Diamictites are unsorted sedimentary rocks that typically include a spectrum of grain sizes from clay particles to boulders (Moncrieff, 1989) through deposits that bear blocks and megablocks (Terry and Goff, 2014). Some of the diamictites in the Cryogenian are argued to have been deposited in an exclusively non-glacial context as olistostromes or slope deposits (Nascimento et al., 2016, Damara Orogen, Namibia), whereas many are interpreted to contain a mixed slope-glacial signature (Macdonald et al., 2013; Le Heron et al., 2017). In the Cryogenian glacial record of Death Valley, California (United States), Busfield and Le Heron (2016) pointed to the repeated occurrence of lonestones associated with impact structures in shales, which they interpreted as ice-rafted dropstones. A similar pattern was observed in the central Kingston Range (Le Heron et al., 2014) and in the southern Kingston Range (California, United States) where dropstone-bearing horizons are intimately associated with both stratified and massive diamictites. It is this context that allowed them to emphasize a strong glacial influence on sedimentation, although further studies in the southern part of that range by Le Heron et al. (2018) demonstrated pervasive evidence for a glacial influence on sedimentation. Latterly, Kennedy and Eyles (2021) took a much more skeptical view, arguing that “dropstones” across a number of outcrops in the eastern Death Valley region could be ascribed to deflated debris, left behind while the main flow moves downslope. Although this might be regarded as a somewhat controversial view, the interpretations of Kennedy and Eyles (2021) challenge glacial geologists to consider the very foundations of evidence for a glaciogenic signature in Earth’s glacial record. Consequently, a careful outcrop description should ideally be supplemented with a quantifiable approach to determine the origin of diamictites as precisely as possible. In this paper, we demonstrate how diamictite texture can be objectively quantified based on clast orientations, at both macroscale and microscale (micromorphology), guiding interpretations. We carry out this using a hitherto undescribed section from the Valjean Hills area of Death Valley.

With few exceptions (Denis et al., 2010; Fleming et al., 2016; Busfield and Le Heron, 2018), there have been very few attempts to integrate micromorphology into the interpretation of Neoproterozoic diamictites or deep-time diamictites in general in spite of the technique being of proven value in the interpretation of Quaternary diamictons (Lachniet et al., 2001; van der Meer and Menzies, 2011). For sections that were long considered to be of glaciogenic origin, Delpomdor et al. (2017) showed that micromorphology allowed a mass flow interpretation to be supported on account of the relationship of ductile and brittle deformation, grain-to-grain relationships, and high strain rates. Similarly, for the Grand Conglomérat (within the Congo Basin), Kennedy et al. (2019) identified a dominant mass flow influence on sedimentation by analyzing an extensive core dataset. Consequently, stepping beyond the confines of mere outcrop description is essential for the rigorous interpretation of unsorted deposits in deep time. The problem is a vital one to solve, owing to the potential for erroneous interpretation of paleoclimate when the sedimentary textures are not properly understood. Here, we show how microfacies analyses can supplement a better understanding.

The Valjean Hills belong to an area of mountain ranges in the Death Valley region, which consist of tilted blocks associated with rifting of the south-western Laurentian margin during the Neoproterozoic. According to Calzia and Troxel (2002), the southern parts expose Late Proterozoic rocks from the Johnnie Formation overlain by Miocene carbonate breccias. The northeastern parts comprise the Noonday Dolomite and the upper member of the Neoproterozoic Kingston Peak Formation (KPF). Overall, the KPF, as described by Prave (1999), Macdonald et al. (2013), and Le Heron et al. (2014), is subdivided into four members (KP1–KP4). These consist of the basal KP1 pre-glacial succession, followed by Cryogenian aged rocks of the Sturtian glaciation (KP2 and KP3), and finish with glacial rocks of the Marinoan (KP4).

Deposition of the KPF in both the Sperry Wash region (Busfield and Le Heron, 2016) as well as in the central (Le Heron et al., 2014) and southern Kingston Range (Le Heron et al., 2018) illustrates a common SE-oriented paleoslope on the basis of sediment structures like current ripples, flute casts, grooves, and slump folding. This is also underpinned by regional thickness changes which increase progressively toward the southeast (Macdonald et al., 2013; Lechte et al., 2018). In contrast, the Silurian Hill section illustrates paleocurrent flow in an opposite orientation (toward the north), which in conjunction with compositional differences of clasts (the northern and southern facies of Troxel, 1966 and Troxel, 1982) was interpreted by Le Heron et al. (2017) to represent the convergence of ice masses flowing in opposite orientations. This interpretation developed the regional paleogeographic model of a Mojave upland to the south and a Nopah upland to the north by Wright et al. (1974) (Figure 13). The Valjean Hills comprise an outlier of the KPF that is overlain by breccia facies of the Noonday Dolomite (Calzia and Troxel, 2002) (Figure 1; see Figure 2A for further details). The overall character of the Valjean Hills section (red, iron rich, and dominated by turbidite–debrite intercalations) closely resembles that described in both the Sperry Wash section (Busfield and Le Heron, 2016; Kennedy and Eyles, 2021) and the sections of the Saddle Peak Hills (Macdonald et al., 2013; Le Heron et al., 2019; Le Heron and Vandyk, 2019). Interpretations of these sections range from glaciomarine to non-glacial debris flow deposits for Sperry Wash and (distal) glacial to non-glacial slope failure deposits at the Saddle Peak Hills. Thus, a key question is how the Valjean Hills section fits into this paleogeographic framework.

