Dancing serpentine columns of dust tear through the surface of Mars almost daily, leaving behind dark, filamentous trails that are easily recognizable from orbit by both experts and enthusiasts of the Red Planet. A new comprehensive analysis of two European orbiters shows that these vortices are not just an exotic pastime of the wind, but a fundamental link in the Martian climate system. By tracking them over twenty years, planetologists have reconstructed how dust rises, how it travels and where it most often settles, and in doing so, they have also recorded a surprising fact: surface winds on Mars are often faster than what previous models and measurements from the surface suggested. This is the most extensive tracking of “dust devils” to date, with a systematically built catalog that combines observations from different seasons, areas and geological environments, and offers concrete guidelines for planning future missions that will have to work on Mars in an environment of unpredictable gusts and persistent, sticky dust.

What are “dust devils” and why are they important

Dust vortices, in popular terminology “dust devils”, are created when the ground heats up faster than the air immediately above it due to solar radiation. The warm air begins to rise abruptly and creates an updraft column, while colder, denser air flows in from the side and closes the circulation. A sufficient horizontal component of the wind turns this upward flow into a rotating funnel which, in the dry and dusty conditions of Mars, very easily picks up fine particles and lifts them tens or even hundreds of meters high. The vortices are transient: they last from just a few minutes to, in rare cases, a little longer; but their combined effect on the atmosphere is surprisingly large because they occur very frequently, over large areas and during parts of the day when the surface is warmest.

Unlike Earth, where rain and moisture quickly “wash away” floating particles, on Mars, dust can remain in the atmosphere for a long time and be transported over thousands of kilometers. This affects the distribution of temperature (it weakens heating during the day and “covers” the surface at night), the formation of clouds and the water balance, because vortices and storms can accelerate the loss of water vapor into space. In practice, this means that understanding small, local vortices leads to more accurate global climate models and safer planning of space missions – from scheduling self-cleaning of solar panels to assessing the risk to optics and mechanisms on rovers and landers.

Two spacecraft, twenty years and the first global movement catalog

A team of researchers analyzed archives of images taken by Mars Express (in orbit since 2004) and ExoMars Trace Gas Orbiter (in orbit since 2016). Instead of manual counting, a computer approach was applied: a deep learning model was trained that automatically recognizes the distinctive “signature” of active vortices in images – a brighter cloud of elevated dust and a dark shadow it casts on the ground – and extracts their geometries and dynamics from a multitude of other visual patterns. The result is a publicly available list with 1039 active vortices, for which the direction and speed of movement across the ground were also estimated for a large number, which allowed the researchers to “draw” wind fields on a planetary scale.

The mapping shows that the vortices are spread almost everywhere – from lowlands to volcanic slopes – but they often “originate” in specific source zones with an abundance of fine material. One such is Amazonis Planitia, a vast area of the northern lowlands covered with fine sand and dust, where frequent vortices literally “suck up” thousands of tons of particles and send them into the air. The sample of sizes is diverse: vortices with a diameter of only a few tens of meters were recorded, as well as those with a width of several hundred meters, and their heights sometimes exceed everything we would expect by observing only from the ground. This spatial diversity is crucial for understanding the transport of dust and its deposition in different geomorphological niches of Mars.

Faster winds than we thought

By tracking the movement of vortices between multi-channel images, researchers directly measured surface winds reaching up to about 44 m/s, or approximately 158 km/h. This is significantly higher than typical measurements from previous landers and rovers, whose observations are, understandably, limited to individual locations and time periods. Although these numbers sound intimidating, it should be kept in mind that the thin atmosphere of Mars (about 1% of Earth's density) also means a drastically lower dynamic pressure – a human on Mars would barely feel such a “hurricane” wind on their skin, but at the same time it is quite sufficient to lift and transport dust over large areas and block sunlight for rovers that rely on solar panels.

