: Analysis of extreme rainfall episode in short period over Kolkata & neighbourhood, India and identification of meteorological precursors

Article References: Bondyopadhyay, S., & Biswas, H. R. (2026). Analysis of extreme rainfall episode in short period over Kolkata & neighbourhood, India and identification of meteorological precursors. Theoretical and Applied Climatology, 157(11), Article 708. https://doi.org/10.1007/s00704-026-06630-z

Image Credits: AI Generated

DOI: 10.1007/s00704-026-06630-z

Keywords: Kolkata, extreme rainfall, flash flooding, nowcasting, numerical weather prediction, ERA5 reanalysis, radiosonde, thermodynamic indices, convection, India Meteorological Department, urban flooding, monsoon extremes

The core challenge the team set out to address is one of the most consequential blind spots in atmospheric science. Aerosols, the tiny solid and liquid particles suspended in the air, are far from uniformly mixed. Their optical and microphysical properties can change dramatically with height, and particles in the planetary boundary layer often behave very differently from those in the free troposphere above. These vertical contrasts govern how aerosols scatter and absorb sunlight, how they interact with clouds, and how long they persist in the atmosphere. Yet the tools available to profile them each carry significant drawbacks. Ground-based and satellite lidars suffer from signal attenuation in dense layers and depend on retrieval assumptions about aerosol type. Research aircraft provide the most direct measurements but at operational costs that rule out routine use. UAVs and tethered balloons are limited in vertical range and payload.

AeroSonde was designed to fill that gap with an approach that is almost aggressively simple. The system pairs a standard Vaisala RS41 radiosonde with a compact add-on unit weighing just 175 grams including its housing, measuring 90 by 60 by 40 millimeters. At its heart sits a Sensirion SPS30 particulate matter sensor, a device that uses a 660-nanometer laser diode, a photodiode detector, focusing optics, and a small integrated fan to measure particle mass concentrations in the PM1, PM2.5, PM4, and PM10 fractions, along with data in five particle size bins. A Raspberry Pi RP2040 microcontroller manages data acquisition and streams measurements at one hertz to a ground station over a LoRaWAN radio link operating in the EU 863 to 870 megahertz band. The team confirmed reliable data reception even when the probe drifted more than 100 kilometers horizontally and more than 25 kilometers vertically from the receiver.

Crucially, the researchers did not simply trust the sensor out of the box. Over 22 days during the summer of 2025, five SPS30 units were run side by side in a laboratory at the Institute of Geophysics, University of Warsaw, and then calibrated against reference-grade instruments: an Aurora 4000 polar nephelometer for scattering measurements and a TSI Laser Aerosol Spectrometer 3340 for particle size. The inter-sensor agreement was strikingly tight, with average Pearson correlation coefficients of 0.97 for PM10 and PM1 when one-hour averaging was applied, and mean relative differences of less than 9 percent. Using a non-linear regression model grounded in Lorenz-Mie scattering simulations, the team converted the sensor’s mass readings into estimates of the aerosol scattering coefficient, achieving a coefficient of determination of 0.96 against the nephelometer with a root mean square error of just 5.1 inverse megameters, or 3.8 percent.

The calibration also revealed honest limits. The scattering Ångström exponent, an indicator of particle size distribution, and the effective particle radius could be estimated from the sensor’s typical particle size parameter, but with substantially larger uncertainties, limiting their use to qualitative discrimination between fine-mode and coarse-mode particles. The team also verified that the sensor’s fan maintained a stable speed of roughly 6000 revolutions per minute across pressures from 1000 down to 50 hectopascals, confirming that airflow through the measuring chamber requires no correction at altitude, and that the sensor’s response time of under five seconds supports a vertical resolution of 15 to 20 meters at typical balloon ascent rates of 3 to 4 meters per second.

The field campaigns put all of this to the test under genuinely challenging conditions. Three launches from the SolarAOT Radiative Transfer Station in Strzyżów, a rural background site on Niebylecka Hill, captured two episodes of long-range transport of biomass-burning aerosol from North America in June and August 2025, and a Saharan dust intrusion in late August. Back-trajectory modeling with the LAGRANTO tool, driven by ECMWF wind fields, traced the smoke plumes across central Canada and the Atlantic, while the ICAP multi-model ensemble indicated smoke optical depths exceeding 0.1 over Poland, accounting for 61 percent of the total column loading. During the dust event, trajectories showed air masses arriving from northern Africa, with dust contributing 58 percent of the total aerosol optical depth.