The technique of tracing the longest axes (or a-axes) of grains to interpret the spatial relationship of fabrics presented in this study uses a refined methodological approach originally proposed by Phillips et al. (2011) and further exemplified by Gehrmann et al. (2017) and Phillips et al. (2018) in Quaternary studies. We apply this method to three samples from Neoproterozoic diamictites from the Valjean Hills (35.662322° and −116.120398°), systematically analyzing the 2D fabric on three perpendicularly oriented thin sections. These are finally assembled into a 3D representation as a sample cube to reveal the true fabric in a graphical depiction and acquire fabric data that are plotted in a stereographic projection.

Samples have been collected from the base of three diamictite beds along the measured section (Figure 2). During sampling, areas of seemingly homogeneous diamictite were chosen, to minimize the potential effect on the microfabric caused by large boulders. Three oriented thin sections were cut from a single sample (in x-, y-, and z-direction). One thin section is oriented N–S, and the second is oriented perpendicular to that in an E–W orientation. Both with the bottom and top edges oriented horizontally. Perpendicular to these, a top view thin section was cut horizontally with edges exactly oriented north to south and east to west. In the following discussion, these three views are referred to as NS, EW, and TOP. When reassembled, the three 2D planes act as the basis for the fabric reconstruction, being a representation of the sample cube.

Each thin section was scanned with 2400 dpi via a 9000F Mark II CanoScan flatbed scanner under transmitted light. The color-corrected images were imported to a vector graphics software (CorelDRAW) to trace the longest axis of each individual skeleton grain (>0.05 mm, depending on scan resolution) with a vector line by hand. This enabled the systematic surveying of grain-fabric associations within the diamictite. It is noteworthy that all such axes of skeleton grains on a thin section only represent the apparent a-axes and not true in situ a-axes, thus constituting an artifact of the cutting orientation of the thin section. The basic workflow is subdivided into seven steps (Figure 3):

(1) Tracing of the apparent a-axes of skeleton grains on a high-resolution scan of an oriented thin section (Figure 3, step 1): Skeletal grains comprise all components larger than 35 µm that are clearly distinguishable from the plasma. Together, all grains represent the S-matrix. Particles smaller than 35 µm form the plasma or plasmic microfabric (van der Meer and Menzies, 2011). Only well-visible clasts with a length/width ratio >1.1 have been traced. The next step involves systematic surveying of clast orientations to map possible microstructures. If observable, structural features like rotational structures, folds, faults, and textural characteristics like layering, clast shape, and clast distribution are used to define a potential depositional or deformation history, and the nomenclature of Menzies and Meer (2018) is followed.

After completing every step, a fabric interpretation and derivation of possible depositional or deformation directions is possible.

To maximize fabric data, we deploy the use of outcrop-scale fabric data from diamictites, obtained through terrestrial structure from motion (TSfM) photogrammetry models, which complement and develop the micromorphological approach. Acquiring measurements from well-lithified rock outcrops is a laborious and time-consuming part, yet the diamictites in the study area have an advantage as they exhibit well-weathered-out clasts and a general appearance that enable virtual measurements. Therefore, meter-scale ground-based photogrammetry models of three outcrops corresponding to the exact location of the sampled diamictites have been computed. All of these models serve the purpose of extracting macroscopic clast orientations for comparison with the microfabric mapping. Several precautions need to be taken into consideration, since the correct dimensions and real-world orientations are crucial for obtaining solid data from virtual outcrop models. For that purpose, a 40 cm-long, horizontally laid out level, oriented exactly north, was placed on the outcrop, accompanied by labels that are marked with the dip and dip direction directly drawn onto the rock. These arrangements served as markers to orient and scale the models according to the real-world properties. Several hundred photos were taken with a 16MP Fujifilm S1 digital camera from a variety of viewing angles to recreate a representation of the actual outcrop topology. For the generation of the 3D outcrop models, the software Agisoft Metashape version 1.7.4 was used. The general workflow for TSfM photogrammetry (and its advantages over traditional compass-clino measurements) was well discussed by Cawood et al. (2017) and can be subdivided into an initial alignment of all photos and a computation of a tie-point cloud, followed by the computation of a dense cloud. This dense cloud acts as the basis for the following computation of a textured polygon mesh that was used to analyze the macrofabric. Measurement of clast orientation was performed in CloudCompare (2022) by importing the mesh and using the compass plugin (Thiele et al., 2017). The build-in lineation tool allows measuring the longest axis of each representative clast and provides accurate readings of dip and dip direction.

