Stratigraphic change in flow transformation processes recorded in early post‐rift deep‐marine intraslope lobe complexes

The Early Jurassic Los Molles Formation in the Neuquén Basin of western Argentina is a rare example of well‐exposed syn‐rift to post‐rift stratigraphy. In the Chachil Graben, the onset of the early post‐rift stage is marked by drowning of a carbonate system and the development of two deep‐marine intraslope lobe complexes. This field‐based study in the Chachil Graben involved field mapping and correlating eleven stratigraphic logs, and petrographic analysis to document how grain size and texture within intraslope lobe sandstones change from the lobe centre to their frontal pinch‐out. Eight different bed‐scale facies are identified and inferred to be formed by turbulent (turbidites; Type A and B beds), transient turbulent–laminar (transitional flow deposits; Type C, D, E and F beds), laminar gravity flows (debrites; Type G) and post‐depositional clastic injections (injectites; Type H beds). Fifteen lobes form two stacked lobe complexes that show stratigraphic evolution from a lower argillaceous sandstone‐dominated lobe complex, built by transitional flow deposits, to an upper coarser‐grained, sandier lobe complex largely constructed by turbidites. Petrographic analysis quantified sandstone mineralogy, matrix content, grain size and sorting, revealing that both lobe complexes are volcanic arc‐sourced. This study proposes that the differences in the character of the two lobe complexes are due to maturation of sediment transport routes through progressive healing of the intraslope relief, with a concomitant decrease in substrate erosion and flow bulking. Also proposed here is a model for intraslope lobe complex development that accounts for the impact of flow‐confinement on flow behaviour and transformation induced by the inherited topography. Bed type distribution suggests that high‐density flows terminate more abruptly against confining slopes and produce greater depositional variability than lower‐density flows. This integrated petrographic, architectural and sedimentary process model provides new insights into how post‐rift intraslope lobe systems may act as hydrocarbon reservoirs, aquifers and carbon storage sites.

Sediment gravity flows interacting with topography can change both flow character (velocity, concentration) and direction (Kneller et al., 1991; Al-Ja' Aidi et al., 2004;Soutter et al., 2021); the resultant deposits differ from those in unconfined settings, which generally have more predictable facies variability (e.g. Pr elat et al., 2009;Spychala et al., 2017a,b). The influence of seafloor relief on bed types, facies transitions and the architecture of lobes bounded by steep slopes (>3°) has been well-documented in exhumed, highly confined settings (e.g. McCaffrey & Kneller, 2001;Sinclair & Tomasso, 2002;Patacci et al., 2014;Soutter et al., 2019). In contrast, there are fewer documented exhumed systems in which intraslope lobes developed above subtle slope changes (Smith, 2004;Burgreen & Graham, 2014;Spychala et al., 2017a,b;. This is partially attributed to more subtle bed type changes and pinch-out terminations, which might have been under-recognized in exhumed and subsurface analogues. Petrographic analysis and bed-scale characterization can help us to understand flow transformations recorded at the event bed scale, which represents a single depositional event, and the physiographic configurations of intraslope settings. Syn-rift and post-rift lobe systems are typically deeply buried, and outcrop analogues are rarely well-preserved due to later deformation during basin inversion. Therefore, our understanding of their bed types, facies transitions and depositional architecture remains limited. Early Jurassic strata in the Neuqu en Basin (Fig. 1A) provide a rare example of an exhumed, syn-rift to post-rift sedimentary succession that is well-preserved and well-exposed. In the Chachil Graben, located in the south-west of the Neuqu en Basin (Fig. 1B and C), the onset of the early-post rift stage is recorded by the drowning of a carbonate system and the subsequent development of an intraslope lobe system in the Los Molles Formation ( Fig. 2; . The objectives of this paper are to: (i) characterize the bed types of the early post-rift submarine lobe complexes and their variability towards pinch-outs; (ii) document the petrography and texture of the different bed types to better understand flow processes and environments of deposition; and (iii) discuss the stratigraphic evolution of deepmarine lobes deposited in a confined post-rift intraslope setting. A spatially well-calibrated sedimentary and architectural model will improve our understanding of early post-rift intraslope lobes and potentially be used as an analogue for lobes in similar settings elsewhere.

GEOLOGICAL SETTING
The Neuqu en Basin is located between 32°S and 40°S in central-western Argentina and centraleastern Chile, covering >160 000 km 2 . It was bounded to the north-east by the Sierra Pintada System, to the south by the North Patagonian Massif, and to the west by the Andean magmatic arc ( Fig. 1A; Legarreta & Uliana, 1996;Franzese & Spalletti, 2001). The stratigraphic record of the Neuqu en Basin is characterized by Mesozoic (starting in the Late Triassic) to early Cenozoic, continental and marine siliciclastic, carbonate, evaporitic and volcanic rocks (Legarreta & Uliana, 1996;Howell et al., 2005). Three tectonic phases are recognized in the development of the Neuqu en Basin (Vergani et al., 1995;Franzese & Spalletti, 2001: (i) Triassic to Early Jurassic rifting, characterized by the development of narrow and isolated depocentres; (ii) Early Jurassic (Toarcian) to Early Cretaceous post-rift thermal subsidence, associated with the development of the back-arc basin; and (iii) the Late Cretaceous to early Cenozoic formation of a foreland basin related to Andean compression. This study focuses on the earliest Los Molles Formation (Toarcian; Weaver, 1931), deposited during the early post-rift stage in the Chachil Graben.

STUDY AREA: CHACHIL GRABEN
The Chachil Graben is a 15 km long, 10 km wide, NNW-SSE trending normal fault-controlled depocentre. It is bounded to the north-west and south-west by the south-east and north-west dipping Chihuido Bayo Fault system (Fig. 1B). The initial syn-rift volcano-sedimentary deposits (Lapa Formation; Precuyano Cycle) thicken northward from <400 m at the basin margins to 2 km near its centre (Franzese et al., 2006). These thickness changes were controlled by fault systems striking parallel or oblique to the Chihuido Fault system along the horst forming the south-western border of the depocentre (Fig. 1B). Late syn-rift extension promoted the development of footwall highs (Paine Milla, El Luchador, Puesto Alfaro and Morro del Aguila) and lows inside the Chachil Graben (Mirador de Chachil and Pic un Leuf u;Figs 1C and 2;Franzese et al., 2006;. Regional subsidence associated with the late syn-rift phase led to the flooding of the rift depocentres,  . The onset of the early post-rift stage is recorded by the drowning and deepening of a carbonate system and the subsequent development of a deep-water lobe system. The palaeoflow direction is towards north-east (from left to right; see rose diagrams in Fig. 13). Note that the debrites pinch out in a more proximal position than the lobe complexes, which show a progressive basinward advancement of their pinch-outs from the base to the top (more detail in Fig. 13).
promoting the onset of carbonate sedimentation (Chachil Formation;Weaver, 1942;Leanza et al., 2013). Rapid basin deepening and drowning of the carbonate system led to the deposition of organic-rich, calcareous mudstones that draped inherited rift topography (basal Los Molles Formation; Fig. 2; . Healing of rift topography and the onset of extra-basinal siliciclastic supply promoted the development of early post-rift intraslope lobe complexes . Differential compaction of organic-rich mudstones over inherited rift topography promoted a long-lived seabed relief in the Chachil Graben ( Fig. 2; . Similar differential compaction has also been documented in the subsurface (Cristallini et al., 2009) and exhumed depocentres elsewhere in the Neuqu en Basin (Veiga et al., 2013). In the Chachil Graben, a south-dipping compaction hinge over the El Luchador fault block acted as an oblique counterslope to the north-east-directed sediment gravity flows, controlling the architecture and pinchout location of two submarine lobe complexes ( Fig. 2; dirty and cleaner lobe complexes, sensu . Both lobe complexes are 6 to 8 km long, 5 km wide and <50 m thick and are bound at their base and top by a >4 m thick mudstone interval or an erosionally-based debrite (Figs 1C and 2;. This study refers to them as the Lower and Upper Lobe Complex, respectively (Fig. 2). Their constituent bed types (Fig. 3) and stratigraphic architecture are described and compared to propose an overall depositional model.