The comparison against independent remote sensing was the decisive test. Because the balloon drifts horizontally during its one-to-two-hour ascent, the team developed a spatiotemporal matching procedure that aligns each AeroSonde measurement with the lidar profile at the time the sampled air parcel passed over the station, using wind data from the co-launched radiosonde. When the resulting scattering coefficient profiles were compared with aerosol extinction profiles retrieved from a Raymetrics lidar using two variants of the Klett inversion, the Pearson correlations ranged from 0.74 to 0.81, with the agreement holding across all altitude ranges from the boundary layer up to above 8 kilometers. The positive bias between extinction and scattering was consistent with the expected contribution of aerosol absorption. Column-integrated aerosol optical depth derived from the AeroSonde profiles agreed with AERONET sun photometer values to within 0.01 to 0.05 for the biomass-burning cases and 0.11 for the dust event.

The case studies also revealed a physics-driven subtlety that the authors address candidly. During the Saharan dust sounding, both scattering and optical depth were underestimated relative to the sun photometer and lidar. Several factors conspire here: mineral dust absorbs light, the SPS30 measures at a scattering angle of roughly 90 degrees where the dust phase function reaches a minimum, and the sensor’s design is better suited to fine-mode particles than to the coarse particles above 5 micrometers that dominate dust. The team argues that aerosol-type-dependent calibration against reference instruments is essential, particularly in dust-influenced environments, and points to future upgrades including a single-channel aethalometer to retrieve absorption directly and an optical particle sizer with an extended size range.

What makes the result genuinely exciting is what it unlocks. Unlike lidar, the AeroSonde is not hampered by the geometric overlap limitations that blind ground-based systems in the lowest few hundred meters, allowing more reliable characterization of boundary-layer aerosol contributions, which the measurements showed ranged from 1 to 36 percent of total column optical depth during the transport events. Unlike drones, the passive ascent avoids flow-induced sampling artifacts and can reach into the stratosphere, opening a window on volcanic plumes and stratospheric aerosol that UAV platforms cannot access. The expendable nature of the platform, each launch sacrifices the sensor, is the trade-off, but at forty dollars per unit the economics remain compelling compared with aircraft campaigns costing orders of magnitude more.

The researchers frame the work as a proof of concept, and the limited number of profiles means broader validation across more aerosol regimes is still needed. But the implications reach well beyond one Polish hilltop. Vertical profiles of this kind provide an independent check on reanalysis products and model simulations of aerosol transport, and combining AeroSonde scattering measurements with Raman lidar extinction retrievals could eventually yield single-scattering albedo profiles, a key quantity for climate forcing calculations. If a forty-dollar sensor, a microcontroller, and a weather balloon can deliver scientifically useful profiles of the particles that shape our climate, the era of dense, affordable, three-dimensional aerosol monitoring may be closer than anyone expected.

Subject of Research: Balloon-borne aerosol profiling using calibrated low-cost optical particle sensors

Article Title: A new balloon-borne system for measuring the vertical variability of aerosol optical properties using low-cost sensors

Article References: Broda, M., Zawadzka-Mańko, O., Chiliński, M. T., Nurowska, K., Kłapiński, S., Makuch, P., & Markowicz, K. M. (2026). A new balloon-borne system for measuring the vertical variability of aerosol optical properties using low-cost sensors. Atmospheric Measurement Techniques, 19(18), 5989-6014. https://doi.org/10.5194/amt-19-5989-2026

Image Credits: AI Generated

DOI: 10.5194/amt-19-5989-2026

Keywords: aerosols, balloon-borne measurements, low-cost sensors, SPS30, lidar, atmospheric profiling, biomass burning aerosol, Saharan dust, aerosol optical depth, planetary boundary layer, radiosonde, atmospheric measurement techniques

The study area itself is almost engineered for fog. The Eastern Gangetic Plain spans roughly 94,000 square kilometres of flat land lying mostly below 100 metres, wedged between the Himalayas to the north, with peaks exceeding 8000 metres, and the Chotanagpur Plateau to the south. This basin-like geometry traps moisture and aerosols, while westerly and northwesterly winds in winter funnel pollutants from western India and Pakistan into the region like a low-level atmospheric river, supplying abundant condensation nuclei. Add elevated soil moisture from irrigation and river systems, recurrent cold-wave episodes, and weak pressure gradients that keep horizontal winds sluggish, and the ingredients for persistent dense fog assemble almost automatically each December and January. At Jay Prakash Narayan International Airport in Patna, situated centrally in the basin at about 53 metres above sea level, runway visual range regularly falls below operational minima, producing cascading delays, diversions, and cancellations.