Surprisingly high speeds were recorded in more areas and times of the year than existing models suggested. This further clarifies the so-called “mystery of sand transport on Mars”: laboratory thresholds for moving sand grains – a skipping motion that we call saltation – are often higher than the speeds measured by instruments on the ground. New data from orbit show that fast winds are actually common, which means that dust lifting is likely more frequent than we previously assumed. This has direct consequences for estimates of visibility, heating and cooling, as well as for models of the “lifespan” of dust clouds in the atmosphere.

When and where Mars is “dustiest”

Seasonality is pronounced: most activity occurs in the spring and summer of each hemisphere, during the late morning and early afternoon, approximately between 11 a.m. and 2 p.m. local solar time. This is when the temperature contrast between the ground and the air is strongest, and the surface winds are stable and strong enough to “power” the rotation. In lowland areas with fine sediment – especially in the northern plains – the frequency of vortices is higher, but they have also been recorded on steep volcanic slopes and along the edges of craters, where the relief further intensifies the flow. Such a distribution is reminiscent of patterns in deserts on Earth, but Mars, due to its dryness and thin atmosphere, shows a wider spatial and temporal “loop” of activity.

How “noise” became data

Neither Mars Express nor ExoMars TGO were originally designed to measure wind speed. The key is in the way their cameras – HRSC on Mars Express and CaSSIS on TGO – assemble one final image from several separate channels. Since the channels are recorded with a time lag of one to several tens of seconds, anything that moves in the scene leaves subtle shifts in color or geometry. The researchers turned this “artifact” into a meter: by measuring the shift of the dust cloud between channels, they obtained the speed and direction of the vortex's movement and, indirectly, the speed and direction of the wind at the ground. This is an elegant example of how engineering limitations (channel delay) can become a scientific opportunity.

Automatic vortex recognition is based on neural networks trained on thousands of labeled examples. The model recognizes the characteristic contrast of a brighter cloud of elevated dust and a dark shadow, circular and spiral patterns that differ from clouds, drifts or hill structures, as well as the spatial relationships between these elements. Such an approach allows for a systematic “combing” of planetary archives, and its effect is multiplied as fresh images arrive and as the models are further refined. Ultimately, this means that the wind map will become increasingly detailed over time, and predictions of the local microclimate will be more reliable.

What the discoveries mean for future spacecraft and rovers

More detailed wind maps by region and season directly help in choosing landing sites and planning the work of spacecraft on the surface. The amount of dust that settles on solar panels and instrument optics is one of the main threats to the energy and scientific budget of missions. If we know in advance how often fine dust “falls” in a certain area, we can predict the schedule for self-cleaning, the orientation of cameras, the planning of critical measurements during the hours when the clouding is minimal, and better dimension the filters, seals and moving assemblies exposed to abrasion.

In practice, such insights are already being incorporated into preparations for the next European steps on the Red Planet. The Rosalind Franklin rover is planned for launch in 2028, with a landing in 2030, with the goal of drilling to a depth of about two meters for the first time and searching for chemical traces of ancient life in protected, underground layers. A new European platform with precise retro-rockets and advanced parachutes is being built for a safe descent; in addition, new landing plans and the schedule of operations on the surface already take into account the finer map of winds, dust and seasonal windows of favorable weather.

Rovers as “meteorologists”: lessons from the ground

In addition to data from orbit, important clues are provided by observations from the ground: NASA's rovers regularly record transient vortices, sometimes more than one at a time, which provides a local reference for the directions and changes in wind speed. Such scenes show how vortices can interact – merge or “swallow” each other – and confirm that they are one of the main sources of dust in the atmosphere. By combining local observations and the global catalog, a coherent picture of dust circulation on different scales emerges, from a grain of sand hopping across the ground to planetary currents carrying dust clouds across the continents of Mars.

From local meteorology to global climate

One of the most important benefits of the new study is the calibration of climate and forecast models of Mars. Until now, they often “guessed” wind speeds from indirect measurements of temperature and pressure, with very few direct confirmations on a wide spatial scale. Now, for the first time, a set of measurements is available that covers various geographical latitudes, altitudes and seasons. Models that systematically underestimate wind speed in certain regions will have to be revised, and this will consequently change estimates of how much dust circulates, when clouds form and what the consequences are for the energy balance of the atmosphere and surface.