DATA AND METHODS
The dataset comprises 11 sedimentological logs 1 : 25 scale (Fig. 1C), collected over a horizontal distance of 4 km. Logs were correlated, and the overall depositional architecture was constrained by walking out individual sandstone-rich packages and several stratigraphically distinct marker beds. Data collected included lithology, bed thickness and contacts, sedimentary structures, and palaeocurrent measurements from grooves, flutes, crossbedding, flame structures and ripple crosslamination. Uncrewed Aerial Vehicle (UAV) photogrammetry was used in conjunction with mapping and logging to capture the architecture of the sandstone and mudstone packages within the investigated stratigraphic interval.
Petrographic analysis was performed on 40 orientated and polished thin sections collected in 16 sandstone beds logged at a 1 : 2 scale (Figs 4,5,6,7,8 and 9). To understand the longitudinal character of flows, up to five samples were taken from the base to the top of the sandstone beds (see similar methods in Kane et al., 2017;Southern et al., 2017;Bell et al., 2018;Fildani et al., 2018). The long axes of 200 grains per thin section were measured (see Table S2), enabling quantification and analysis of sorting (Folk & Ward, 1957;F & W method). Detrital components were quantified using the Gazzi-Dickinson point counting method (Gazzi, 1966;Dickinson, 1970;Zuffa, 1985). Percentage abundance was quantified by counting 400 points per thin section using an automated and motorized stepping stage controlled by PETROG software (PETROG System, Conwy Valley Systems Limited, UK). The points were stochastically distributed across an aleatory grid, where the minimum interpoint distance was larger than the coarsest grain fraction to avoid repetition ( Van der Plas & Tobi, 1965). Point counting enabled the classifications of grains into 11 classes, matrix (particulate matter <0.062 mm) and cement. The results were plotted on a matrix versus mean size grain diagram (Fig. 10) and three ternary diagrams ( Fig. 11; Qm-F-Lt, Dickinson, 1983;Lm-Lv-Ls, Ingersoll & Suczek, 1979;Qp-Lv-Lsm diagram, Dickinson, 1985) because combinations of specific end-members discriminate between different grain properties (Hulka & Heubeck, 2010;Ciccioli et al., 2014).

BED TYPES
The sandstone beds were divided into eight bed types ( Fig. 3; Type A-H beds) based on their thickness, texture (for example, matrix content, grading, mudstone clast content), grading and sedimentary structures. Petrographic analysis was used to quantify the mineralogy, matrix content, grain-size range and sorting Figs 4,5,6,7,8 and 9) to support the outcropbased interpretation. Samples with >15% matrix content were considered matrix-rich, while those with <15% were considered matrix-poor, based on the threshold of Pettijohn et al. (1972), which is similar to the one proposed in Sylvester & Lowe (2004;ca 16%). The authors did not differentiate between detrital and authigenic matrix, but the abundance of carbonaceous matter, micas, fine quartz and feldspar grains suggest that the matrix is predominantly detrital (Lowe & Guy, 2000). The deformation of weak detrital grains such as sedimentary lithoclasts (Ls) and    devitrified volcanic clasts (Lv) contributed to the formation of the detrital matrix (pseudomatrix; Dickinson, 1970). Two additional bed types had no petrographic analysis because they were not genetically related to the flows responsible for the formation of intraslope lobes (Type G beds) or do not represent depositional event beds (Type H beds; Fig. 3).

Type A beds
Description: Thin-bedded (<0.2 m thick), matrixrich (>15%), planar-laminated and ripplelaminated with sharp, planar bed bases and tops. Matrix content increases upward within the bed ( Fig. 4; S13 sample bed: 17 to 33%). Beds are poorly to moderately well-sorted and normally  graded, comprising very fine to medium-grained sandstones. When analysed under the microscope, millimetre-scale mud chips are aligned along the laminae ( Fig. 4G; CH-26). Ripples suggest an overall NNE palaeocurrent. However, ripples show a broad divergence, even developing bidirectional ripple laminations. Near the pinch-outs, ripples are commonly replaced by convolute lamination. Interpretation: These bed types represent deposition from low-density turbidites (sensu Lowe, 1982). Grains are suspended due to flow turbulence and deposited under aggradation rates that are sufficiently slow to permit differential grain-size settling and tractional reworking (Best & Bridge, 1992;Hiscott, 1994a;Sumner et al., 2008). Bidirectional ripple laminations suggest flow deflection and reflection against seafloor relief (Kneller et al., 1991;Edwards et al., 1994). Convolute lamination suggests high sedimentation rates and has been attributed to the development of reflected bores and internal waves, which might be related to interaction with seafloor relief (Tinterri et al., 2016). The upward increase in matrix content reflects the differential grain-size settling from waning flows, which deposits coarser material near bed bases and finer material near bed tops (Mulder & Alexander, 2001).

