Lockne is a concentric impact structure due to a layered target where weak sediments and seawater covered a crystalline basement. A matrix-supported, sedimentary breccia is interlayered between the crystalline breccia lens and the resurge deposits in the crater infill. As the breccia is significantly different from the direct impact breccia and the resurge deposit, we propose a separate unit name, Tramsta Breccia, based on the type locality (i.e., the LOC02 drilling at Tramsta). We use granulometry and a novel matrix line-log method to characterize the sedimentology of the Tramsta Breccia. The obliquity of impact combined with the layered target caused an asymmetric, concentric transient crater, which upon its collapse controlled the deposition of the breccia. On the wide-brimmed downrange side of the crater where the sedimentary target succession was removed during crater excavation, wide, overturned basement crater ejecta flaps prevented any slumping of exterior sediments. Instead, the sediments most likely originated from the uprange side where the brim was narrow and the basement crater rim was poorly developed, sediment-rich, and relatively unstable. Here, the water cavity wall remained in closer proximity to the basement crater and, aided by the pressure of the collapsing water wall, unconsolidated black mud would flow back into the crater. The absence of interlayered resurge deposits in the Tramsta Breccia and the evidence for reworking at the contact between the overlying resurge deposits and the Tramsta Breccia indicate that the slumping was a rapid process (<75 s) terminating well before the resurge entered the crater.
The Lockne crater (Fig. 1) is a well-preserved and well-documented marine-target impact structure in central Sweden (e.g., Lindström et al. 1996, 2005a; Sturkell 1998a). The impact occurred about 458 Ma during the upper Sandbian Stage of the early Upper Ordovician (Ormö et al. 2010a, p. 1217) in the epicontinental sea that covered much of Baltoscandia at that time (Ormö et al. 2007). At the Lockne impact site, the sea was about 500 m deep and deepening toward the west where a foredeep formed in front of the approaching Caledonian orogeny (Lindström et al. 2005; Shuvalov et al. 2005; Heuwinkel and Lindström 2007; Ormö et al. 2002, 2007 and references therein). Below the seafloor the target included 50 m of mainly consolidated limestone with marly beds, 30 m of a dark mud (today “alum shale”), and a crystalline basement with a horizontal peneplain. The crater has a concentric morphology (cf. Quaide and Oberbeck 1968) with a shallow outer crater in the sedimentary strata surrounding a 7.5 km wide, deeper, nested crater in the basement (Lindström et al. 1996, 2005a). Concentric crater morphologies can be caused by primarily two different processes: Either due to extensive slumping of a poorly consolidated target sequence surrounding the direct area of impact (e.g., Chesapeake Bay impact structure), or by a shallow excavation flow developing in the upper weaker target layer (see discussion in Ormö and Lindström 2000; Ormö et al. 2012). The Lockne crater is an example of the latter process (Ormö and Lindström 2000).
The Lockne crater is characterized by its extensive water-transported resurge deposits (definition in Ormö et al. 2007) that inside the basement crater is represented by an up to several tens of meters thick, normally graded unit beginning with coarse clastic Lockne Breccia fining into sand-siltstone Loftarstone both containing mixed clasts from the whole target stratigraphy (Lindström et al. 1996, 2005a; Ormö et al. 2007). The crystalline rock and limestone have easily distinguishable clasts in the resurge deposits, whereas the dark mud was mainly blended into the matrix or transported as clay balls. Cores drilled in the basement crater found that the graded resurge deposits show a rapid transition to an underlying, matrix-supported breccia (Ormö et al. 2007) that in drill core LOC01 is 1 m thick, but that in LOC02 reaches a thickness of about 91 m (this study). This breccia was previously thought to represent a lower part of the resurge deposits (e.g., Lindström et al. 1996). However, as this report will show, the breccia is in many ways strongly distinctive from the resurge deposits, suggesting a different mode of formation, and, thus, a separate unit and a different name. As customary for lithologies related to the Lockne crater (cf. Lindström et al. 2005a), we give it the local name Tramsta Breccia based on the name of the area where the drill cores LOC01 and LOC02 containing the type lithology are located (Fig. 2). At present, the Tramsta Breccia has only been observed in drill cores. We agree with Lindström et al. (2005a) on the advantages of using a local name before a generic name as the latter can be used regardless of changes in interpretation. The sedimentology and mode of formation of the Tramsta Breccia is the main objective for this study.
