A fluvio-lacustrine environment preserved in the Jezero crater inlet channel, Neretva Vallis (Mars)

Introduction:

Fluvial valley networks on Mars provide key sites for investigating the role of surface liquid water in Mars’ ancient past. Between June and August 2024, the Perseverance rover spent 85 sols exploring a series of outcrops in Neretva Vallis, one of two inlet channels which provided water to the Jezero crater paleolake. These predominantly fine-grained sedimentary rocks comprise the “Bright Angel” formation [1], and the collection of an astrobiologically compelling sample from these rocks [2], motivates the development of a robust stratigraphic model to understand their emplacement. Here, we document the sedimentology and stratigraphy of the Bright Angel formation using Mastcam-Z landscape mosaics and derived 3D Digital-Outcrop-Models (DOMs) [3, 4], to constrain its depositional environment and sedimentary evolution.

Sedimentology and stratigraphy:

The Bright Angel formation comprises seven members (Figure 1). The base of the sequence comprises 1–10-cm-thick, planar-stratified rocks of the Tuff Cliff member, which exhibits at least one example of sub-rounded, pebble-sized clasts up to 2.5-cm-diameter (Figure 1B). These dark-toned rocks fine upwards into an ~8–10-m-thick succession of laminated mudstone (Figure 1C-D) variably cut by Ca-sulfate veins and nodules, with rarer dm-thick pebbly interbeds in its lower ~2m, termed the Walhalla Glades member. ~500 m west, at the northern contact of the Bright Angel formation with the Margin unit, a ~0.5-m-thick lens of monomict, matrix-supported olivine granule conglomerate is observed, named the Fern Glen Rapids member. This is overlain by a ~0.5 m thick mudstone with rare, coarser-grained, cm-thick olivine-rich horizons (Figure 1E), termed the Cheyava Falls member, which is overlain by the Walhalla Glades member. The Tuff Cliff member was not observed here. This succession is overlain and incised by an unstratified, unsorted, matrix-supported conglomerate exposed in the south of the channel (Figure 1F), termed the Mount Spoonhead member. The Serpentine Rapids (SR) member comprises a cross-stratified pebbly conglomerate (Figure 1G), lining the southern channel margin and appears to cap the sequence.

Figure 1: Geologic map and outcrop images of the members of the Bright Angel formation. (A) Structural-geologic map, (B) Contact of the Tuff Cliff and Walhalla Glades members, (C) Pebble-sized clasts in the Tuff Cliff member, (D) Contact of the Cheyava Falls and Fern Glen Rapids members, (E) plane-parallel lamination in the Walhalla Glades member, (F) Matrix-supported, conglomeratic texture of the Mount Spoonhead member, (G) Cross-stratified pebbly conglomerate of the Serpentine Rapids member. Image credits: NASA/JPL-Caltech/ASU/MSSS

Structural analysis and sedimentary successions:

Cross-sections were constructed using structural mapping from Mastcam-Z DOMs (Figure 2). Long-channel profile (A-A’) shows the Tuff Cliff member exposed in the east is likely buried beneath other members in the western Bright Angel outcrop. Cross-channel profile (B-B’) shows a fining-up sequence (from the Tuff Cliff to Walhalla Glades member), truncated by the Mount Spoonhead member conglomerates and Serpentine Rapids member cross-stratified conglomerates. A close-up of the northern exposure reveals an open, synform structure, with beds dipping into the channel. This apparent channel-infilling structure of the unit is supported by ground-penetrating radar acquired by RIMFAX [5], and appears consistent with the Bright Angel formation being younger than the Margin unit (although the contact is very poorly exposed at the surface). Structural data were projected across this structure onto an inferred fold axis, and assuming a constant bedding thickness, allowed the thickness of the mudstone succession to be estimated to at least ~10 m. Profile C-C’ shows that within ~2 m of the northern contact, bedding dips increase from 30˚ to >50˚ into the channel, exceeding the expected angle of repose and implying localised deformation [6].

Figure 2: Structural analyses of the Bright Angel formation, including long-channel profile A-A’, cross-channel profile B-B’ (with zoom of the northern section, showing a channel-filling bedding geometry), and zoomed profile C-C’ across the northern contact. Stereonet shows poles to bedding planes across the B-B’ section, and inferred “fold” axis.

Interpretation: The Tuff Cliff, Fern Glen Rapids, Cheyava Falls and Walhalla Glades succession is consistent with lacustrine deposition. The Tuff Cliff member conglomerates may reflect subaerial or proximal lacustrine deposition, and the fining-up into ~10 m of predominantly laminated mudstones (consistent with subaqueous suspension settling) may reflect a lake transgression. The matrix-support and compositional similarity of clasts in the Fern Glen Rapids and Cheyava Falls members to the Margin unit is consistent with local derivation from the valley walls, potentially as small debris flows. The very poor sorting and matrix-support of the Mount Spoonhead member is consistent with a mud-rich debris flow, with diverse clasts potentially sourced from the crater rim and beyond [7]. The Serpentine Rapids member cross-stratified conglomerates are consistent with unidirectional bedload transport in a fluvio-alluvial or fluvio-deltaic environment. Abundant Ca-sulfate veins in the lower stratigraphy supports burial depths sufficient for hydrofracturing.

This lacustrine environment occurs 10–50 m above the paleolake level implied by the Jezero western fan [8], suggesting the Bright Angel formation was deposited either (1) during a lake highstand, pre-dating breach of the eastern crater rim and western fan deposition [9], or (2) in a later-stage, valley-confined lake. The latter are common in terrestrial valley networks, where slope failure leads to channel blockage and formation of a lake upstream. These can produce successions similar to the Bright Angel formation [1,10], and may explain the absence of Bright Angel-like materials beyond Neretva Vallis. However, it remains unclear whether sufficiently thick deposits could accumulate to enable hydrofracturing of the lower members. Thicker deposits may accumulate in a more stable, pre-delta, lake highstand phase as proposed by [9]. In this model, the Bright Angel sediments may have been deposited up to 100 m below the highstand lake level [9].

References: [1] Jones et al., LPSC (2025), [2] Hurowitz et al., LPSC (2025), [3] Bell et al., Space Sci Reviews 217 (2021), [4] Paar et al., Earth and Space Science, 10 (2023), [5] Russell et al., AGU (2024). [6] Barnes et al., LPSC (2025), [7] Treiman et al. LPSC (2025), [8] Fassett and Head, Geophysical Research Letters, 32 (2005), [9] Salese et al., Astrobiology, 20 (2020), [10] Fort et al., Quaternary Research, 31 (1989).