Type B beds
Description: Medium to thick-bedded (0.3 to 1.2 m thick), matrix-poor (<15%) sandstones, with sharp planar tops and sharp erosive bed bases with abundant multidirectional grooves with a ENE-WSW oriented dominant trend. Type B beds are moderately sorted and weakly normally graded, passing upward from coarsegrained (locally very coarse and granular) to medium to fine-grained sandstone ( Interpretation: These sandstones were deposited from erosive high-concentration flows where particles are suspended by grain-to-grain interactions and fluid turbulence (Kuenen, 1951;Middleton & Hampton, 1973). Therefore, they are interpreted as high-density turbidites (sensu Lowe, 1982) deposited incrementally (Kneller & Branney, 1995) under aggradation rates high enough to suppress tractional reworking Fig. 10. Graph showing matrix content as a function of the mean grain size of each sample (n = 40) according to the bed type. The dashed line at the 15% matrix represents the threshold for matrixpoor and matrix-rich intervals (Pettijohn et al., 1972). Note that the high-density turbidites (Type B beds) are the coarsest and less argillaceous, in contrast to most transitional flow deposits (Type C, D, E and F beds) and low-density turbidites (Type A bed).  Dickinson, 1985) showing volcanic arc source. (Lowe, 1982;Sumner et al., 2008). The tractional structures and normal grading observed at bed tops indicate differential grain settling and effective shearing from waning gravity flows (Sumner et al., 2008;Talling et al., 2012;Terlaky & Arnott, 2014). High-angle cross-bedding developed over scours supports scour-fill rather than dune formation (see pseudo-dunes of Arnott & Al-Mufti, 2017, and references therein). Inverse grading in the lower division (for example, S4 sample bed) represents deposition under traction carpets (Lowe, 1982;Hiscott, 1994b;Sohn, 1997). The upward increase in matrix content reflects differential grain-size settling from waning flows, which deposit coarser material near the bed base and finer material near the bed top ( Fig. 5; S6 and S14 sample beds; Mulder & Alexander, 2001). The high sediment fallout rate can prevent the segregation of flocs from returning into suspension, resulting in enhanced trapping of clay flocs at the flow and bed base (e.g. Middleton & Neal, 1989). Sand-sized mudstone clasts in intermediate positions within the sandstone beds suggest that either they remained in suspension for longer than particles of a similar size and density (for example, quartz and feldspar grains; Mulder & Alexander, 2001) or that they were entrained and deposited later in the lifespan of the flow. In contrast, deposits with an upward decrease in matrix content are associated with developing a cohesive viscous near-bed sublayer (near-bed laminar plug; Lowe & Guy, 2000). Grain-size breaks ( Fig. 5; S4 sample bed) suggest that flow stratification developed between the cohesive viscous layer and the overlying turbulent flow division (Lowe & Guy, 2000). The grooves represent tool-like clasts that gouged the substrate  and indicate high-concentration flows. Groove development might be linked to the passage of a forerunning, fast-moving and bypassing debritic head, responsible for rafting mudstone clasts and transferring them as bedload, developing grooves (Baas et al., 2021). Disaggregation of mudstone clasts could trigger the development of turbulence-enhanced transitional flows or even lower transitional plug flows (LTPFs; Baas et al., 2009Baas et al., , 2011Baas et al., , 2016) over short distances.

Type C beds
Description: Medium-bedded (0.2 to 0.4 m), matrix-rich (>15%) sandstones, forming tripartite deposits with sharp boundaries between constituent divisions ( Fig. 6; S11 sample bed). Bed bases are sharp and conformable and contain abundant ENE-WSW-oriented discontinuous tool marks (skim marks) with high divergence. The lower division of Type C beds (5 to 20 cm thick) is structureless, weakly normally graded, moderately sorted and matrix-rich (27%). The middle division (10 to 15 cm thick) is ungraded and poorly sorted, with the highest matrix content of the three divisions (37%), and characterized by low-amplitude, undulating microbanding (<1 cm spacing; Stevenson et al., 2020), with alternating planar sub-parallel to undulatory, light (i.e. mud-poor) and dark (i.e. mud-rich) bands containing rare, millimetre-scale mud chips ( Fig. 6F; S11 sample bed). The dark argillaceous bands are responsible for the poor sorting within this division (Fig. 6). The upper division (<5 cm thick) is normally graded matrix-rich (30%) and moderately sorted. NNE-oriented ripple lamination is common, with millimetre-scale mud chips aligned within the ripple laminae (Fig. 6).
Interpretation: Due to its similar matrix content and lack of bounding fine-grained interval, the development of the three divisions is interpreted as being deposited from the same transitional flow. The normally graded lower division suggests deposition from a flow comprising a turbulent component to permit differential grain-size settling (Lowe, 1982). The lack of tractional structures and high matrix content suggest deposition from transitional flows under moderate to high aggradation rates. The banding in the intermediate division results from deposition and reworking from transitional flows, with a sustained traction period formed within the upper-stage plane bed flow regime (Baas et al., 2009(Baas et al., , 2016Stevenson et al., 2020). Alternatively, the banding can represent episodic near-bed turbulence suppression (Lowe & Guy, 2000). Lower structureless and intermediate banded divisions suggest deposition from LTPFs (sensu Baas et al., 2009), where mixing with ambient water developed the ripple-bearing upper division. However, the basal skim marks suggest that a bypassing laminar and concentrated head could have preceded the LTPFs because of the incompatibility between the skim marks and the turbulence state Baas et al., 2021). This tripartite structure is comparable to the H1 (lower division), H2 (intermediate division) and H4 (upper division) described in Haughton et al. (2009;Fig . 3). The lack of the linked debrite division (H3) suggests that the flows were able to deliver sand in a basinward direction (flow efficiency, sensu Mutti, 1979), and that they neither decelerated rapidly nor were very cohesive (Stevenson et al., 2020). The divergence between the tool marks (ENE-WSW) and the ripples (NNE) suggests the existence of seabed relief where the forerunning bypassing head was more influenced by seabed relief than the more dilute depositional part of the flow, although it could be related to two different flows .

Type D beds
Description: Thin-bedded (<0.2 m thick), matrixrich (>15%) sandstone, forming bipartite deposits, with sharp boundaries between its constituent divisions ( Fig. 7; S2, S7A and S7B sample beds). The lower divisions (<0.2 m thick) are matrix-rich (26 to 30%), structureless, fine-to medium-grained, weakly normally graded and moderately sorted (Fig. 7), and grooves with a wide range of directions. The upper divisions (ca 0.1 m thick) are massive and characterized by argillaceous sandstone comprising abundant mudstone clasts and plant fragments (0.5 to 2.0 cm in length; Fig. 7). Upper divisions lack thin section observation due to the fissile character of the argillaceous sandstones.
Interpretation: The juxtaposition of the lower and upper divisions without an intervening mudstone layer suggests that they represent the H1 and H3 divisions of hybrid event beds (HEBs; see Haughton et al., 2009;Fig. 3) genetically related to a single depositional event. The internal layering reflects discrete rheological changes within the source flow due to deceleration related to lateral spreading and/or increased near-bed flow concentration or cohesion . However, the lack of abrupt boundaries suggests that the flow did not develop stable internal stratification . The high matrix content (26 to 30%) and the crude normal grading within the lower divisions (H1) suggest the shearing of mud flocs and mudstone clasts caused by a transitional flow with a turbulent component. The upper argillaceous divisions (H3) suggest deposition from intermediate to high-yield strength debris-flow divisions (Talling et al., 2012). The preservation of plant fragments supports the interpretation of a laminar state within the upper divisions (e.g. Hodgson, 2009;D2 Type). Therefore, Type D beds are interpreted as thin-bedded HEBs formed under upper transitional plug flows (UTPF) and quasi-laminar plug flows (QLPF; sensu Baas et al., 2009Baas et al., , 2011Baas et al., , 2016. The development of H3 division, and lack of H2, suggest abrupt deceleration rates (Baas et al., 2011(Baas et al., , 2016Stevenson et al., 2020) related to the distal collapse of the flow . The grooves at the base indicate that UPTF/QLPFs were preceded by bypassing a forerunning debritic head responsible for incorporating cohesive clay from the substrate (Baas et al., 2021).
Interpretation: The internal layering suggests different rheologies related to HEB development (e.g. Lowe & Guy, 2000;Haughton et al., 2003Haughton et al., , 2009Talling et al., 2004;Barker et al., 2008;Kane & Pont en, 2012). The lack of abrupt boundaries suggests that the flow did not develop stable internal stratification . Loaded bases and mudstone rafts suggest shearing and entrainment of semi-consolidated substrate material and transported in suspension by flow-turbulence (see undersaturated flows of Eggenhuisen et al., 2017;Kane et al., 2017). The incorporation of clay-rich material and increased flow concentration promoted transitional flow behaviour (Kane & Pont en, 2012). Normal grading within the lower division (H1) suggests that the basal layer was fluid and turbulent enough to allow particle settling of the denser/coarsergrained fraction according to grain size/density (Baas et al., 2009) or a low yield strength basal division . Therefore, these matrix-rich basal divisions are not comparable to the clean sandstone turbidite base of HEBs in other systems (Haughton et al., 2003Talling et al., 2004;Amy et al., 2006;Barker et al., 2008;Davis et al., 2009) suggesting a different style of flow transformation. The abundant grooves indicate that these flows were preceded by a highly erosive bypassing debris flow and later infilled by the deposition of H1. The weakly normally graded, matrix-rich and mudstone clast-rich intermediate divisions (H3) suggest hindered settling as a sediment transport mechanism (Mulder & Alexander, 2001) and capacity-driven deposition (Hiscott, 1994a) rather than transport as a result of the matrixstrength of a high yield strength debris-flow division and cohesive freezing (Talling et al., 2012). The poorly-developed lamination on the upper divisions (H4) suggests high aggradation rates with poorly developed tractional reworking on the upper-stage plane bed. Alternatively, they may be related to deposition from parts of the flow that did not incorporate enough substrate and develop a poorly debritic division (Hussain et al., 2020;Baas et al., 2021), such as the tail. Additionally, they can be caused by mixing ambient water (Marr et al., 2001;Talling et al., 2002;Sequeiros, 2012). Therefore, Type E beds are interpreted here as thick-bedded HEBs (e.g. Kane et al., 2017) deposited from UTPF to QLPFs under high deceleration rates (Baas et al., 2009(Baas et al., , 2011(Baas et al., , 2016.

