Atmospheric Dust Causes Darkness to Fall Rapidly on Mars

On mission sol 783 (at Martian solar longitude (L_S ) of 217°), Mars Science Laboratory's (MSL's) MastCam instrument imaged a faint glow, extending upwards from the horizon (Figure 1(a)). The images were acquired from 18:58 to 19:01 LTST, nearly an hour after sunset occurred at 18:05 LTST, providing an opportunity to examine illumination in the Martian sky around sunset. Figure 1(a) displays a mosaic of this image sequence (5 images in total), where the data are separated between the red, green and blue filters. Images were captured using the MastCam L0 filter (Bell et al. 2017) with a 5 second exposure at an azimuth of −55°. The images were taken in quick succession with increasing elevations: 59, 189, 319, 449, and 579; this sequence was executed with the intention of positioning the camera to observe C/2013 A1 Siding Spring. The images were processed using the method detailed in Smith et al. (2019), such that the color axis in Figure 1(a) corresponds to pixel radiance.

Zoom In Zoom Out Reset image size

The radiance of the observed glow is small when compared to daytime sky brightness values, with an average value between 20 and 35 μW m⁻² Sr⁻¹, and appears brightest in the blue filter (the average B/R ratio is 1.25). The glow is most concentrated at approximately 15° of elevation, but extends from 5° to 40°. The light spans the azimuth range of the images, extending from −625 to −475, northwest of the rover (0° is north) and increases in intensity toward the west.

There are several possible sources for this illumination. These include: atmospheric scattering from light reflected by Phobos (Phobos-shine), light scattered by dust particles in the plane of the solar system (zodiacal light), or light scattered by dust in the Martian atmosphere (twilight).

Two of these phenomena can be immediately eliminated. Phobos is below the horizon in these images and therefore cannot explain the glow. The zodiacal light is both too dim and has the wrong mixture of wavelengths. Examining observations of zodiacal light seen at the Earth, this source would have a mean radiance of ∼0.0054 μW m⁻² Sr⁻¹ if observed by the MastCam L0 filter's red bayer filter, ∼0.0041 μW m⁻² Sr⁻¹ in the green and ∼0.0053 μW m⁻² Sr⁻¹ in the blue (Bernstein et al. 2002) giving a B/R ratio of close to unity. This source is therefore too poor in blue light to explain the observations. Furthermore, there is no reason to expect the zodiacal light to appear brighter at Mars than it does at the Earth, even without considering the opacity of the Martian atmosphere at visible wavelengths. Therefore, this source is four orders of magnitude smaller than what would be required to explain the observations.

The light source is most likely Martian twilight, indeed, Martian sunset illumination (Figure 1(c)) is known to be especially rich in blue light compared to the conditions at mid-day due to the long scattering path which preferentially forward-scatters blue light as compared to red, which is scattered more broadly and dominates the illumination of the sky away from the Sun. We can examine this hypothesis by comparing the magnitude of twilight illumination on Earth for an equivalent solar zenith angle and adjusting the parameters for Mars. The solar elevation angle of the MastCam images is −14° below the horizon, a solar zenith angle (SZA) of 104°. For an equivalent SZA on the Earth, the upwelling reflected twilight flux has been measured by the CERES instrument to be 0.09 ± 0.05 W m⁻² (Kato & Loeb 2003), averaged over all terrain types. This value may be divided by the albedo of the Earth's surface (∼0.4) to derive the downward propagating irradiance, 0.23 W m⁻². When adjusted to the Mars-Sun distance at L_S  = 214° of 1.40 au, the expected twilight flux integrated over all solid angles for an earth-like atmosphere at the distance of Mars would be close to 0.12 W m⁻².

How does this compare with the observation of Figure 1? As the solar azimuth at the time of the observation is −107°, this implies that the MastCam images are capturing the rightmost extent of the twilight radiation, and the center of this radiation is ∼44° west from the region of the sky imaged by the MastCam, directly above the Sun. As is shown in Figure 1(b), the glow appears to increase steadily in brightness moving toward the west (left), with a ∼13 μW m⁻² Sr⁻¹ linear increase over a 15° azimuthal interval. The center of the source is therefore 44 μW m⁻² Sr⁻¹ brighter than the left end of the image, which ends at 35 μW m⁻² Sr⁻¹. These images were used to create a model of the sky twilight radiance (Figure 1(d)). The center of the source has a radiance estimated at ∼73 μW m⁻² Sr⁻¹. To calculate the total radiance in the sky, we can use this linear rate of decrease from the source and integrate with respect to azimuth and elevation. Integrating all the flux from the sky in our model yields a value of 4.5 × 10⁻⁵ W m⁻² on the surface from this illumination. However, the spectral range of the MastCam L0 blue filter only captures 10.72% of the solar spectrum (Bell et al. 2017) which suggests that the total flux at the time of the measurement was closer to 4.2 × 10⁻⁴ W m⁻², nearly three orders of magnitude dimmer than what would be expected for an Earth-like Atmosphere.

The three order of magnitude difference in observed twilight brightness at the same SZA is likely the result of the large number of scattering dust particles in the atmosphere. On the sol this data was acquired, the vertical optical depth of dust was 1.14 (Vasavada et al. 2017), more than five times the typical optical depth under clear sky conditions on Earth. During the day, the forward scattering nature of the dust means that a large fraction of the radiance scattered from the direct beam of the Sun goes into illuminating the rest of the sky and continues to propagate toward the ground.

However, after sunset, the optical path of sunlight through the atmosphere becomes exceptionally long, causing dust particles to obscure one another and producing a twilight radiance equivalent to the solar disk observed through 25 optical depths. This is equivalent to a path length through the atmosphere of approximately 200–300 km long under the typical range of Martian near-surface conditions, which is comparable to the ∼275 km path length calculated for the geometry of the situation, taking atmospheric refraction into account. The net result of this effect is that an observer on the surface of Mars would find twilight becoming deeper much more rapidly than on the Earth.

The authors would like to thank the MSL Science Team for the use of MSL data, and the Canadian Space Agency Participating Scientist Program for funding this research.