InSight Pressure Data Recalibration, and Its Application to the Study of Long-Term Pressure Changes on Mars

Abstract

Observations of the South Polar Residual Cap suggest a possible erosion of the cap, leading to an increase of the global mass of the atmosphere. We test this assumption by making the first comparison between Viking 1 and InSight surface pressure data, which were recorded 40 years apart. Such a comparison also allows us to determine changes in the dynamics of the seasonal ice caps between these two periods. To do so, we first had to recalibrate the InSight pressure data because of their unexpected sensitivity to the sensor temperature. Then, we had to design a procedure to compare distant pressure measurements. We propose two surface pressure interpolation methods at the local and global scale to do the comparison. The comparison of Viking and InSight seasonal surface pressure variations does not show changes larger than ±8 Pa in the CO₂ cycle. Such conclusions are supported by an analysis of Mars Science Laboratory (MSL) pressure data. Further comparisons with images of the south seasonal cap taken by the Viking 2 orbiter and MARCI camera do not display significant changes in the dynamics of this cap over a 40 year period. Only a possible larger extension of the North Cap after the global storm of MY 34 is observed, but the physical mechanisms behind this anomaly are not well determined. Finally, the first comparison of MSL and InSight pressure data suggests a pressure deficit at Gale crater during southern summer, possibly resulting from a large presence of dust suspended within the crater.

Keywords: CO2 ice; Mars; atmospheric mass; cap sublimation; pressure.

Figure 1

(a) Schematics of the problem of interpolation with slope winds between the bottom of Gale Crater (point B) and the rim of the crater (point A). Colored dots illustrate the different paths that can be taken to integrate the hydrostatic equation. (b) Relative error of the interpolated pressure from point B to point A and the exact pressure at (a) The black curve is the relative error when point B is not interpolated to point A, while colored curves are for the relative errors when using different altitudes for the temperature. The air temperatures T used in the interpolations are computed at an altitude z above point (b) (c) Relative error on the local interpolation when using the temperature at 1 km above the surface at point A when considering several kinds of weather scenarios. Results are given for the mesoscale simulation that ran for L _s  = 180°, but similar results (i.e., same magnitudes) are obtained at other L _s . (d) Weather‐induced uncertainty of the Viking surface pressure interpolated to InSight landing site computed with extreme dust scenarios when compared to clim dust scenario (red and blue curves).

Figure 2

(a) Diurnal averaged surface pressure computed from the 20 Hz data acquired during the two years of the mission (red and blue), with pre‐landing surface pressure predictions (black curve) and baroclinic waves amplitudes (gray filled area) from Spiga et al. (2018). (b) Diurnal averaged temperature of the pressure sensor. Red dots are for the first year of the mission, while blue dots represent the measurements taken during the second year of the mission. Dashed black lines highlight the significant correlations between the pressure sensor temperature and the pressure measurements.

Figure 3

Correction of the thermal sensitivity E applied on the pressure measurements versus sensor temperature T.

Figure 4

(a) Diurnal averaged surface pressure computed from the raw pressure data. (b) Diurnal averaged surface pressure after applying the thermal correction. Error bars represent the uncertainty on the measurements after the correction at 3 − σ. The details of the uncertainty computations are described in the text. Red dots are for the first year or the mission, while blue dots are for the second year.

Figure 5

(a) Monte Carlo analysis to retrieve σ _{E(T = 260K)}. The histogram of the samples is presented in gray and is normalized to obtain a probability density function. The fitted normal law is illustrated in red and has as parameters the mean μ and the standard deviation of the distribution σ. (b) Empirical law for σ _E(T) (T) obtained from Monte Carlo analysis (black cross) and third order polynomial fit of this law (red line).

Figure 6

(a) Comparison between the surface pressure measured by InSight and that measured by MSL but interpolated to the InSight landing site, for MY 34, 35, and 36. The filled box around the plain line depicts the 3 − σ uncertainty of the interpolation due to weather‐induced uncertainty and MSL absolute errors, following the methodology presented in Section 2.2. Pressure interpolated is averaged over a period of 15 sols to remove atmospheric tides and baroclinic activity. InSight measurements are diurnally averaged thus still indicate baroclinic activity with periods of several sols. The error bars correspond to the 3 − σ on InSight corrected pressure measurements as described in Section 3.4. (b) Evolution of the ratio of MSL REMS pressure measurements interpolated to the InSight landing site, and InSight pressure measurements. Dots correspond to the ratio using the interpolation method described in Section 2.2.

Figure 7

(a) Evolution of the ratio of MSL REMS pressure measurements interpolated to the InSight landing site, and InSight pressure measurements. Dots correspond to the ratio using the interpolation method described in Section 2.1, that is, neglecting atmospheric dynamic effects, during MY 34 (green), MY 35 (gray), and MY 36 (blue). (b) Anomaly between the temperature of the GCM at an altitude of 1 km above the surface, and the temperature T _* that gives a ratio of 1, as a function of L _s (colored curve) for MY 35. (c) Comparison between InSight surface pressure over a complete sol and MSL pressure interpolated at InSight landing site between L _s  = 275° and 280°, during MY 35. (d) Extract of THEMIS image V63417011 of Gale Crater (center of the original image: 4.9°S; 137.0°E) taken at L _s  = 130°, LTST = 7.2 hr, with a solar incident angle of 74.5°. (e) Extract of THEMIS image V65575024 at the same location, taken at L _s  = 229°, LTST = 7.2 hr, with a solar incident angle of 71.3°. The black arrow on (e) points to the suspected aerosols, whereas the red arrows on (d) and (e) point to the same crater for a comparison of the perceptibility of the ground. White arrows point to the position of the Sun in the sky.

Figure 8

Comparison between the surface pressure by Viking 1 interpolated at the InSight landing site for MY 34, 35, and 36. The filled box around the plain line depicts the 3 − σ uncertainty of the interpolation detailed in Section 2.2. Pressure interpolated is averaged on a period of 15°to remove atmospheric tides and baroclinic activity. InSight measurements are diurnal averaged and still keep baroclinic activity. The error bars correspond to the 3 − σ on InSight corrected pressure measurements as described in Section 3.4.

Figure 9

Comparison of the SSPC images taken by (a) Viking orbiter during MY 12, L _s  = 192.6° (extracted from James et al. (1979)); (b) MARCI during MY 33, L _s  = 192.3°; and (c) MARCI during MY 35, L _s  = 192.9°. Blue arrows flag characteristic surface features for the comparison like craters. Orange arrows indicate a possible difference between the Viking images and MARCI images while green arrows indicate a good match between the images. (d) to (f) are zoom on the lowest flagged craters of (a), (b), (c). The 60°S circle of latitude on image (d) extracted from James et al. (1979) is misplaced, but arrows point to the same elements.

Figure 10

(a) Comparison of the surface pressure measured by Phoenix (green dots), MSL (blue dots), and InSight (red dots), to Viking 1 measurements (interpolated at each landing sites), from MY 29 to MY 36. Yellow boxes correspond to periods of local dust storms at landing sites (Holstein‐Rathlou et al., ; Ordonez‐Etxeberria et al., 2019), while the orange box corresponds to the period of MY 34 global dust storm (Lemmon et al., ; Viúdez‐Moreiras et al., 2019). (b) to (d) Evolution of the relative difference between Viking 1 interpolated and MSL (blue) and InSight (red), as a function of martian year at L _s  = 20°(b), L _s  = 170°(c), L _s  = 310°(d). The error bars indicate the sensitivity of the comparison with regards to the interpolation uncertainty at 3 − σ, as described in Section 2.2.