Type F beds
Description: Medium to thick-bedded (0.3 to 1.0 m thick) and generally matrix-rich (>15%) sandstones. The matrix content increases upward within a bed ( Fig. 9; S3 sample bed; 13 to 27%). Bed bases are sharp, loaded and erosive, comprising abundant ENE-WSW oriented grooves and planar to undulatory sharp tops. Beds are structureless, with poorly-developed diffuse planar lamination (S10 sample bed) that usually occupies the upper half of beds, although it can be developed from base to top. Beds are coarsegrained to medium or fine-grained, weakly normally graded to ungraded and poorly sorted, comprising a wide grain-size range, especially at the lower half (S3 sample bed; from very fine to granule). Locally, faint planar laminations or hummock-like bedforms develop near bed tops ( Fig. 9; S5 sample bed).
Interpretation: The high matrix content (up to 40%) and poor sorting suggest deposition from UTPF/QLPFs (Baas et al., 2009(Baas et al., , 2011(Baas et al., , 2016 characterized by an abrupt loss in capacity (Hiscott, 1994a), high sediment fallout and reduced tractional reworking (Talling et al., 2012). The crude normal grading and upward increasing matrix suggest some turbulence at the base of the parental flow, although the dominant rheology was likely an intermediate yield strength laminar flow (Talling et al., 2012). However, the abundance of grooves can be linked to the development of a laminar plug at the base of these flows or a forerunning and bypassing debris flow (Baas et al., 2021), which leads to entrainment of sediments (flow bulking, sensu Haughton et al., 2009). Structured tops, if present, suggest mixing with ambient water at the rear or upper parts of the flow, which reduces viscosity and yield strength (Talling et al., 2002;Sequeiros, 2012). Hummock-like bedforms are attributed to the development of combined flows with an oscillatory component, indicative of complex interaction between the parental flow and topography (Tinterri, 2011;Tinterri et al., 2016) or supercritical flow development (Cartigny et al., 2014, and references therein).

Type G beds
Description: Thick-bedded (0.5 to 6.0 m) sandy mudstone with sharp, erosive bases and undulatory tops. Type G beds contain abundant outsized, subangular sandstone clasts (0.05 to 1.0 m) supported by a poorly sorted and argillaceous matrix (Fig. 3). A medium-grained massive sandstone can locally underlie the argillaceous thick-bedded deposits with basal grooves containing deformed mudstone clasts (0.5 to 8.0 cm diameter; . Interpretation: The outsized clasts supported within the argillaceous matrix suggest cohesive freezing from intermediate to high yield strength debris flows ( Fig. 3; Talling et al., 2012;Sohn, 1997). Erosive bed bases and the presence of sandstone clasts suggest that the source flows entrained compacted substrate (Dakin et al., 2013;Hodgson et al., 2019;Mart ınez-Doñate et al., 2021). Alternatively, the debrites can also represent scour-fills post-dating previous erosive and bypassing flows. Mixing the cohesive laminar flow with ambient water might have diluted the basal part of the flow (Hampton, 1970;Mohrig et al., 1998;Marr et al., 2001). Therefore, both divisions are co-genetic (HEB type 1 of , with the basal sandstone representing deposition from a flow cohesive enough to develop grooves , such as lower to UTPF (Baas et al., 2009). Due to their high matrix strength and to avoid confusion with other bed types (especially Type D, E and F beds), Type G beds are referred to as debrites, despite the local development of transitional flow behaviour near their base.

Type H beds
Description: Thin to thick (0.1 to 4.0 m) sandstones with sharp bases and tops, containing abundant grooves on both margins. These sandstones are massive, fine to medium-grained and well-cemented, although rare parallel striations are present on blistered surfaces (1 cm high and <2 cm diameter; see Cobain et al., 2017) are also observed. They contain abundant mudstone clasts (2 to 5 cm diameter) and angular heterolithic rafts (10 to 50 cm length) that are oriented parallel to the bedding contacts (Fig. 3). Subhorizontal thick sandstones cross-cut the stratigraphy at low angles (<15°). Thick sandstones develop branching of thin-bedded sub-vertical bodies (<0.2 m thick) of similar lithology that cross-cut the stratigraphy at various angles (15°t o 90°).
Interpretation: These features are clastic injectites associated with compaction-driven fluid expulsion (Hurst et al., 2011). The subhorizontal thick sandstones are clastic sills with associated clastic dykes. Grooves at the base and top suggest laminar flow through lateral pressure transfer (Cobain et al., 2015).

PETROGRAPHY
Petrological analysis of the 40 thin sections allowed us to determine the matrix content, grain size and mineral composition of Type A to F beds (Figs 4, 5, 6, 7, 8 and 9). Matrix represents a significant component in the analysed samples, ranging from 3.5 to 40.3% and 19.8% on average. Most sampled beds show an increase in matrix content from base to top, with a few exceptions where the matrix content decreases upward (S4 and S8 sample beds; Figs 5 and 8). The thick-bedded sandstones, where point contacts dominate, contain less matrix than the thin-beds, which are matrixsupported. Samples are characterized by a wide grain-size range, varying from silt to granular, with the mean and median grain size being fine to medium-sand. Overall, bed bases are generally coarser, and the mean grain size decreases upward within the same event bed. This pattern is well-developed in low-density and high-density turbidites (Type A and B; Figs 4 and 5) and banded TFDs (Type C; Fig. 6). In contrast, HEBs (Type D and E; Figs 7 and 8) and TFDs (Type F; Fig. 9) can show ungraded profiles. When traced laterally, the grain size also decreases across the vertical profile of the event beds, developing (very) fine-grained and matrix-rich pinch-outs. High-density turbidites ( Fig. 5; Type B) and thick-bedded TFDs ( Fig. 9; Type F) are the deposits that comprise the coarsest grain sizes (medium to coarse sand), whereas the rest is finer (very fine to medium sand; Fig. 10).

