Diagram of the Insight lander and graphs from the study.
(Left) Artist's Concept of NASA’s InSight Lander on Mars.(Right) Comparison of wind speed throughout the InSight mission measured by TWINS (top; first 750 sols), and retrieved from seismic data using Machine Learning (bottom; first 1400 sols). The Y-axis denotes the local true solar time (LTST) while the X-axis denotes the mission day (sol) and solar longitude (Ls). The color denotes the wind speed and saturated colors indicate missing data. Credit: NASA/JPL-Caltech (left) and Stott et al. [2025], Figure 4 (right, modified)
Editors’ Highlights are summaries of recent papers by AGU’s journal editors.
Source: Journal of Geophysical Research: Planets

Despite providing critical insights into atmospheric dynamics and weather patterns, wind observations on the surface of Mars remain relatively rare. The Temperature and Wind for InSight (TWINS) instrument onboard NASA’s Insight mission was designed to measure wind speed and direction winds. However, due to power constraints caused by increasing dust accumulation on InSight’s solar panels, TWINS primarily operated during the first 750 Martian days (sols) of the mission. In contrast, the Seismic Experiment for Interior Structure (SEIS) instrument operated almost continuously until the mission’s final transmission on Martian day 1440.

Since winds are the dominant source of energy in the seismic data, Stott et al. [2025] developed a machine learning model, WindSightNet, to map seismic data to wind speed and direction, nearly doubling the coverage of TWINS. The authors find an overall good agreement between both datasets during the first 750 sols, increasing confidence in WindSightNet data for the remaining Martian Days. Using this validated dataset, the authors analyze the interannual (one year on Mars is 669 sols) variability of wind speed and direction, as well as large-scale weather patterns and the height of the lower atmosphere throughout the Insight mission.

This dataset delivers a precious long-term and continuous record of Martian winds for the atmospheric community to refine their atmospheric models and better understand how dust is lifted on Mars. While the approach by the authors cannot capture the fastest wind variations or highest wind speeds recorded by TWINS due to a lower sampling rate, nor accurately predict wind speeds near 0 meters per second due to SEIS’s noise level, this study opens new possibilities for planetary instrumentation.

Citation : Stott, A. E., Garcia, R. F., Murdoch, N., Mimoun, D., Drilleau, M., Newman, C., et al. (2025). WindSightNet: The inter-annual variability of martian winds retrieved from InSight’s seismic data with machine learning. Journal of Geophysical Research: Planets, 130, e2024JE008695. https://doi.org/10.1029/2024JE008695   

—Germán Martínez, Associate Editor; and Beatriz Sánchez-Cano, Editor, JGR: Planets

Text © 2024. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.