Two contrasting fog episodes dominated the study window, and their differences turned out to be scientifically revealing. The first, from 16 to 19 December, was synoptically forced: a deep, positively tilted upper-level trough anchored over northwest India drove strong southwesterly jet flow above 20 metres per second, cold-air advection, and a spatially coherent signature of subsidence at 700 hectopascals, with pressure velocities between minus 10 and minus 20 times ten to the minus two pascal-seconds. This descending branch of an active western disturbance reinforced a capping inversion over the fog layer. The second episode, from 23 to 27 December, unfolded after the trough moved eastward and quasi-geostrophic forcing for ascent weakened. Vertical motion over Patna became fragmented and nearly neutral, and the boundary layer shifted into a radiative-thermodynamic regime, exquisitely sensitive to the diurnal cycle, yet paradoxically more stable and longer-lived than the first.

The turbulence diagnostics from the second episode were extraordinary. Turbulent kinetic energy, which peaked near 60 square metres per second squared around 23 and 24 December, collapsed to nearly zero by 25 December and stayed there. Wind shear fell from about 10 times ten to the minus three per second to between 3 and 5 times the same unit. The bulk Richardson number, a dimensionless ratio weighing buoyant suppression of mixing against mechanical shear production, climbed to roughly +300 on 24 December and then held between +100 and +200 for three days, the highest sustained values in the entire record. Ventilation coefficients, the product of boundary-layer height and 10-metre wind speed, crashed from about 6000 square metres per second to 1000 to 1500, indicating the most severe atmospheric dispersion limitation observed. Mixing heights were squeezed to just 250 to 350 metres. In effect, every mechanism that could erode the fog was switched off simultaneously.

From the radiosonde analysis emerged what the authors call a double-lock mechanism for fog persistence. Persistent dense fog required not one condition but a coordinated set: surface-based temperature inversions with bases below roughly 150 to 200 metres, moderate thermal strengths of 2 to 6 degrees Celsius, vertical depths exceeding 2000 metres, bulk Richardson numbers above 10 to 20, and ventilation coefficients below about 5000 to 6000 square metres per second. When normalized inversion strength and depth both exceeded 0.5 alongside these stability and ventilation constraints, persistent fog reliably followed. The 2000-metre depth threshold proved especially diagnostic, separating brief events lasting under six hours from prolonged ones exceeding fifteen hours, consistent with mixed-layer theory in which deeper capping inversions resist turbulent erosion far longer than shallow ones.

Perhaps the most striking single observation came on 29 December, after the main episodes had ended. A powerful inversion with a thermal strength of 10.5 degrees Celsius sat with its base above 700 metres, and it produced no fog at all. The reason is geometric rather than thermodynamic: an elevated inversion decouples the stable layer from surface moisture, so however strong the temperature jump, it cannot trap water vapour where fog droplets need to condense. Inversion placement, the study concludes, matters more than inversion intensity. This finding directly challenges forecasting practice that weights inversion magnitude alone, and it underscores the primacy of surface-atmosphere coupling in continental fog formation. During the same transition, the Richardson number swung from +60 to -50 within twenty-four hours, ventilation coefficients surged to 8000 to 10000 square metres per second, and inversion strength collapsed from 10.5 to below 2 degrees Celsius, a coordinated multi-parameter disaggregation triggered by synoptic rearrangement.

To understand how fog actually dies, the researchers built a rule-based classification framework assigning each of 49 dissipation events to radiative, subsidence-driven, mechanical, or mixed categories, ordered by precedence from synoptic vertical motion through surface wind and shear to residual radiative signatures. The result was unambiguous: mechanically driven processes, whether acting alone or jointly with other boundary-layer dynamics, dominated fog erosion. Mixed-mechanical events accounted for 51.0 percent of cases and pure mechanical dissipation another 20.4 percent, meaning wind shear intensification and turbulent mixing governed roughly 71 percent of all fog clearance at Patna. Mixed-subsidence pathways contributed 14.3 percent, while pure radiative, pure subsidence, and mixed-radiative mechanisms together managed only about 8 percent, a reflection of winter’s short days, low solar angles, and optically thick fog layers that starve the surface of warming radiation. A small residual of 6.1 percent remained diagnostically ambiguous, clustered under extreme stability where all dissipative signals fell below detection thresholds.

The framework was stress-tested and verified rather than merely asserted. A sensitivity analysis perturbing each classification threshold by plus or minus 10 and 20 percent of its interquartile range showed that radiative thresholds were highly stable, with zero reclassification for several parameters, while mechanical thresholds, ironically governing the dominant category, were the least robust, reclassifying up to 39 percent of events under the largest perturbation. Binary verification of the fog-favourable criterion, a transitional inversion with base between 50 and 500 metres, against 384 matched surface observations yielded an accuracy of 0.77, precision of 0.76, recall of 0.33, an F1-score of 0.46, and a Heidke Skill Score of 0.34, confirming genuine predictive skill beyond climatology. The high-precision, low-recall profile positions the tool as a conservative, low-false-alarm mechanism-attribution instrument rather than an exhaustive all-height fog forecaster, a distinction with real operational consequences for aviation safety managers who must balance disruption costs against missed warnings.