Practical implications: energy, optics, logistics

For missions powered by solar panels, dust deposition is a matter of life and death. Some spacecraft ended their mission when a dense “blanket” obscured the panels to the point where there was no longer enough power for communication or heating. Planning the schedule for self-cleaning – from shaking and rotating the panels to targeting “windy windows” when natural scouring is more frequent – is only possible when we understand the cycles of vortices and local wind roses. Optical instruments, from cameras to laser spectrometers, also suffer the consequences: fine dust settles on optical surfaces and changes their reflection spectra, so new strategies for protection and calibration are also needed.

For more complex operations on the ground, logistics are also important: the schedule for driving, drilling and sampling must be coordinated with local meteorological signals. If an increased “dusty” background is expected during the most active hours of the day, it makes sense to shift sensitive tasks to the earlier morning or later afternoon hours. The assessment of the risk of static electricity – which can affect instruments and communication – also changes with the seasonal and daily patterns of vortices, while mechanical assemblies (joints, gears, seals) are dimensioned bearing in mind the abrasive properties of the minerals in the floating particles.

Open data and advanced processing

A special value of the project is that the catalog is prepared as an open data set that is gradually supplemented with new entries. As Mars Express and TGO continue to record daily, the observation network becomes denser, and the neural networks are further refined. This creates a positive loop: better detections feed the models, and better models help find “hot spots” and times of day when there is the highest probability that the camera will capture a transient vortex. In the coming months, targeted imaging campaigns are also expected to lead to a comparison of measurements of the same vortices from different platforms in order to further reduce errors in estimating speed and direction.

Why it is important to update hours and seasons — today's date and current dynamics

As today is October 9, 2025, the northern hemisphere of Mars is entering a calmer part of the year, when the summer maximum of vortex activity is gradually subsiding. However, new measurements of higher surface speeds suggest that the transport of dust will continue in the transitional period, especially along the edges of plains and on steep slopes where topography encourages flow. In the coming weeks, the first updated parameters for numerical models are expected to be published, which will take into account the newly set limitations of wind speeds in the ground layer, which will enable more precise forecasts of visibility and dust deposition on instrument surfaces.

Comparisons with helicopter and surface measurements

The Ingenuity helicopter, which has performed a number of historical flights this year, occasionally recorded wind gusts stronger than expected, which indicated that short-term bursts can exceed the “smooth” average values provided by meteorological stations. The new orbital catalogs provide context: strong wind bursts do not occur in isolation, but in belts that follow thermal gradients and relief channels. Together, orbital and surface measurements form the backbone of a new, synoptic meteorology of Mars, in which small scales (vortices) and large scales (regional currents) finally sit on the same “wind map”.

Where to go next: focused imaging campaigns

Knowing at what hours and in which seasons vortices most often form, camera teams can plan sequences with a maximum time between channels to get a larger “baseline” for measuring the movement of dust clouds. Coordinated imaging of the same vortex from two spacecraft – from different geometries and with different time lags between channels – will enable the validation of methods and the reduction of errors. The combination of stereo pairs and multi-spectral data will further open up the possibility of estimating the vertical structure of the dust clouds and the size of the particles, which is crucial for modeling the optical thickness of the atmosphere and energy flows above different types of terrain.

Context: history and legacy of European missions

Since the arrival of Mars Express at the end of 2003, Europe has been systematically building a unique photographic record of the Red Planet. The longevity of the mission and the careful calibration of the instruments have enabled comparisons ranging from days to decades – a rare privilege in planetary science. ExoMars TGO, for its part, brought high sensitivity for trace gases and the modern CaSSIS camera, which – along with the skill of creative teams – made this new way of “wind-graphing” Mars possible. Together, these missions show how much science can be extracted from the “side effects” of instruments when the perspective is changed and a signal is sought in seemingly random noise.