Compositional trends
Description: The analysed samples were plotted on standard monocrystalline quartz-feldsparlithoclasts (Qm-F-Lt; Dickinson et al., 1983), volcanic-metamorphic-sedimentary lithoclasts (Lv-Lm-Ls; Ingersoll & Suczek, 1979) and polycrystalline quartz-volcanic lithoclasts-sedimentary and metamorphic lithoclasts (Qp-Lv-Lsm; Dickinson, 1985) ternary diagrams (Fig. 11). In the Qm-F-Lt diagram (Dickinson et al., 1983), nine samples plot in the recycled orogenic field, nine in the mixed field, and twenty-two in the magmatic arc field (Fig. 11A). From those twenty-two samples, nineteen correspond to the dissected arc field, whereas three are in the transitional arc field. The Lv-Lm-Ls diagram (Ingersoll & Suczek, 1979) reveals the dominance of volcanic lithoclasts over the other two lithic fragment types by plotting thirty-seven samples on the back-arc basin field and three on the forearc basin field (Fig. 11B). The Qp-Lv-Lsm diagram (Dickinson, 1985) plots twenty-seven samples on the magmatic arc field and thirteen on the mixed field (Fig. 11C).
Interpretation: The petrographic work and later plotting of the collected samples into different ternary diagrams reveal a consistent provenance field for the Early Jurassic Los Molles Formation in the Chachil Graben corresponding to the magmatic arc ( Fig. 11; Burgess et al., 2000) and supported by the north-east-directed palaeocurrents. Therefore, mixed source areas are unlikely to be responsible for the observed differences in depositional architecture and bed types within and between the Lower and Upper Lobe Complexes.

LOBE SUB-ENVIRONMENTS
The studied bed types are stacked into 3 to 7 m thick composite stratal units, bounded at the base and top by 0.2 to 1.0 m thick regional and walkable mudstones and/or debrites (Figs 12  and 13). These sandstone-rich packages represent intraslope lobes . Three bed type associations are identified based on interpreted sedimentary processes and depositional environments. These bed type associations represent lobe sub-environments, including lobe axis, lobe off-axis and lobe fringe. The lobes of each lobe complex show a different transition from axial to fringe sub-environments, which impacted the architecture of the individual lobes and the lobe complexes. In the lobes of the Lower Lobe Complex, the lobe axis passes to the lobe off-axis in ca 1 km and into the lobe fringe over 0.5 km and pinching out in ca 1 km (Fig. 12), whereas within the Upper Lobe Complex, the lobe axis passes into the lobe off-axis over a horizontal distance of ca 1.2 km, and from off-axis to fringe environments over 0.8 km and pinching out in ca 1.3 km (Fig. 12).

Lobe axis
Description: This lobe sub-environment is characterized by amalgamated medium to thick- Flow transformation processes in intraslope lobes 1397 bedded sandstones (<1.2 m thick) forming stratal packages with rare mudstone intervals between the sandstone beds (<0.1 m thick; Fig. 12). The lobe axes within the Lower Lobe Complex are ca 4.5 m thick and dominated by TFDs, with Type F beds dominating over Type E beds (Fig. 12). In contrast, the lobe axes of the Upper Lobe Complex are ca 7 m thick and entirely dominated by thick-bedded, matrix-poor high-density turbidites (Type B beds; Fig. 12).
Interpretation: The highly amalgamated thickbedded package represents a proximal environment dominated by deposition from energetic and high-concentration flows, supporting a lobe axis interpretation (Hodgson et al., 2006;Pr elat et al., 2009;Spychala et al., 2015). In the Lower Lobe Complex, the elevated concentration, cohesiveness and high-deceleration rates promoted the development of the transitional flow deposit (TFD)-dominated lobe axis comprising thick, amalgamated beds. In contrast, the axial parts of lobes in the Upper Lobe Complex were formed under erosional high-density turbidity currents. Despite the abundant mudstone clasts in Type B beds, no critical flow transformation developed due to the limited disaggregation of any contained clay, enabling preferential deposition of the coarsest grain fraction, and producing partial bypass of the finer sediments to more distal, down-current areas (Mutti & Ricci Lucchi, 1972;Mutti, 1992;Kneller & McCaffrey, 2003;Stevenson et al., 2015).

Lobe off-axis
Description: Well-stratified, heterolithic successions of mudstones (<0.1 m thick) and mediumbedded sandstones characterize the lobe off-axis sub-environment. The lobe off-axis shows a lower sandstone-mudstone proportion than the lobe axis due to the lower bed thickness and a reduced level of bed amalgamation (Fig. 12). Sandstone beds in the Lower Lobe Complex are thinner (0.1 to 0.5 m thick) than in the Upper Lobe Complex (0.1 to 0.7 m thick). The lobe offaxis deposits of lobes in the Lower Lobe Complex are ca 4 m thick and dominated by thin to medium-bedded HEBs (Type D and E). However, subsidiary thick-bedded TFDs (Type F), banded TFDs (Type C) and low-density turbidites (Type A) are also observed. The lobe off-axis deposits of the lobes within the Upper Lobe Complex are ca 4.5 m thick and show less variability and thicker deposits. It is dominated by thick-bedded TFDs (Type F) and medium to thickbedded HEBs (Type E), and rare low-density turbidites (Type A).
Interpretation: The range of thin to thickbedded argillaceous deposits documented within the lobe off-axis suggests a subenvironment of deposition characterized by high-concentration flows with variable internal stratification and rheology. The dominance of TFDs with mudstone clasts indicates substrate entrainment and flow bulking , which increases flow cohesion and concentration. Entrainment and disaggregation of clay promote the formation of oversaturated and cohesive flows  that could not transport their sediment load to distal parts of the lobe, thus promoting capacity-driven deposition (Hiscott, 1994a). Type C beds are interpreted as the product of a more moderate flow transformation and deceleration rate, where the cohesive forces are not dominant enough to produce flow collapse (Stevenson et al., 2020). The differences between the lobe off-axis deposits of the Lower and Upper Lobe Complexes (amalgamation, thickness and variability of the deposits) lie in the behaviour of the source flows and their depositional character in the proximal sub-environment (lobe axis). In the Lower Lobe Complex, as the thick-bedded TFDs and HEBs were already deposited in the lobe axis, the flows reaching the lobe off-axis developed thinner deposits (Type C, D and E beds). In contrast, in the Upper Lobe Complex, the high-density turbidity currents transformed into transitional flows over longer distances, leading to highdensity turbidite-dominated (Type B beds) lobe axis and medium to thick-bedded TFDdominated (Type E and F beds) lobe off-axis.