The practical payoff is a companion Fog Dissipation Potential Index that distills the multi-parameter picture into short-lead operational guidance, distinguishing transient from persistent fog and clarifying where predictability simply runs out under extreme stability. The authors benchmarked their diagnostic approach against a complementary decision-tree nowcasting system developed for Delhi’s Indira Gandhi International Airport, noting that the two converge on similar wind-speed regime boundaries and could be paired operationally, since the decision tree’s known failure mode, fog onset without relative humidity saturation at tower level, is precisely the signature this framework would attribute to subsidence or mechanical forcing. The team is candid about limitations: aerosol indicators such as PM2.5 and cloud condensation nuclei were deliberately excluded, turbulent kinetic energy cannot be resolved from standard ERA5 fields, thresholds are site- and season-specific, and some moisture parameters are collinear by construction. Even so, for one of the most fog-burdened and economically vital regions on Earth, the study delivers something rare: physically interpretable, quantitatively verified thresholds that explain not just when the fog comes, but why it stays, and what finally sets the plains free.

Subject of Research: Boundary-layer thermodynamic controls on winter fog persistence and dissipation over the Eastern Gangetic Plains

Article Title: Fog persistence and dissipation over the Eastern Gangetic Plains insights from boundary layer thermodynamics during 16 to 31 December 2025

Article References: Shankar, A., Kant, S., Anand, A., & Sarthi, P. P. (2026). Fog persistence and dissipation over the Eastern Gangetic Plains insights from boundary layer thermodynamics during 16 to 31 December 2025. Discover Geoscience, 4(1), Article 307. https://doi.org/10.1007/s44288-026-00684-2

Image Credits: AI Generated

DOI: 10.1007/s44288-026-00684-2

Keywords: fog, Indo-Gangetic Plain, boundary layer, temperature inversion, Richardson number, ventilation coefficient, radiosonde, ERA5 reanalysis, aviation meteorology, fog dissipation, Patna Airport, western disturbances

The problem boils down to geometry. When a cloud radar’s beam is perfectly vertical, the Doppler velocity it measures is the sum of just two things: the speed at which hydrometeors such as ice crystals and raindrops fall, and the vertical motion of the air itself. Separating those two contributions is already one of the fundamental challenges in cloud radar analysis, and decades of techniques have been built around Doppler spectra and their moments to tease them apart. But if the beam tilts even slightly away from zenith, a third term sneaks in: the horizontal wind, projected onto the tilted beam. A radar that thinks it is measuring falling snow may actually be measuring a slice of the jet stream.

The size of this contamination is startling given how small the tilts are. The researchers’ simulations show that a pointing offset of just 0.5 degrees, with a horizontal wind of 10 meters per second, injects a Doppler velocity signal of roughly 0.1 meters per second, comparable to the fall speeds of the smallest ice crystals. Crank the wind up to 40 meters per second, typical of upper-level jet conditions, and the bias grows to about 0.3 meters per second. Doubling the tilt to one degree roughly doubles the amplitude of the spurious signal, and a two-degree offset quadruples it. For radars designed to detect the feeble velocity perturbations of boundary-layer turbulence, these are not rounding errors; they are systematic biases that can propagate into retrievals of vertical air motion, turbulence, and cloud microphysics.

Min Deng of Brookhaven National Laboratory and her colleagues developed a technique they call the U-normalized velocity-direction display, or UN-VDD, which turns this contamination into a diagnostic. The key insight is that the wind-induced Doppler velocity does not vary randomly. It follows a cosine dependence on wind direction, peaking when the wind blows along the radar’s tilt azimuth and reversing when it blows the opposite way. This is the same geometric principle that underpins the velocity-azimuth display method used since 1968 to retrieve winds from scanning Doppler radars, but run in reverse: instead of assuming a vertical beam to find the wind, the team assumes a known wind to find the beam.

Normalizing is what makes the method work in practice. Because horizontal wind speed changes with height and time, raw Doppler velocities mix the geometric signal with variations in wind magnitude. Dividing the measured Doppler velocity by the radiosonde-measured wind speed strips that dependence away, leaving a quantity dominated by the beam-pointing geometry itself. Particle fall velocities and vertical air motions do not vary systematically with wind direction, so they appear as scatter and a roughly constant offset rather than a coherent cosine wave. The amplitude of the fitted cosine then yields the off-vertical tilt angle, and its phase gives the azimuth of the tilt. In upper-level ice clouds, where fall speeds are weak compared with synoptic-scale winds, the approximation is particularly clean.