The target material (here: water, unconsolidated sediments, lithified sediments, and crystalline rock) has a strong influence on the cratering and subsequent modification processes resulting in a spectrum of crater morphologies and lithologies. Here we compare the sedimentary infill sequence of the Lockne crater with the observations of the nearly contemporaneous Kärdla crater, Estonia, and the Paleogene Chesapeake Bay impact structure, Virginia, USA, (henceforth CBIS) (Fig. 1). With good knowledge of the target (e.g., different units of different strength), the parameters used in modelings of the impact process are better constrained, thus further improving the modeling result.
The target material also has a great influence on the formation and modification of the crater rim. In craters such as Lockne, with a relatively thick, weaker layer covering the basement, the nested crater rim may lack the structurally uplifted part causing it to obtain only a fraction of the height of a rim of a similar-sized crater formed in crystalline rock (Ormö and Lindström 2000). Likewise, a large amount of poorly consolidated sediments sandwiched into a crystalline crater rim may decrease the stability. This can lead to slumping of the whole rim or sections thereof (e.g., the Wetumpka crater, King et al. 2006; the Kärdla crater, Puura and Suuroja 1992; CBIS, Gohn et al. 2009). Consequently, the crater rim construction has a direct influence on the processes during the modification stage and, thus, the geology and morphology of the final crater. A rim that reaches higher than the top of the target succession (sea level) can block the entry of a resurge. This can be the case if the impact takes place in shallow target water. In these cases, there is a combined effect of a more pronounced structural uplift of the rim as well as the relatively shallow water in relation with the crater size (Ormö and Lindström 2000). This will efficiently cause the rim to block all entry of the seawater into the crater directly after the event, e.g., Gardnos (Kalleson et al. 2008, p. 40). The crater will eventually be water-filled as water penetrates through the fractured rock, but this process cannot be described in cataclysmic terms. A different scenario occurs when a rim is formed that rises higher than the water surface, but starts to collapse directly into the newly formed crater, depositing slumped rim material before the resurge enters the crater, e.g., the Wetumpka and Kärdla craters (Puura and Suuroja 1992; King et al. 2006). When the restraining effect of the rim has been removed by the collapse, any poorly consolidated sediments in its vicinity will lose their support and slump toward the open crater, e.g., the Wetumpka (King et al. 2006) crater and CBIS (Gohn et al. 2009). This will continue until the slope is stable, but with the inclusion of the rim wall, a considerable amount of material can be emplaced (slumped), and the collapsed zone can expand to a great distance outside the nested, basement crater. The process of slumping can occur prior to (e.g., Kärdla, Suuroja et al. 2002), or simultaneously (e.g., CBIS, Gohn et al. 2009; Ormö et al. 2010a), with the resurge, depending on the relative size of the collapsed zone and the magnitude of the resurge (e.g., CBIS, Gohn et al. 2009; Ormö et al. 2010a). In such cases, resurge deposits may occur interlayered with slumped material, which, in turn, may be incorporated in the resurge and redeposited as disintegrated, smaller bodies that can be difficult to distinguish from other ripup material.
The drilling in the crater moat (LOC02) was done in 1994 and aimed to penetrate the thickest sequence of crater infill before reaching the basement crater breccia lens. This crystalline breccia as well as crystalline basement crater ejecta is collectively called the Tandsbyn Breccia by Lindström and Sturkell (1992) from the type locality at the village of Tandsbyn. What was suggested to be Tandsbyn Breccia was reached at the 310.92 m level and until the drilling commenced at the 335 m level (Lindström et al. 1996). However, it is uncertain whether this is the top of the autochthonous breccia lens or a huge block of crystalline breccia forming part of what we are here defining as the Tramsta Breccia. This can be compared to what has been described from the LOC01 core drilled in the same year just inside the northwestern rim of the basement crater (Fig. 2) (Sturkell et al. 1998). In this core, Tandsbyn Breccia was reached at 90 m depth after passing through the postimpact sediments and the resurge sequence. The drilling continued into the Tandsbyn breccia, but at 129.3 m depth, black shale was encountered; however, this layer was only one meter thick. It was suggested that this represented the true crater floor (Sturkell 1998a). Below the Tandsbyn Breccia the drilling encountered an intensively fractured crystalline rock and ended at 225.15 m. Later on, Sturkell et al. (1998) interpreted the core sequence such that the Tandsbyn Breccia in the 90–129.3 m core interval is a part of the crater wall that slid down over the clay that formed a lining of the true crater floor. The wall of the basement crater shows fault terraces that can be seen today just inside the apparent basement crater rim. This deformation of the crystalline upper part of the crater would have taken place just before the resurge entered the crater.