Lobe fringe
Description: This lobe sub-environment is characterized by a well-stratified, fine-grained heterolithic succession composed of mudstones (<0.25 m thick) and thin sandstone beds (<0.3 m thick). The sandstone-mudstone proportion is lower than the lobe axis and off-axis subenvironments (Fig. 12). Overall, lobe fringes from the Lower and Upper Lobe Complex are similar, being ca 3 m thick (Fig. 12). Both comprise thin-bedded HEBs (Type D beds), banded TFDs (Type C beds) and low-density turbidites (Type A beds). However, the Lower Lobe Complex shows a higher proportion of thin-bedded HEBs (Type D beds), whereas the Upper Lobe Complex comprises a higher proportion of lowdensity turbidites (Type A beds) and rare medium-bedded HEBs (Type E beds).
Interpretation: The thin-bedded heterolithic packages and lack of bed amalgamation suggest deposition in distal lobe settings. Thin-bedded HEBs (Type D beds) support flow transformation across the entire lobe, with flows that could not carry the sediment for a longer distance due to low flow cohesiveness, high deceleration rate, or both . The other fundamental deposit, the low-density turbidites (Type A beds), are related to the source flow's high flow efficiency and dilute nature, which largely bypasses proximal lobe subenvironments and are deposited in distal areas. Such flows are deflected and reflected from intrabasinal relief without collapsing, reworking the sediment and developing bidirectional ripple lamination (e.g. Kneller & McCaffrey, 1999). In these relatively distal sub-environments, banded TFDs (Type C beds) developed through flow deceleration, enhancing the cohesive forces over the turbulence (Stevenson et al., 2020). The rare presence of mediumbedded HEBs (Type E) in the Upper Lobe Complex is likely to represent flow transformation over a long distance from efficient high-density turbidity currents, enabling sediment transport and deposition in distal areas.

INTRASLOPE LOBE COMPLEX ARCHITECTURE AND BED TYPE DISTRIBUTION
The bed type associations described in the previous section are spatially and genetically related, representing different sub-environments within intraslope lobes. These intraslope lobes are stacked, forming two lobe complexes (Lower and Upper Lobe Complex), each bounded at the base and top by a >4 m thick regional mudstone or an erosionally-based debrite (Type G beds; Fig. 13; log 1 to log 7).
The sandstone-mudstone proportion was quantified for each bed type within each lobe for both lobe complexes in three different areas based on 764 sandstone beds (Fig. 14A). Logs 1 to 4, 5 to 6 and 7 to 8 represent the proximal, medial and distal locations, respectively (Fig. 1C). Logs 9 to 11 were not included in this analysis due to limited coverage of the investigated stratigraphic interval.

Lower Lobe Complex
The Lower Lobe Complex is ca 45 m thick and contains 10 stacked intraslope lobes. Due to outcrop limitations, the study of the stacking patterns within the Lower Lobe Complex is limited; however, the juxtaposition of lobe axis/off-axis and lobe fringe environments have been demonstrated to record compensational stacking of lobes (Pr elat & Hodgson, 2013), albeit with a strong aggradational component (see 'jig-saw architecture'; Marini et al., 2015;Fig. 13). Additionally, an overall basinward stepping stacking pattern of successive lobe pinch-outs is observed, recording the forestepping trend for the lobe complex (Fig. 13). Lobes achieve a maximum thickness of 4.5 m and comprise the thickest (<1 m thick) sandstone beds in their proximal parts (logs 1 to 4). These lobes are almost tabular, showing thinning rates of ca 0.4 m/km (Fig. 12). However, lobes pinch out abruptly north-eastward (between logs 7 and 9) over a horizontal distance of 2.4 to 2.7 km (Figs 1 and 2; Pic un Leuf u).
Apart from a complicated stacking pattern, the Lower Lobe Complex also records variability in sandstone content and bed type from proximal to distal (Fig. 14A). When traced down-dip over 2.4 km, amalgamation and sandstone-mudstone ratio decrease from proximal and medial areas (0.65) to distal areas (0.21; Fig. 14A). In proximal (logs 1 to 4) and medial (logs 5 and 6) parts, TFDs (Type D, E and F) dominate over lowdensity turbidites (Type A bed; proximal: 84% and medial: 87%). In the distal parts (logs 7 and 8), beyond the pinch-out of most HEBs and thick-bedded TFD (Type D, E and F), lowdensity turbidites (Type A bed) dominate over TFDs (75%), with occasional thin-bedded banded sandstone (Type C bed; Fig. 14A).

Upper Lobe Complex
The Upper Lobe Complex is ca 30 m thick and contains five aggradationally stacked intraslope lobes (Fig. 13). The lobes are up to 7 m thick in proximal areas (logs 1 to 4), showing their highest amalgamation ratio and the thickest beds (up to 1.2 m thick; Fig. 12). The lobes thin north-eastward from ca 7 m to ca 3 m thick, showing thinning rates of 1.2 m/km (Fig. 12) and pinch out over a horizontal distance of ca 3.3 km (Fig. 11)  In proximal parts, the Upper Lobe Complex contains a sandstone-mudstone ratio of 0.66 and is dominated by high-density turbidites (Type B beds; 98%; Fig. 14A). In medial areas, the sandstone-mudstone proportion is 0.75, and TFDs (Type E and F) dominate over turbidites (Type B; 87.7%; Fig. 14A). In distal parts, there is an abrupt decrease in sandstone-mudstone   (Log 1 to 4), medial (Log 5 to 6) and distal (Logs 7 to 8) parts of the Chachil Graben (see Fig. 1C for location). (B) Summary diagram of different flow evolution showing deposit variability from proximal to pinch-out areas. Denser flows tend to decelerate more abruptly and produce strong deposit variability, unlike dilute flows, which can maintain their efficiency (Mutti, 1979) for longer distances. The greater the incorporation of cohesive mud, the shorter the distance over which flows transform into low-efficiency cohesive flows, resulting in rapid flow collapse. proportion (0.18), and low-density turbidites dominate (Type A beds) over TFDs and HEBs (55.8%; Fig. 14A).

Controls on bed types and flow transformation
The flow transformations responsible for the reported variability in bed types are proposed to have been controlled by the incorporation of clay from the substrate and the flow deceleration rates (Fig. 14B). It is suggested that the flow transformations observed here were characterized by the interplay between complex longitudinal and vertical flow segregation processes. Longitudinal flow processes are characterized by high concentration bypassing flow heads (e.g. Sohn et al., 2002) responsible for the bedload transfer of mudstone clasts to slower and more depositional parts of the flow (Baas et al., 2021). 'Self-accelerating' flow cells at the front of the flow have been reported in physical modelling (sensu Sequeiros et al., 2009Sequeiros et al., , 2018 and are increasingly recognized in modern systems (Azpiroz-Zabala et al., 2017;Paull et al., 2018;Pope et al., 2022). However, in the rock record, evidence for the erosional and bypassing flow head is the abundance of grooves and high matrix content in Type B, C, D, E, F and G beds, as reported in this study. The formation of H1 divisions in Type C, D and E beds is inferred to be a product of transitional flows (Baker & Baas, 2020;Hussain et al., 2020) rather than turbidites (Haughton et al., 2003Davis et al., 2009). The inherited early post-rift topography ( Fig. 15A; Privat, 2019; likely promoted higher rates of flow deceleration, which could have led to the development of TFDs in proximal lobe sub-environments ( Fig. 14B; e.g. Fonnesu et al., 2015Fonnesu et al., , 2018Terlaky & Arnott, 2016;Brooks et al., 2018aBrooks et al., ,b, 2022Mueller et al., 2021). In contrast, the nonerosional low-density turbidites remained almost tabular without any transition into other bed types (Fig. 14B), indicating more stable flow conditions and efficiency than the denser transitional flows (Mutti, 1979).