The team put the method through its paces during the Cloud and Precipitation Experiment at kennaook (CAPE-k) in Tasmania, where a Ka-band ARM Zenith Radar (KAZR) and a Marine W-band ARM Cloud Radar (MWACR) operated within 100 meters of each other. A telling case from September 2024 showed the two radars observing the same ice cloud layer between 5 and 8 kilometers altitude, yet reporting systematically different Doppler velocities: the KAZR registered apparent upward velocities approaching 2 meters per second near cloud top, physically implausible as sustained air motion, while the MWACR showed weak downward values consistent with small ice crystals. Applying UN-VDD to the full campaign revealed why. The KAZR beam was tilted about 2.5 degrees from vertical, toward an azimuth near 140 degrees, while the MWACR was nearly true at about 0.6 degrees off zenith.

The most convincing validation came from subtracting one radar’s velocities from the other. Because the two instruments observe nearly the same atmospheric volume, the contributions of particle fall speed and vertical air motion largely cancel in the difference, leaving a signal dominated by their relative pointing error. The normalized velocity difference showed a strikingly clear cosine dependence with a near-zero offset, yielding a relative tilt of about 1.5 degrees at an azimuth of roughly 140 degrees, consistent with the individual radar fits. It is a neat piece of self-consistency: two independent radars, one analysis, one answer.

The method proved robust well beyond Tasmania. Applied to the Bankhead National Forest campaign in the southeastern United States, it retrieved a smaller offset of about 0.5 degrees, and sensitivity tests showed the retrieved angle varied by less than 0.3 degrees across different radar averaging windows and by less than 0.25 degrees across different reflectivity thresholds used to select ice-cloud samples. The overall uncertainty is estimated at roughly 0.5 degrees. Across recent ARM campaigns including SAIL, TRACER, EPCAPE, and CoURAGE, most deployments showed tilts within about half a degree, with CoURAGE showing a moderate 1.2-degree offset and CAPE-k the largest at 2.5 degrees for the KAZR. Notably, the method even caught history: at the Eastern North Atlantic fixed site, the analysis detected a growing pointing deviation during 2017 and 2018 that matched a documented data quality issue, and after a radar levelling procedure in January 2019, the retrieved offset dropped to about 0.1 degrees.

What makes the technique genuinely powerful is that it requires no new hardware and no special campaigns. Radiosondes are launched routinely at ARM sites and at weather stations worldwide, providing high-resolution wind profiles that are independent of radar calibration. The main caveat is balloon drift: a radiosonde can travel 5 to 30 kilometers horizontally by the time it reaches 10 kilometers altitude, introducing representativeness uncertainty in sheared environments. But because the wind-direction signature is geometric rather than local, aggregating many launches across a campaign builds up enough directional coverage to constrain the fit reliably.

The researchers also lay out how the retrieved angles could be used to correct archived data. For a small tilt, the maximum wind contamination is approximately the wind speed times the tilt angle in radians, so a 0.5-degree offset can bias velocities by up to 0.17 meters per second in 20-meter-per-second winds and 0.35 meters per second in jet-level winds. A correction formula removes the projected wind component, though the authors caution that corrections should only be applied when the tilt is statistically distinguishable from zero and the projected bias exceeds what the intended product can tolerate. For a global network of cloud radars feeding climate models and process studies, a quality-control tool that costs nothing but arithmetic on data already being collected may prove one of the quiet upgrades with the loudest consequences.

Subject of Research: Estimating beam pointing errors of vertically pointing cloud radars using radiosonde wind measurements

Article Title: Estimating beam pointing of vertically pointing cloud radars using radiosonde measurements

Article References: Deng, M., Giangrande, S. E., Theisen, A. K., Johnson, K. L., Lindenmaier, I. A., Wendler, T. G., Comstock, J., Rocque, M., Zhu, Z., & Matthews, A. (2026). Estimating beam pointing of vertically pointing cloud radars using radiosonde measurements. Atmospheric Measurement Techniques, 19(19), 6327-6339. https://doi.org/10.5194/amt-19-6327-2026

Image Credits: AI Generated

DOI: 10.5194/amt-19-6327-2026

Keywords: cloud radar, radiosonde, beam pointing, Doppler velocity, ARM, KAZR, MWACR, vertical air motion, atmospheric measurement, CAPE-k, radar calibration, wind projection

VITAL I, short for Vertical profiling of the troposphere: Innovation, optimization and application, Part I, took place in August 2024 at the Jülich Observatory for Cloud Evolution, or JOYCE, a National Facility of the ACTRIS European Research Infrastructure Consortium in western Germany. The campaign is a core activity of the Hans Ertel Centre for Weather Research, a joint initiative between German atmospheric research institutions and the German Meteorological Service, known by its German abbreviation DWD. That institutional pairing matters, because the campaign was not an academic exercise conducted in isolation from operational needs. The instruments under test are being considered for the next-generation DWD observational network, and the evaluation was carried out explicitly against the World Meteorological Organization’s Rolling Requirements Review, the international benchmark that defines how well a measurement system must perform before it is useful for short-term, high-resolution weather forecasting.

