The Mars 2020 Perseverance rover has explored the escarpment at the front of the western fan in Jezero crater, Mars, where it encountered a variety of rock units as in-place outcrops and as loose pieces of rock separated from outcrops, or “float” rocks. Comparing float rocks to in-place outcrops can provide key insights into the crater’s erosional history and the diversity of units in the Jezero watershed that the Perseverance rover cannot visit in-situ. Here, we used multispectral observations from Perseverance’s Mastcam-Z instrument to investigate the lithology and origin of float rocks found on the western Jezero fan front (sols 415-707). We identified four textural classes of float rocks (conglomerates, layered, massive, and light-toned) and investigated their physical characteristics, spectral properties, and distribution to interpret their source and constrain their mode of transport. We found that the conglomerate and layered float rocks are highly spectrally variable and altered with differing ferric and ferrous signatures, and they likely derived from local sedimentary outcrops in the western fan front. Massive float rocks are the least altered, exhibit ferrous signatures, and could have derived from local outcrop sources or more distal sources in the Jezero watershed. Massive float rocks separate into two subclasses: massive olivine and massive pyroxene, which likely derived from the regional olivine-carbonate-bearing watershed unit and the crustal Noachian basement unit respectively. The unique light-toned float rocks have variable hydration and low Fe-abundance, but there are no local outcrop equivalent of these rocks in the western Jezero fan or crater floor.

The first samples collected by the Perseverance rover on the Mars 2020 mission were from the Maaz formation, a lava plain that covers most of the floor of Jezero crater. Laboratory analysis of these samples back on Earth will provide important constraints on the petrologic history, aqueous processes, and timing of key events in Jezero. However, interpreting these samples will require a detailed understanding of the emplacement and modification history of the Maaz formation. Here we synthesize rover and orbital remote sensing data to link outcrop-scale interpretations to the broader history of the crater, including Mastcam-Z mosaics and multispectral images, SuperCam chemistry and reflectance point spectra, RIMFAX ground penetrating radar, and orbital hyperspectral reflectance and high-resolution images. We show that the Maaz formation is composed of a series of distinct members corresponding to basaltic to basaltic andesite lava flows. The members exhibit variable spectral signatures dominated by high-Ca pyroxene, Fe-bearing feldspar, and hematite, which can be tied directly to igneous grains and altered matrix in abrasion patches. Spectral variations correlate with morphological variations, from recessive layers that produce a regolith lag in lower Maaz, to weathered polygonally fractured paleosurfaces and crater-retaining massive blocky hummocks in upper Maaz. The Maaz members were likely separated by one or more extended periods of time, and were subjected to variable erosion, burial, exhumation, weathering, and tectonic modification. The two unique samples from the Maaz formation are representative of this diversity, and together will provide an important geochronological framework for the history of Jezero crater.

The Mastcam-Z radiometric calibration targets mounted on the NASA’s Perseverance rover proved to be effective in the calibration of Mastcam-Z images to reflectance (I/F) over the first 350 sols on Mars. Mastcam-Z imaged the calibration targets regularly to perform reflectance calibration on multispectral image sets of targets on the Martian surface. For each calibration target image, mean radiance values were extracted for 41 distinct regions of the targets, including patches of color and grayscale materials. Eight strong permanent magnets, placed under the primary target, attracted magnetic dust and repelled it from central surfaces, allowing the extraction of radiance values from eight regions relatively clean from dust. These radiances were combined with reflectances obtained from laboratory measurements, a one-term linear fit model was applied, and the slopes of the fits were retrieved as estimates of the solar irradiance and used to convert Mastcam-Z images from radiance to reflectance. Derived irradiance time series are smoothly varying in line with expectations based on the changing Mars-Sun distance, being only perturbed by a few significant dust events. The deposition of dust on the calibration targets was largely concentrated on the magnets, ensuring a minimal influence of dust on the calibration process. The fraction of sunlight directly hitting the calibration targets was negatively correlated with the atmospheric optical depth, as expected. Further investigation will aim at explaining the origin of a small offset observed in the fit model employed for calibration, and the causes of a yellowing effect affecting one of the calibration targets materials.