In situ recording of Mars soundscape

Before Perseverance, no pressure fluctuations on Mars had been monitored above 20 Hz (the acoustic domain). Given Mars’ low surface pressure (~0.6 kPa) and CO2-dominated atmosphere, theory predicted lower sound levels than on Earth, a near-surface sound speed around 240 m s−1, strong high-frequency attenuation, and frequency-dependent dispersion of the sound speed. Yet, models diverged because of limited low-pressure experimental data and challenges in characterizing attenuation and turbulence in closed environments. Recording sounds with SuperCam’s microphone (up to 100 kHz sampling) enables probing wind, turbulence, and acoustic wave propagation at unprecedented, centimetre-to-metre scales to test these predictions and characterize Mars’ acoustic environment.

Prior work modeled acoustic propagation and attenuation in CO2 atmospheres and discussed expected turbulence spectra on Mars, but with large discrepancies, especially below 1 kHz (for example, Petculescu and Lueptow 2007; Bass and Chambers 2001; Williams 2001). Boundary layer structure and turbulence on Mars have been studied using meteorological packages (Viking, Curiosity, InSight) and models (Savijarvi 1991; Petrosyan et al. 2011; Martinez et al. 2017; Banfield et al. 2020), but lacked high-frequency pressure fluctuation data. Acoustic relaxation theory indicates that vibrational modes in CO2 produce frequency-dependent adiabatic indices and sound speeds (Bass et al. 2007), but in situ validation at Martian surface conditions was missing. Experimental microphone work under simulated Martian atmospheres informed sensor response and wind noise (Chide et al. 2021; Murdoch et al. 2019) yet did not capture real surface propagation or turbulence in situ.

• Instrumentation: The Perseverance rover’s SuperCam instrument carries an electret microphone capable of recording pressure fluctuations from 20 Hz up to 12.5 kHz or 50 kHz (sampling at 25 kHz or 100 kHz). Environmental context (pressure, temperature, wind) was provided by the MEDA suite.
• Acoustic datasets: Multiple sols’ recordings were analyzed. A representative dataset (Sol 38b) was examined via time series and spectrograms to identify wind-driven acoustic fluctuations and bursts up to ~300 Hz.
• Cross-correlation with meteorology: Microphone time–frequency content was compared with MEDA wind speeds to assess correlations between acoustic intensity and wind variability, including short gusts (≈10 s timescales). Power spectral densities (PSDs) of microphone, MEDA pressure, and MEDA wind were computed over overlapping intervals to identify spectral regimes and slopes and to locate the transition from shear-dominated turbulence to dissipative behavior.
• Point acoustic sources: Two controlled/known sources were used to measure sound speed and attenuation.
  • Laser-Induced Breakdown Spectroscopy (LIBS) sparks from SuperCam provided pulsed broadband signals above ~2 kHz (frequencies f > relaxation frequency), enabling repeated local sound speed estimations via time-of-flight between the spark and the microphone and retrieval of frequency-dependent attenuation with distance (>2 kHz).
  • The Ingenuity helicopter’s blade passage frequency (~84 Hz) generated a quasi-harmonic source at f < relaxation frequency. The received tone exhibited Doppler shifts modulated by changing range and wind; fitting the Doppler curve yielded the effective sound speed at low frequency. MEDA wind along the line of sight was used to correct for advection.
• Thermodynamic comparison: Sound speeds predicted from temperatures measured by MEDA at different heights (0.85 m, 1.45 m) and by the Mars Climate Database at the surface and 2 m were computed using appropriate adiabatic indices (γ = 1.4 for f > f_relax; γ = 9/7 for f < f_relax) for comparison with acoustic measurements.
• Spectral interpretation: Relaxation theory was applied to interpret the observed sound speed dispersion, with relaxation frequency scaling with pressure, placing the vibrational relaxation crossover near ~240 Hz at ~0.6 kPa. Acoustic attenuation coefficients above 2 kHz were derived and compared to propagation models including CO2 vibrational relaxation losses.