The technological lineup at the heart of the campaign reads like a catalogue of the tools many meteorologists believe will define the coming decades of atmospheric observation. On the ground, the researchers deployed Doppler lidars, which measure wind by detecting the Doppler shift of laser light scattered back from aerosol particles drifting with the air; water vapor lidars, which profile humidity by measuring the absorption and scattering of laser light by water molecules; and microwave radiometers, which infer temperature and humidity profiles from the faint microwave radiation the atmosphere emits at different frequencies. In the air, the team flew multi-copter uncrewed aerial systems, essentially instrumented drones capable of climbing through the boundary layer carrying sensors that measure temperature, humidity and wind directly along their flight paths.

What makes VITAL I scientifically valuable is that none of these instruments was evaluated in a vacuum. The two-week campaign produced vertical profiling datasets of temperature, humidity and wind that were compared against radiosondes, the established balloon-borne gold standard, and against measurements from a 120-meter meteorological tower at the site, which provides high-frequency in-situ readings at fixed heights. Comparisons were also carried out against satellite sounders, the spaceborne instruments that provide global but vertically coarse profiles of atmospheric temperature and moisture. This layered evaluation design allowed the researchers to ask a deceptively simple question with real operational consequences: when a new instrument disagrees with an old one, which one is right, and how much of the disagreement is a genuine error versus a difference in what each system can actually see?

The headline finding is encouraging for the future of ground-based profiling. Most of the tested ground-based systems were shown to fulfill the WMO breakthrough requirements, the ambitious performance thresholds the World Meteorological Organization has identified as targets that would transform forecasting capability if met. Equally important, the instruments proved capable of capturing the diurnal cycle of the atmospheric boundary layer, the full day-night evolution of the shallow, well-mixed layer of air near the surface that grows during daytime heating and collapses again after sunset. This cycle drives the formation of morning fog, the triggering of afternoon thunderstorms, and the dispersion or accumulation of pollutants, yet it unfolds largely between the twice-daily radiosonde launches, making continuous ground-based profiling a genuinely new observational capability rather than merely a cheaper substitute for balloons.

The technical challenges involved in this kind of intercomparison are considerable, and they explain why campaigns like VITAL I are needed at all. Each instrument family measures the atmosphere in a fundamentally different way. A microwave radiometer infers profiles indirectly from brightness temperatures and depends heavily on the retrieval algorithm that converts radiances into atmospheric states, with uncertainties that grow in cloudy or precipitation conditions. A Doppler lidar provides excellent wind information but needs sufficient aerosol loading to scatter light back, and its range can collapse in very clean air. Water vapor lidars offer high vertical resolution but demand careful calibration. Drones provide direct, high-accuracy in-situ measurements but sample one point in space at a time and are subject to aviation regulations and weather limits. Reconciling these different sampling geometries, resolutions and error characteristics against the tower, the radiosondes and the satellites is precisely the work that turns a collection of promising instruments into a validated observing system.

The campaign also had an educational dimension that its organizers built in from the start. Alongside the measurements, VITAL I encompassed an international summer school held simultaneously at JOYCE, designed to educate students from the bachelor to the doctoral level in modern methods of atmospheric profiling. This is more than a pleasant add-on. The transition from balloon-based to sensor- and laser-based observing networks will require a workforce fluent in radiative transfer, signal processing, data assimilation and the operational constraints of uncrewed aviation. Training the next generation of atmospheric scientists on live, cutting-edge instrumentation, while the instruments themselves are being validated, creates a pipeline of expertise that national weather services will need when these systems move from research sites into routine operations.

The data itself has been made publicly available through the research repository Zenodo, under the dataset title VITAL I: Vertical profiling of the troposphere: Innovation, optimization and application, Part I. Open release of campaign datasets of this kind has become a defining feature of modern atmospheric science, and it multiplies the value of the original effort. Model developers can use the co-located, multi-instrument observations to evaluate how well weather models represent boundary-layer structure. Retrieval scientists can test new algorithms against independent reference measurements. Instrument engineers can quantify the strengths and weaknesses of competing systems. And national meteorological services beyond Germany, all of which face the same questions about modernizing their observation networks, can draw on the evaluation results without repeating the entire campaign themselves.

