Mars' Winds Can Reach ~100 mph — Deep-Learning Study Maps Global Gusts

How strong are the winds on Mars?

Mars may look calm from afar, but its thin red atmosphere can produce surprisingly powerful near-surface winds. A recent study led by Dr. Valentin Bickel at the University of Bern used deep learning on orbital images to reveal gusts and dust-devil winds reaching roughly 44 m/s (≈100 mph) in many regions — far stronger and more widespread than earlier rover-based measurements suggested.

How the study worked

The team applied a neural network to more than 50,000 images from two orbital instruments: CaSSIS aboard the ExoMars Trace Gas Orbiter and the HRSC camera on Mars Express. The algorithm automatically identified dust devils and their surface tracks. From those detections, researchers selected about 300 high-quality stereo-image sequences to track dust-devil motion, estimate near-surface wind velocities, and map wind directions across broad areas.

Key findings

• Near-surface winds associated with dust devils can reach up to ~44 m/s (≈100 mph).
• These high wind speeds were observed across wide swaths of Mars, not only at isolated sites.
• Earlier in-situ rover measurements typically reported lower speeds (mostly under ~30 mph, with occasional peaks near 60 mph), indicating orbital surveys reveal a more complete global picture.

Why this matters

Winds on Mars are a primary driver of dust lifting. Airborne dust affects how much sunlight reaches the surface, redistributes heat in the atmosphere, and influences storm development. Greater dust lifting changes the planet's radiative balance and weather patterns and can produce hazards for spacecraft.

Operational impact: Stronger and more frequent winds increase the risk of dust accumulation on solar panels and instruments, complicate landing dynamics, and influence long-term hardware durability — factors mission planners must consider for robotic and human missions.

Practical applications and future work

The machine-learning approach provides a new, data-driven way to build global wind maps from orbit. These wind catalogs can inform landing-site selection, descent and touchdown models, dust mitigation strategies for solar-powered systems, and climate-model inputs. As more stereo imagery and targeted observations are collected, the wind maps should gain higher spatial resolution and accuracy, improving predictions of surface conditions for future missions.

Example of historical relevance: The Opportunity rover’s long-term operations were severely affected by a global dust storm in 2018, demonstrating how airborne dust can endanger surface missions by blocking sunlight.

Note: The study demonstrates the value of combining modern machine-learning tools with long-running orbital image archives to reveal dynamic processes on Mars that are difficult to capture with localized lander and rover measurements alone.