No suevite-like breccia has so far been observed in any of the Lockne drillings and if present it must be located deeper than the 335 m level reached by LOC02. Likewise, no major melt body is expected as no diagnostic magnetic anomalies are observed within the crater (Sturkell and Ormö 1998). At the moment, without any additional available information from deep drilling, we consider that the crystalline basement crater breccia lens was reached by the LOC02 drilling and that the slumped Tramsta Breccia is situated above that breccia lens, but below the resurge deposits. The depth to the true crater floor remains unknown for the deeper parts of the Lockne crater.
The aim of this study is to illustrate the development of the Lockne marine-target crater during the earliest phase, i.e., preresurge, of the crater modification stage.
Geological Setting of the Compared Craters
The basement in the Lockne area is dominated by a K-feldspar megacryst bearing (Revsund) granite, dated to 1.86–1.85 Ga (Högdahl 2000). This granite grades in places into a medium grained type of similar age. The oldest crystalline rocks (probably >1.9 Ga) consist of a suite of metavolcanic rocks (felsic–mafic), which can be found in the southeastern part of the structure. Metasedimentary rocks, para and orthogneissic rock of similar age can be found directly south of the structure. The youngest of the Proterozoic rocks are the 1.2 Ga dolerite sills. A peneplain developed on the Proterozoic basement before the onset of early Paleozoic marine sedimentation in the area. The complete preimpact sedimentary strata (Cambrian-Middle Ordovician) measure about 81 m, reconstructed from a drill core (Brunflo-2) just outside the crater and observed sections (Sturkell 1998b; Ormö et al. 2007, p. 1933). The lower Cambrian sediments rest directly on the weathered peneplain, which is commonly covered by residual material with grain sizes spanning from clay to blocks in size. This material makes up the base of the lower Cambrian transgressive deposits that frequently begin with a thin sand layer. The thickness of this basal conglomerate unit is often <0.5 m. The middle and upper Cambrian is completely dominated by a black shale, which is often referred to as “alum shale” in literature. We choose to use this term in this study. At the time of impact (Upper Ordovician), the Cambrian deposits were still organic rich, argillaceous, black mud, which had not yet been lithified into today's “alum shale” (cf. Kirsimae et al. 1999). The about 32-m thick Cambrian strata also include thin limestone beds or limestone lenses with their numbers increasing in the upper Cambrian. This limestone is bituminous (commonly called “stink stones”) and can easily be distinguished from the Ordovician limestone. The Upper Cambrian is terminated by an erosive surface at different stratigraphic levels, which is overlain by Ordovician strata. The upper-most part of the Tremadocian beds (Lower Ordovician) is observed at some localities, whereas at other localities, the lowermost Floian limestone (2.4 m) directly overlies the Cambrian beds. The sedimentation during Floian to mid-Darriwilian time is not disrupted by any major erosion events. It consists of green shale (7.3 m), bedded limestone (6.8 m), and a marly gray limestone (16.7 m). The upper Darriwilian stage consists of a bedded limestone, which is both red and gray in color. The thickness of this unit is about 17 m. The thickness estimate is not as precise because no drilling penetrates the uppermost part of the Darriwilian stage. The upper-most meter of the sedimentary part of the target is in the marly Dalby Limestone (Sandbian). In total, the preimpact Ordovician is about 50 m thick of which 26 m consists of “orthoceratite limestone” that was lithified at the time of impact (von Dalwigk and Ormö 2001), 17 m of marly limestone that was easily disintegrated by the impact (Lindström et al. 2008), and 7 m of a greenish mud that today occur as shale.