The broader context is a worldwide rethink of how the lower atmosphere should be observed. Radiosonde networks provide indispensable accuracy and vertical detail, but their cost and logistics limit their temporal and spatial coverage, and the boundary layer’s rapid evolution between launches is a persistent blind spot for high-resolution forecasting. Satellites provide continuity and global reach but struggle with vertical resolution in the lowest layers, where temperature and humidity gradients are sharpest and where clouds interfere most. The emerging answer, which VITAL I was designed to stress-test, is a hybrid architecture: dense networks of continuous, ground-based active and passive remote sensors, augmented by targeted drone flights and constrained by satellite observations, all feeding data assimilation systems that can exploit high-frequency, high-resolution profiles. The finding that most of the ground-based systems tested at JOYCE meet WMO breakthrough requirements is a concrete step toward showing that such an architecture is technically feasible, not just conceptually appealing.

For the German Meteorological Service, the path from campaign to network runs through further evaluation, optimization and, ultimately, procurement decisions, and the VITAL campaign’s title promises a Part II. For forecasters and the public, the payoff would arrive in the form of better short-range forecasts of the phenomena that matter most at ground level: fog that closes airports, convective storms that flood streets, and the low-level winds that spread wildfire smoke or carry wind energy to the grid. The VITAL I dataset, gathered over two intensive weeks at a single observatory, demonstrates that the instruments to fill the boundary-layer gap already exist, that they can meet international performance standards, and that the community of scientists trained to use them is being built at the same time. What remains is the institutional work of turning a validated research dataset into an operational observing network, a transformation that, if it succeeds, will quietly reshape the accuracy of every forecast that begins with a measurement of the air above us.

Subject of Research: Evaluation of novel ground-based and uncrewed aerial profiling instruments for measuring temperature, humidity and wind in the atmospheric boundary layer

Article Title: Vertical profiling of the troposphere: Innovation, optimization and application, Part I (VITAL I)

Article References: Vertical profiling of the troposphere: Innovation, optimization and application, Part I (VITAL I). (n.d.). https://doi.org/10.5194/essd-2026-784

Image Credits: AI Generated

DOI: 10.5194/essd-2026-784

Keywords: atmospheric boundary layer, vertical profiling, Doppler lidar, microwave radiometer, water vapor lidar, uncrewed aerial systems, radiosonde, weather forecasting, WMO Rolling Requirements Review, JOYCE observatory, German Meteorological Service, Earth System Science Data

The study, carried out by Pravat Rabi Naskar of the Meteorological Watch Office Kolkata at the India Meteorological Department and Gyan Prakash Singh of the Department of Geophysics at Banaras Hindu University, draws on radiosonde and radiowind observations collected at 00 UTC for every winter season from December 1990 through February 2022. These balloon-borne instruments, launched routinely to profile temperature, humidity and wind through the depth of the troposphere, provide one of the longest continuous records of vertical atmospheric structure available anywhere in the region. The authors supplemented this archive with ERA5 reanalysis vertical velocity data from the Copernicus Climate Change Service, allowing them to connect the observed inversion behavior to the large-scale motions of the atmosphere above the Gangetic delta.

Thermal inversions come in several flavors, and distinguishing them matters enormously for forecasting. A surface-based inversion forms on clear, calm nights when the ground radiates heat to space and chills the air in contact with it. A subsidence inversion, by contrast, forms aloft when air in a high-pressure system sinks slowly from above; as the descending air compresses, it warms at a rate faster than the environmental lapse rate, creating a warm layer that floats above the cooler air below. The Kolkata analysis found that the winter inversions between the 1000 and 500 hectopascal pressure levels — a layer stretching from near the surface to roughly five kilometers altitude — are overwhelmingly of the subsidence type, pointing to the dominant role of large-scale atmospheric descent rather than purely local nighttime cooling.

One of the most striking trends the researchers documented concerns multiplicity. The percentage of winter days on which more than one thermal inversion appears within that 1000 to 500 hectopascal column is increasing at a rate of approximately one percent per year. Multiple stacked inversion layers are a well-known signature of complex vertical structure in the lower troposphere, and studies from eastern China and northwest China have linked such multilayer configurations to severe haze episodes, because each warm layer acts as a separate barrier to vertical mixing. Over Kolkata, the steady rise in double-inversion days suggests that the lower troposphere is becoming increasingly stratified in winter, a structural change that could compound the region’s already serious wintertime pollution problems.

