Pressure readings from the InSight lander indicate that a near-daily atmospheric wave called K1 appeared only intermittently, but reached about 8 pascals during the C34 dust storm. Clearer seasonal detections were about 1 pascal in late northern spring and roughly 1 to 2 pascals in northern autumn; during C34, two other extracted pressure components, S1 and S2, reached about 39 and 13 pascals.
Separating the near-daily rhythms
To separate the overlapping patterns, the researchers applied singular spectrum analysis, a method for teasing apart repeating signals in a time series, to InSight pressure intervals longer than 10 sols. The record covered InSight sols 15 to 824 and was divided into 99 continuous intervals; K1 was detected in 17 of them. The analysis retained the first 30 reconstructed components and estimated their frequencies with a Fourier transform.
The signal was seasonal
The pattern varied with season. K1 was mainly detected in late northern spring and northern autumn. Significant summer detections were absent, but the authors attribute that partly to the difficulty of separating K1 when its frequency is close to the near-daily signal and to the short record.
A shift during the C34 storm
During C34, K1 and S1 changed their relative phase. While S1 shifted toward earlier local times, the two signals were in phase, meaning their cycles lined up; during S1's return toward its original phase, K1 was approximately out of phase.
The study also found a temporary change in S1's apparent period, measured from the interval between successive local peaks. It fell below 24 Mars hours starting at sol 42, reached about 23.5 Mars hours at sol 45, rose to about 24.5 hours at sol 51, and returned to around 24 hours after sol 54.
Dust timing and a possible resonance
Dust records showed their zonal wavenumber-1 and -2 components, measures of uneven loading around Mars, peaking around sols 46 to 47, before global-mean dust peaked around sols 48 to 51. That period of high nonuniformity coincided with a larger S1 amplitude and an apparent S1 period shorter than 24 Mars hours.
The authors read the phase alignment and shortened S1 period as consistent with resonance, an interaction in which linked rhythms can reinforce one another, between K1 and DE1, the eastward nonmigrating diurnal tide. They propose that transient zonally uneven dust loading may be associated with amplification of K1 and DE1 when their phases align. The study does not establish that dust caused the amplification.
The model matched only part of the picture
The Mars Planetary Climate Model gave a partial match. Simulations showed clear K1 signals during both C-storm periods, with in-phase enhancement followed by rapid decay, but they did not clearly reproduce the observed shortening of S1's period. The modeled K1 frequency was about 1.02 cycles per sol, compared with about 1.1 cycles per sol in the InSight data; the authors say the reason for the mismatch remains unclear.
A signal, not a cause
The analysis is observational, so its timing patterns show association rather than causation: they do not establish that dust caused K1 or DE1 to strengthen. K1 detection also depends on interval length and frequency separation, and no formal uncertainty estimate was reported. For that reason, the lack of a significant summer detection does not establish that the wave was absent.
Paper data and sources
Original title: Quasi-diurnal Kelvin Wave Observed in the Atmosphere of Mars by the Pressure Sensor on the InSight Lander
Authors: Anzu Asumi, Jorge Hernández-Bernal, Kaoru Sato, Aymeric Spiga
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text