Unlike the well-exposed and coherent sedimentary successions that crop out at Sperry Wash or in the type section of the KPF (Macdonald et al., 2013; Busfield and Le Heron, 2016; Kennedy and Eyles, 2021), exposed rocks in the Kingston Range in the northern Valjean Hills have a more fragmentary record. As the Valjean Hills section is dissected by numerous medium- to small-scale faults (offset ranging from <dm to >m), compiling a composite section of accurate thickness is challenging. However, in spite of the evidence for faulting and some structural complexity, many beds like those in Sperry Wash have a pristine character making them excellent candidates for thin-section-based fabric analysis. This is in stark contrast to the greenschist to amphibolite grade metamorphism and highly strained characteristic of the KPF in the Panamint Range (Hodges et al., 1990; Petterson et al., 2011).

Underlying carbonate platform deposits of the Beck Spring Dolomite do not crop out as in other Death Valley outcrop areas (Le Heron and Vandyk, 2019). The upper erosional contact with the overlying Noonday Dolomite is, however, present and was mapped by Calzia and Troxel (2002). The surveyed rocks on the southerly exposed valley flank (see Figure 1 for section locality; Figure 4 for a schematic overview) dip within the outcrop toward the south or a southeasterly direction with angles between 30° and 70°. Although fault surfaces themselves are rare, their presence is suggested by highly variable dips across the outcrop belt. Incisions in between valley flanks mark hidden faults, with horizontal offsets exceeding tens of meters and varying inclinations and tilt of beds. Therefore, bedwise recording of the orientation is especially important for the reconstruction of the true thickness and possible depositional directions. Generally, three main types of lithofacies associations are distinguished across the outcrop:

(1) Breccia of light ochre angular to very angular dolostones with a mean grain size of very coarse pebbles and dark gray matrix cement (Figure 2A): Dolostones are finely crystalline and featureless under freshly hammered conditions; otherwise, they exhibit microkarst weathering in exposed areas. These correspond to the Noonday Dolomite.

Diamictite LD (Figures 2D, 5F) shows approximately 86% red matrix supporting angular to subangular clasts of varying lithologies. Dominant clasts are siltstone, sandstone, dolostone, and quartzite. Additionally, a variety of igneous rocks are recorded: basalt, diabase, and granitoid. Separated by ca. 18 m of stratigraphy, the overlying MD (Figures 2C, Figures 5C, D) shows a similar composition regarding the aforementioned lithologies. Clast shapes match the lower diamictite, although grain sizes are generally smaller, on average ranging from medium- to coarse-grained gravel, with one prominent dolostone outlier (70 cm). The red iron-rich matrix (93%) is massive and lacks internal structures. Solid information on spatial extent is vague due to limited exposure of this unit.

Diamictite UD is massive at the top, red to dark gray, and variably matrix- to clast-supported, making it hard to estimate the matrix/clast ratio (Figures 2B, Figures 5A, B). Clasts range in size between very coarse gravel and cobbles. Contact to the directly underlying sandstone is undulating and gradational in a 10 cm-thick band (see Figure 5A). Clasts are generally better rounded and show distinct facetting and striations. The majority of lithologies are light yellow dolostones, silt and sandstones, quartzites, and accessory components in the form of igneous rocks like diabase, basalt, and granites. As a whole, from a purely lithological standpoint, the three present diamictites are very similar.

Reassembling the thin section into a “sample cube” allows the spatial relationship of each fabric component to be revealed (see Figure 10). To acquire a mean orientation for each fabric component, all domain polygons are represented by a line matching the longest axis of a domain polygon. By plotting the average orientation of these lines of every fabric component (S1 and S2) as great circles and linears into stereographic projection, all three sample blocks show that the S1 foliation forms a planar fabric (all linears lie on a great circle). However, the orientation of the S2 foliation always plots in one quadrant and, thus, forms a linear or cluster fabric, indicated by the area where the great circles intersect. The dip and dip direction of both fabric components show a similar orientation for the in situ position as well as after bedding correction for each diamictite. The linear component is oriented toward the SE, whereas the planar component is oriented toward NW to W.