The geometry of the inversions is changing too, and in a consistent direction. Over the 33-winter record, the height of the inversion top has been dropping at roughly 28 meters per year, while the base of the inversion has descended at about 20 meters per year. Because the top is falling faster than the base, the inversion layer itself is thinning by approximately 7 meters annually. A thinner inversion layer is not a weaker one; on the contrary, compressing the temperature contrast into a narrower vertical span can sharpen the stability of the layer, making it a more effective lid on the atmosphere beneath. The authors also found that inversion intensity — the temperature difference across the inversion — is increasing, although the inversion gradient, the temperature change per unit height, shows no statistically significant trend, presumably because the intensifying temperature jump is being spread across a shrinking layer.

What physical mechanism could drive this systematic downward migration and intensification? The most likely candidate is a change in the large-scale subsidence pattern over the region, which the ERA5 vertical velocity data were brought in to examine. Sinking motion in the mid-troposphere warms the air through compression, and the altitude at which that warming is strongest determines where the inversion forms. If the descending branch of the regional circulation has shifted lower, or if subsidence has strengthened in the layers just above the boundary layer, the inversion would be expected to form closer to the ground and to grow warmer relative to the air beneath it. Such changes in atmospheric stagnation have been flagged in recent work on Indian air pollution, including studies linking circulation-induced stagnation changes to surface PM2.5 concentrations across the subcontinent.

Perhaps the most consequential finding of the study concerns the planetary boundary layer — the shallow, turbulent skin of the atmosphere that responds directly to the surface and within which nearly all human emissions of pollutants are first released. The researchers found that the planetary boundary layer height over Kolkata airport sits consistently below the base of the subsidence inversion, and that it has been decreasing significantly at a rate of around 5.5 meters per year over the study period. This relationship is physically coherent: a subsidence inversion whose base lies just above the boundary layer suppresses the turbulent entrainment that normally allows the boundary layer to grow during the day. As the inversion base sinks, it effectively presses down on the boundary layer, compressing the volume of air into which vehicle exhaust, industrial emissions and dust are injected.

The implications for air quality are direct and sobering. A lower boundary layer means a smaller mixing volume, so the same quantity of emissions produces higher surface concentrations. A stronger, thinner inversion overhead makes that confinement even more effective by preventing any exchange of air across the stable layer. Research from around the world — from the Sichuan Basin in China to Hanoi, Istanbul, Tehran and the Canary Islands — has repeatedly documented the tight coupling between inversion characteristics and particulate pollution, and a recent national cohort study in China even linked thermal inversion exposure to increased risk of incident cardiovascular disease. For Kolkata, a megacity whose winter air already ranks among the most polluted anywhere, a trend toward stronger and lower inversions represents a slow tightening of the atmospheric vise.

Fog and aviation are equally implicated. The same authors previously documented a dense fog episode over Kolkata airport in January 2025 and traced its synoptic origins, and the present study provides the climatological backdrop against which such events unfold. Inversions trap moisture near the surface overnight, and shallow, intense inversions favor the radiative cooling and moisture convergence that produce dense fog — the single most disruptive weather phenomenon for airport operations in northern and eastern India. A boundary layer that is steadily losing height, capped by an inversion that is steadily losing altitude and gaining strength, is a boundary layer primed for more frequent and more persistent fog, with cascading effects on visibility, flight delays and safety margins during the December-to-February peak travel season.

What makes this study particularly valuable is its length and its focus on a single, operationally critical location. Thirty-three winters of consistent 00 UTC soundings from one site remove much of the noise that afflicts shorter or geographically scattered records, and the airport setting ties the results directly to a real-world stakeholder. The picture that emerges is of a lower troposphere in structural transition over the Bengal delta: inversions multiplying, sinking, thinning and strengthening, while the boundary layer beneath them contracts year by year. Whether these trends are driven by shifting regional circulation, by warming sea surface temperatures in the Bay of Bengal, by aerosol-induced changes in the radiation balance, or by some combination of all three remains an open question — but the direction of change is unambiguous, and it points toward winters in which the air over Kolkata becomes steadily harder for the atmosphere to flush clean.

Subject of Research: Long-term trends in winter thermal inversion characteristics and planetary boundary layer height over Kolkata airport

Article Title: Characteristics of winter thermal inversions over Kolkata airport

Article References: Naskar, P. R., & Singh, G. P. (2026). Characteristics of winter thermal inversions over Kolkata airport. Theoretical and Applied Climatology, 157(10), Article 645. https://doi.org/10.1007/s00704-026-06575-3

Image Credits: AI Generated

DOI: 10.1007/s00704-026-06575-3

Keywords: thermal inversion, Kolkata airport, radiosonde, planetary boundary layer, subsidence, air pollution, fog, ERA5 reanalysis, winter climate, atmospheric stability, India, boundary layer height