Sediment routing pathways and mud availability
The mudstone-dominated substrate could have enhanced the development of matrix-rich deposits (Fig. 14B). The >100 m thick mud-rich package that draped the syn-rift topography (Figs 2, 13 and 15A) was available for entrainment into the overriding flows. The availability and incorporation of the muddy substrate (Fonnesu et al., 2015(Fonnesu et al., , 2018Mueller et al., 2021;Brooks et al., 2022) could have been enhanced by the immature sediment routing pathways , leading to poor channelization. Additionally, if the substrate was poorlyconsolidated, it could have been easily entrained (Terlaky & Arnott, 2014;Kane et al., 2017;Mart ınez-Doñate et al., 2021), promoting flow bulking . Furthermore, poorly-developed sediment routing pathways are unlikely to be as effective as channel-lev ee systems in removing the fines through flow stripping and overspill (Peakall et al., 2000;Hodgson et al., 2016). These effects could explain the dominance of TFDs in the lobe axis of the Lower Lobe Complex (Figs 12  and 15B). This contrasts with previous studies documenting the sandy nature of instraslope lobes (e.g. Spychala et al., 2015;Brooks et al., 2018a,b). These models highlight the deposition of the coarser fraction in the more up-dip locations, while the more dilute finegrained tops of the turbidity currents bypass down-dip (flow decoupling; Kneller & McCaffrey, 1999). However, in transitional flowdominated systems, the cohesive lower part would deposit matrix-rich poorly sorted deposits (TFD) while the upper and dilute division bypasses. This effect could also explain the TFD-dominated nature of lobes within the Lower Lobe Complex. In contrast, in the Upper Lobe Complex, flows travelled over sandier deposits (Fig. 13), which might have limited entrainment of the muddy substrate (Terlaky & Arnott, 2014). This interpretation is supported by the dominance of highdensity turbidites in the lobe axis and TFDs and HEBs in lobe fringes (Fig. 15B). Bed type proportions in medial parts of the Upper Lobe Complex are comparable to the proportions in proximal parts of the Lower Lobe Complex (Fig. 14A). A similar effect is observed at the lobe scale, with lobe axes of the Lower Complex and lobe off-axis deposits of the Upper Lobe Complex comprising similar bed type associations despite the different degrees of bed amalgamation (Fig. 12). This study highlights the role that variable substrate entrainment can play in intraslope lobes with a similar source.

Impact of topography on the pinch-out pattern
Seafloor relief affects the deceleration rates of sediment gravity flows, impacting the character of deep-water deposits (Pantin & Leeder, 1987;Edwards, 1991;Kneller et al., 1991;Edwards et al., 1994;Kneller, 1995;Al-Ja'Aidi et al., 2004). Depending on their flow properties, sediment gravity flows will respond differently when interacting with such relief, and this will influence the distribution of bed types in the lobes (Figs 12, 14B and 15B). The abruptness of pinch-out terminations (Al-Ja' Aidi et al., 2004;Bakke et al., 2013;Soutter et al., 2019) is strongly controlled by the yield strength of the source flow with limited flow efficiency (Mutti, 1979;Hiscott, 1994a). This effect is observed in the Chachil Graben, where the lobe complex terminations offset the pinch-out of the contemporaneous debrites (Fig. 13). This effect is not limited to laminar flows; it also impacts the transitional flows producing thin-bedded, matrix-rich HEBs (Type D) that pinch out abruptly against the frontal topography (Fig. 14B). Denser flows are more sensitive to intrabasinal topography than dilute flows/divisions (e.g. T} ok es . This statement is supported by the high divergence in palaeoflow reported from the multidirectional grooves, indicative of the sensitivity of TFDs  to obstacles (for example, debrite rugosity; Muzzi Magalhaes & Tinterri, 2010) and basin-scale topography (for example, compaction hinge; Fig. 15A). In contrast, the dilute (parts of) flows did not decelerate rapidly enough, or failed to entrain enough substrate mud, to suppress flow turbulence (Al-Ja' Aidi et al., 2004;Talling et al., 2012). This results in extensive thinbedded low-density turbidites (e.g. Smith, 2004) or banded TFDs offsetting the pinch-out of HEBs, depositing on relative topographic highs, as seen with the pinch-out pattern documented in the study area (Fig. 14B). The ripple and flame structures indicate that proximal areas (logs 1 to 4) show an overall north-east trend, with deflection/reflection indicators ( Fig. 13; . Another indicator of complex flow behaviour resulting from the interaction between transitional flows and seafloor relief is the development of hummock-type structures ( Fig. 9; S5 sample bed). These structures are evidence of combined flows characterized by an oscillatory component produced by internal waves and reflected bores (Tinterri et al., 2016). Combined-flow bedforms (S5 sample bed, Type F beds; Fig. 9) are well-developed in the Lower Lobe Complex, where a higher interaction with topography is inferred (Fig. 15A).

Steepness of the confining slope and comparison with other systems
Intraslope lobe complexes of the earliest Los Molles Formation were deposited under the influence of a south-facing, oblique counterslope formed by a compaction hinge above the El Luchador fault block Figs 2, 13 and 15). Reconstructing the steepness of the confining slope during this earliest post-rift stage is not straightforward, where onlap terminations against basin margins are not developed. Deepwater systems interacting with gentle topography (i.e. a fraction of a degree) tend to produce a complex pinch-out of the sand-rich lobe component, characterized by an aggradational succession of tens of metres of thin beds containing climbing ripples. Published examples include the Welsh Basin (Wales, UK) (Smith, 2004) and Karoo Basin (South Africa) (Spychala et al., 2017a,b). In contrast, on steeper slopes (i.e. >10°), the characteristic termination style is onlap of sheet-like sandstone beds unconformably terminating directly onto the basin margin, such as in the Annot Basin, France (Kneller & McCaffrey, 1999;Sinclair, 2000;McCaffrey & Kneller, 2001;Bakke et al., 2013;Soutter et al., 2019) and the Laga Basin in Italy (Marini et al., 2015). In the Chachil Graben, the abruptness of pinch-outs, lack of climbing ripple lamination, and the bed types reported in this contribution suggest that the compaction hinge-related counterslope was steeper (>1°) than those reconstructed for the Welsh (Smith, 2004) and Karoo Basin (Spychala et al., 2017a,b) but less steep than those reported in the Annot (Du Fornel et al., 2004;Soutter et al., 2019) or Laga Basin (Marini et al., 2015). Further quantifying the angle in the Chachil Graben is challenging due to the similarity between the organic-rich calcareous mudstones that draped the rift topography  and the high-efficiency silty (Mutti, 1979), lowdensity turbidites offsetting the sandstones (distal fringe; Boulesteix et al., 2020). In addition, postdepositional compaction-driven deformation (burial-related) and the draping nature of the more distal deposits might lead to overestimating the counterslope angle (Pickering & Hilton, 1998;Bakke et al., 2013). Another indicator of seafloor relief is the presence of an injectite network (Figs 13 and 15A;Cobain et al., 2017;Hansen et al., 2019). This suggests that early post-rift topography is not solely limited to inheritance from the previous extensional phase but can be locally enhanced by compaction-related deformation, promoting long-lived seabed relief that largely influenced the nature of the intraslope lobes.