Industrial mosaic and European-American partnership

The recovery of the ExoMars program after the termination of cooperation with Russia in 2022 brought a new division of roles: European companies took over a larger part of the system, while the United States will provide the launch vehicle and some critical subsystems. Airbus in the United Kingdom is in charge of building the landing platform, with a contract of about 150 million pounds, and the key goal is precise braking and a soft landing of the rover in 2030. In planning these scenarios, new wind fields obtained from the vortex catalog are also being used, which reduces uncertainties in the design of parachutes, retro-rockets and the construction of the platform legs.

Measurement technique: from shadow to speed

With HRSC, up to nine channels record the scene with a time lag of approximately 7 to 19 seconds. In this interval, the vortex moves enough that a discrete color shift or a “ghost” that reveals the movement appears on the combined image. CaSSIS records pairs with a one-second interval for color and about 46 seconds for stereo; this makes it possible to capture larger movements, but sensitivity to very short-term vibrations or accelerations is lost. In both cameras, geometric reconciliation of the channels and precise correction of distortions are crucial to distinguish artifacts from the actual movement of the dust cloud, and additional validations are achieved by comparing with sequences of images from other orbits and, where possible, with on-site videos from rovers.

Risk maps and landing site selection

In standard mission practice, average and extreme wind speeds are included in so-called risk maps. Comparisons between the vortex catalog and topographical maps – for example, the edges of craters or “bottlenecks” between plateaus – indicate corridors of increased flow, where a small pressure drop and changes in relief concentrate gusts. This knowledge is used to optimize the geometry of parachutes, dimension the legs of the platform and define tolerances for lateral wind impacts. The daily cycles are equally important: landings are, where possible, planned outside the strongest daily maximums of vortex activity in order to reduce the chances of uncontrolled lateral impacts and the raising of dust during contact with the ground.

What the archives still hide

The archives of Mars Express and TGO are rich in sequences in which, with robotic eyes focused on geology, meteorological gems are “accidentally” hidden. Systematic retroactive searching, accompanied by continuous training of algorithms, will likely increase the number of recorded vortices and improve spatial coverage. In addition, new programs of coordinated imaging of the same scenes a few days and weeks apart should shed light on how the “paths” of the vortices change throughout the season and how long they remain visible on the ground as dark, sinuous lines. This, in combination with in situ experiments, could also be used to determine the erosion thresholds of different types of soil under the action of short-term but frequent wind impacts.

Mars's dust as a resource and a challenge

Although dust has for years been seen primarily as a nuisance, there are more and more proposals to view this ubiquitous material also as a resource. Knowledge of the granulometry and electrical properties of the particles, in combination with wind maps, could help in the development of passive catchers that “harvest” dust for in situ experimental analyses. This is also partly a matter of the safety of future astronauts: fine dust can be abrasive and electrostatic, so the design of filters, seals and suits must take into account locally expected wind speeds and typical particle sizes. Every new pixel in the wind map thus also becomes a new item in the risk budget of future crews.

How the media reports new findings and why the numbers need to be contextualized

Speeds of 158 km/h sound dramatic, but without context they can be misleading. On Mars, the thin atmosphere means that such a wind carries significantly less energy than a storm of the same speed on Earth. However, it is precisely this low density that is the reason why fine particles, once lifted, stay in the air for a long time and travel long distances. The scientists' message is more nuanced than a “record”: what is important is that for the first time we have obtained a reliable, spatially widespread picture of surface winds that improves models and helps avoid oversights in mission planning – from air currents during landing to strategies for maintaining the energy and scientific capacity on the ground.

Data availability on October 9, 2025

Since both orbiters are still active today, fresh images arrive daily. This also expands the vortex catalog, which is intended for a wide community – from climatologists and geologists to engineers and mission planners. It is expected that by the end of this Martian year, enough unified records will be available for the first series of improvements to global models, including finer parameterization of dust sources, seasonal weather windows for landing and estimates of the energy budgets of missions that depend on the Sun.