• First in situ acoustic characterization of Mars from 20 Hz to 50 kHz, extending pressure variation measurements to 1,000× smaller spatial–temporal scales than previously observed and revealing a dissipative regime spanning five orders of magnitude in energy.
• Clear correlation between microphone-recorded acoustic intensity and MEDA wind speed; short, intense gusts are resolved on ~10 s timescales.
• Spectral regime change: PSD slopes indicate a transition from shear-dominated turbulence to a dissipative regime between ~1 and 20 Hz in the examined dataset (Sol 38b).
• Frequency-dependent sound speed (dispersion) consistent with CO2 vibrational relaxation:
  • High-frequency regime (f > ~240 Hz): LIBS-based measurements yielded sound speeds of 246–257 m s−1 during daytime, with maximum values near 11:00–14:00 LTST and minima around 18:00. Intra-session dispersion decreased from ~1.5% at noon to ~0.5% by 18:00, consistent with waning turbulence.
  • Low-frequency regime (f < ~240 Hz): From Ingenuity flight 4 with emitted frequency 84.43 Hz, a fitted Doppler solution gave c = 237.7 ± 3 m s−1; correcting for a ~2.5 m s−1 wind toward the helicopter gives ~240 m s−1, matching temperature-derived predictions (T ≈ 232–240 K at 1.45 m, γ = 9/7) of 238.8–242.9 m s−1.
  • Net dispersion in the audible range is ~10 m s−1 between the low- and high-frequency regimes at the surface.
• Acoustic attenuation with distance above ~2 kHz was retrieved in situ and shows a strong contribution from CO2 vibrational relaxation in the audible range; attenuation is strong at Martian pressures and temperatures.
• Temperature-derived sound speeds from MEDA and Mars Climate Database agree well with the acoustic measurements, with the effective path-integrated sound speeds corresponding to heights near or above MEDA’s 0.85 m sensor, consistent with vertical thermal structure.
• The measurements validate continuum acoustic theory at small Knudsen numbers on Mars while demonstrating the need to include molecular-scale energy exchange (vibrational relaxation) to model propagation parameters.

The study addresses longstanding uncertainties in Martian atmospheric acoustics by directly measuring high-frequency pressure fluctuations, sound speed, and attenuation under true surface conditions. The observed two-regime sound speed confirms that CO2 vibrational relaxation at low pressure causes measurable dispersion in the audible band, providing crucial ground truth for propagation models previously constrained only by laboratory data and theory. The PSD analyses reveal a clear transition to the dissipative regime above a few Hz, offering new constraints on boundary layer turbulence at scales and frequencies previously inaccessible. These findings improve confidence in modeling of sound propagation in CO2-rich, low-pressure atmospheres (Mars, Venus) and inform planetary boundary layer parameterizations (e.g., in LES), particularly regarding unresolved dissipation at small scales. The agreement between acoustic and temperature-derived sound speeds supports the use of acoustic measurements as probes of near-surface thermodynamics and wind. In addition, controlled sources (LIBS, Ingenuity) enable precise, repeatable diagnostics of propagation and environmental variability.

This work presents the first in situ soundscape of Mars, establishing that: (1) atmospheric pressure fluctuations extend into a high-frequency dissipative regime; (2) sound speed is frequency dependent with a ~10 m s−1 dispersion across the ~240 Hz relaxation crossover; and (3) high-frequency attenuation is strong due to CO2 vibrational relaxation. These results provide essential validation for acoustic propagation models in CO2-dominated, low-pressure atmospheres and open a new window into microscale atmospheric dynamics on Mars. Future work should include systematic sound speed and attenuation measurements across local times, seasons, and sites; detailed characterization of the shear-to-dissipation transition; derivation of dissipation rates relevant to heat diffusion; and expanded use of acoustic diagnostics for both environmental studies and rover health monitoring.

• Frequency coverage: The measurements do not extend to sufficiently low frequencies (<1 kHz for some analyses) to fully resolve discrepancies among existing propagation models in that band.
• Source and environment constraints: Wind flow around the rover/mast can contribute to flow-induced noise; disentangling ambient turbulence from rover-induced effects remains nontrivial.
• Spatial representativeness: Measurements sample a single location and limited heights; derived sound speeds are path integrated and may be biased toward near-surface thermal gradients.
• Temporal sampling: Data span limited durations per sol; turbulence statistics and attenuation may vary with diurnal and seasonal cycles not fully captured here.
• Affiliation mapping and some ancillary instrument details for the highest superscript indices were not available in the excerpt, potentially limiting cross-references.