Since the microfabric analyses of thin sections only provide a small glimpse of the overall outcrop situation, macroscopic measurements of clast orientations were carried out for all three diamictites in the section that were also sampled (Figure 11). The stereographic plots show an overall similar result compared to the microfabric mapping. All three diamictites exhibit a dominant linear clast orientation dipping toward the SE. Compared with the bedding of each diamictite unit, the linear component lies almost in the same level as the bedding or few degrees below the bedding plane (see stereoplots of Figure 10). In contrast to the two observable fabric components of the thin sections samples, only a clustering is notable.

For the Silurian Hills region, Le Heron et al. (2017) proposed that diamictites ultimately derived from a glacial source could be differentiated from those of non-glacial (gravitational collapse) origin on the basis of composition. In that region, dropstones were shown to be of basement crystalline composition (schists, granites, and gneisses). On account of them being sandwiched between diamictites with compositionally identical clasts, a glaciogenic origin was proposed, and they were interpreted as glaciogenic debris flow (GDF) deposits. Mantling the basement, and directly below the KPF in that range, a dolostone-dominated carbonate platform crops out. Thus, large (up to 50 m long) blocks of carbonate were interpreted as olistoliths, in a similar way to the much larger (up to 1 km long: Macdonald et al., 2013) examples in the central Kingston Range. It should be noted that the logic of these interpretations cannot be applied to the Valjean Hills owing to the compositional similarity between the three principal diamictite levels noted previously.

More recent work (Kennedy and Eyles, 2021) has sharply questioned a glaciogenic influence on diamictite deposition, most particularly at Sperry Wash sections. A key question remains whether glaciogenic versus non-glaciogenic diamictites can be differentiated from one another on textural (specifically micromorphological) grounds. The famous catalog of microstructures developed by van der Meer (1993) and Menzies et al. (2006) and later iterations (Menzies, 2012; Menzies and Meer, 2018) show that key structures can be found in both diamictites regardless of their mode of emplacement and, thus, likely only play a supporting role in the development of a final interpretation which must fully consider the geological context. Indeed, investigations of mass flow diamictons in glacial environments show that they may be impossible to differentiate from their glaciogenic counterparts (Phillips et al., 2022). This is also true for moraines that have been remobilized by mass flows (Phillips and Kearsey, 2020). Thus, the study of these “tectonomicts” (Menzies et al., 2006; Menzies, 2012) may play a vital role in understanding the deformation process but may be less useful in resolving the deep-time controversy about glaciogenic influence in deep time, unless the glacial context is already firmly established (Busfield and Le Heron, 2018). There have been few attempts to highlight the formation of cross-cutting fabrics in laboratory experiments (Iverson et al., 2008) to quantify strain rates and clast orientations; the latter is very similar to the findings described within this study. Thus, based on the current understanding of the fabric evolution of modern and Quaternary “tectonomicts,” there is little to support a direct glaciogenic origin for the diamictites in the Valjean Hills.

The real power of fabric analysis of diamictites is their role in piecing together complex paleogeographic puzzles. Our results illustrate that both microfabric and macrofabric analyses of three diamictite intervals in the Valjean Hills reveal a strongly developed SE-oriented linear clast orientation. In neighboring outcrop belts, such as the Sperry Wash sections (Busfield and Le Heron, 2016) and in the central and southern Kingston Range (Le Heron and Busfield, 2016; Le Heron et al., 2018), a strongly preferred SE paleoflow orientation is recognized. In these previous studies, paleoflow was determined specifically from the crest orientation of current ripples or the 3D expression of cross laminae in sandstone beds. In those locations, whose stratigraphy was interpreted as a combination of debrites, hyperconcentrated flow deposits, and turbidites, evidence of emplacement orientation of apparently massive diamictites was lacking. Our new fabric data demonstrate that the linear clast orientation in diamictites—expressed equally well at the micro- and macroscale—can be used to proxy paleoslope, or at least paleoflow. This approach provides a vital data point in the schematic paleogeographic reconstruction (Figure 13). Moreover, we propose that this method—teasing out a well-expressed fabric orientation from deposits that superficially appear structureless—may have a significant role to play in the interpretation of other diamictite-rich successions. This approach may be adopted to solve a number of paleogeographic problems throughout a wide spectrum of the deep-time record. For example, in the Late Paleozoic record of South Africa, Dietrich and Hoffmann (2019) proposed a new model of grounding zone wedge deposition for stacked diamictites that show some similarity to the Valjean Hills succession described herein. Nevertheless, information on ice flow direction and paleoslope orientation is largely missing or anecdotal, yet could potentially be resolved with quantitative fabric analysis. A similar problem concerns basin configuration during deposition of the Grand Conglomérat (Cryogenian) of the Congo Basin, a succession dominated by mass flow diamictites (Kennedy et al., 2019).