Controls on stratigraphic architecture
The differential flow deceleration and deposition produced by the frontal topography in the Chachil Graben promoted a thicker accumulation of lobe-related deposits over the depocentre lows than on the topographic highs (Figs 13, 15A and 15B; Smith & Joseph, 2004). In the Lower Lobe Complex, the abruptness of the pinch-out terminations, the tabularity of the intraslope lobes (4.5 m thick and <1 m thinning), and the multidirectional grooves and ripples suggest that a complex frontal seafloor topography was a strong control on geometry and architecture. The Lower Lobe Complex (ca 45 m thick) is characterized by a progressive down-dip advancement of the pinch-out position from base to top (Figs 13 and 15A). This pinch-out pattern suggests infilling, flattening and smoothing the complex inherited rift topography over time ( Fig. 15B; e.g. Ross et al., 1994;Hay, 2012). The stacking of 10 intraslope lobes shows a compensational pattern (e.g. Mutti & Sonnino, 1981;Deptuck et al., 2008;Pr elat et al., 2009) with a strong aggradational component, where topography confined the lobes frontally but enabled lateral compensation. The authors interpret that during the deposition of the Lower Lobe Complex, the deep-water system was confined (Winker, 1996;Sinclair, 2000;Southern et al., 2015) at the lobe scale but semi-confined (Marini et al., 2015;Dodd et al., 2019) at the lobe complex scale. In the Upper Lobe Complex (ca 30 m thick), the five intraslope lobes are thicker (7 m thick), show lower thinning rates and can be traced over longer distances (>3.3 km) than in the Lower Lobe Complex. This architecture and termination style suggest that seafloor topography impacted individual lobes less. Given that the pinch-out terminations developed over the compaction hinge, flows were likely to deflect towards the east (north-east) after interacting with the south/south-west-dipping oblique-tofrontal oriented (see Kneller et al., 1991;Soutter et al., 2021) compaction hinge. The five intraslope lobes show an aggradational stacking pattern, juxtaposing similar sub-environments on top of one another. This reveals that, even if the Upper Lobe Complex and its lobes recorded a reduction in frontal confinement (compared to the Lower Lobe Complex), the system was more confined laterally at the lobe complex scale, limiting lobe-scale compensational stacking. In other words, the stratigraphic evolution presented here shows an upward decrease in frontal confinement at both lobe and lobe complex scales; however, there is an increase in the degree of lateral confinement. The authors postulate that the increasingly aggradational stacking pattern recorded from the Lower to the Upper Lobe Complex is unrelated to depocentre-scale topographic modification but rather to an increase in sediment input and flow efficiency (Mutti, 1979). Additionally, the continuous slope degradation and emplacement of debrites might have flattened the topography, favouring the connectivity along the slope and efficiency of the flows. This highlights that the concept of confinement is related to the balance between the accommodation and the size of the depositional system and their source flows. Considering the two lobe complexes as a lobe complex set (Pr elat et al., 2009), this larger unit shows a progradational stacking pattern (e.g. Hodgson et al., 2006;Grundv ag et al., 2014) related to the maturation of the sediment routing pathway (Hodgson et al., 2016) and smoothing of the inherited topography ( Fig. 15A and B). This study highlights the impact of inherited early post-rift topography in deep-water siliciclastic systems, from the bedscale to lobe complex-scale (Jackson et al., 2008;.

Los Molles intraslope lobes as an analogue for subsurface lobe systems
This section discusses the Los Molles intraslope lobes as an analogue for subsurface reservoirs for hydrocarbon extraction or carbon storage. Grain size, sorting, matrix content and distribution are key parameters controlling porosity and permeability values (e.g. Bell et al., 2018). Overall, the mean grain size, amalgamation and sandstone/mudstone proportions decrease from lobe axes to fringes, whereas matrix content and sorting increase. The deposits found at the lobe axis are likely to comprise the best reservoir properties of the lobes due to the abundance of highdensity turbidites (only in the Upper Lobe Complex) and the high degree of amalgamation, which promotes communication between sandstone beds (Begg & King, 1985;Desbarats, 1987;Deutsch, 1989). In contrast, the decrease in amalgamation, grain size and the development of argillaceous deposits or divisions in the lobeoff axis and lobe fringe environments may result in decreasing permeability values (intrabed baffles; Porten et al., 2016;Southern et al., 2017;Bell, 2019).
The development of abrupt HEB-rich frontal pinch-outs, and confinement produced by compaction-related topography, suggests a high stratigraphic trapping potential with reduced leakage risk. The coarser grain sizes and less argillaceous deposits of the Upper Lobe Complex suggest higher permeability values than the TFD-dominated Lower Lobe Complex (e.g. Southern et al., 2017). However, in the subsurface, the matrix content does not necessarily hinder the reservoir properties because the volume of clay coating around grains, such as chlorite, can inhibit quartz cement growth and, therefore, preserve primary porosity (Heald & Larese, 1974;Ehrenberg, 1993;Anjos et al., 1999;Dowey et al., 2012). In addition, the development of injectites that pinch out within the mudstones suggests that sand bodies are sealed and may form stratigraphic traps (postdepositional; Hurst et al., 2005Hurst et al., , 2008Scott et al., 2013) although, in some cases, they can also represent leakage paths (Monnier et al., 2014;Cobain et al., 2017). Further studies that integrate detailed petrographic and architectural data in these and similar deposits are needed to improve our understanding of how these systems might act as hydrocarbon reservoirs, aquifers and carbon or hydrogen storage sites.

CONCLUSIONS
Field mapping, log correlation and facies analysis of two exhumed early post-rift intraslope lobe complexes show consistently different characteristics. Characterization at grain-scale using quantitative petrographic methods suggests that transitional flow deposits dominate the argillaceous Lower Lobe Complex, and beds pinch out more abruptly than in the overlying turbiditedominated Upper Lobe Complex. Compositional analysis suggests that both lobe complexes are volcanic arc sourced and, therefore, that the source area was not responsible for the differences in bed types between the lobe complexes. The preferred interpretation is an autogenic response of the depositional system due to the progressive healing of intraslope relief, maturation of sediment transport routes, and reduction in mud entrainment from the substrate. The compaction hinge(s) acted as a barrier for the denser and cohesive (parts of) flows, but the dilute turbidity currents could override and deposit on the intrabasinal slope. This study highlights the primary role of subtle inherited relief developed during the early post-rift stage of basin development. This integrated petrographic, architectural and sedimentary process model provides new insights into how these systems act as hydrocarbon reservoirs, aquifers and carbon storage sites. A better understanding of grain type and grainsize segregation from proximal to distal parts of deep-water systems is critical for carbon sequestration and predicting CO 2 plume dynamics.

CONFLICT OF INTEREST
The authors declare no conflict of interest for this work.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author.

Supporting Information
Additional information may be found in the online version of this article: Figure S1. Density plots of S1-S7B sample beds. Figure S2. Density plots of S8-S15 sample beds.