The layered target consisting of materials of various strength over a rigid horizontal basement contributes to the crater obtaining a concentric morphology. This produced a 7.5 km wide nested crater in the basement surrounded by an approximately 14 km wide outer crater where, in addition to the seawater, most of the sedimentary succession was excavated (See discussion in Ormö et al. 2012). Ormö et al. (2002) used 2-D numerical simulation constrained by Lockne's concentric morphology to show that the target water depth must have been at least 500 m. Similar results were obtained by 3-D numerical simulation of an oblique impact for the Lockne event (Lindström et al. 2005b; Shuvalov et al. 2005). In Lockne, the sedimentary part of the target was progressively eroded at higher stratigraphical levels with increasing distance from the edge of the inner crater (Sturkell 1998b). The lithified, but often marly and bedded, limestone was strongly affected by the outwards-directed, shallow excavation flow as well as the collapsing water cavity wall. The limestones were shattered into what Lindström et al. (2008) described as a “water-blow breccia,” which they named Ynntjärnen Breccia from its type locality at Lake Ynntjärnen. However, parts of the cohesive Cambrian clay clearly withstood the erosive phase prior to the emplacements of the overturned flaps. Excavations in the Nordanbergsberget quarry (NBB in Fig. 2) show the best example of how the proximal part of the nested crater ejecta flap rests on a surface freed of most of the sedimentary target succession except for various amounts of the dark Cambrian alum shale. During ejecta emplacement, the flaps fragmented into large units (see Fig. 2), both because of the force at emplacement and the geometrical spreading effect when moved from the relatively smaller area of the crater toward the periphery. The Cambrian mud from the substrate was pressed up in cracks through the flaps (Fig. 3), giving further evidence for how the alum shale was in the form of soft mud at the time of impact. Possibly, some of the dark mud also squeezed out (Fig. 3) at the hinge zone back into the newly formed crater. However, the contribution of squeezed out material is probably minor considering that only thin beds of alum shale occur under or just outside the crystalline ejecta flaps. The hinge zone of the flaps for most parts slumped back into the crater during the early crater modification.
The basement crater ejecta flap formation was strongly affected by the layered target and obliquity of impact with wider flap on the downrange (western) side, and a much less pronounced flap on the uprange (eastern) side of the structure. Likewise, the outer crater has a down-range shift (Lindström et al. 2005a; Ormö et al. 2012). These parameters were revealed by extensive geological mapping and used as input parameters for numerical modeling of the impact angle and target water depth (Lindström et al. 2005b; Shuvalov et al. 2005). At the time of the impact, the Lockne area was located on the southern hemisphere and possible with a rotated position relative to today. The trajectory of the projectile was from today's east with, most likely, a 45° angle over the horizontal plane (Lindström et al. 2005b). The circumstance of having a thick, weaker layer of seawater and sediments in the upper part of the target caused not only a concentric morphology, but also the near absence of a structural uplift component of the nested, basement crater rim (Ormö and Lindström 2000; Sturkell and Lindström 2004). In comparison with the nearly two kilometer high water cavity wall, the crater rim formed in the sea floor was far too low to provide an obstacle for the resurge.
The 4 km wide mid-Ordovician Kärdla impact structure in Estonia (Fig. 1) is in nearly pristine condition, as it was buried under postimpact sediments soon after its formation. The impact structure was formed in a 50 m deep shallow sea (Suuroja et al. 2002) with a 142 m thick sedimentary cover over a pre-Cambrian peneplain. The target sedimentary sequence consists of an Ordovician sequence comprising 14 m of lithified limestone and 8 m of sandstones on top of 120 m of “weakly cemented” Cambrian sand-, silt-stones, and clays (Suuroja et al. 2002). The “weakly cemented” Cambrian “rocks” behaved during the crater modification as unconsolidated sediments, and could, when not supported by the crater rim, slump into the crater. The fresh crater rim is estimated to have had most parts reaching well above sea level (fig. 2 in Puura and Suuroja 1992). However, soon after its formation, much of the rim gravitationally collapsed into the crater, leaving it intact only in the NE quadrant and parts of the SW quadrant. This allowed the slumping of exterior sediments into the basement crater during the short time interval when it was dry before the onset of the resurge.
The stratigraphy of the crater infill is given by Suuroja et al. (2002). It consists, from bottom to top, of >800 m of brecciated crystalline basement rocks, 44–52 m of suevite-like breccia, 42–140 m of slump breccia, 2–14 m of fallback breccia, and 8–30 m of normally graded resurge deposits. The slumped material originates from the collapsed sections of the crater rim and the unlithified clastic sediments outside it. The slumped sequence is overlain by a layer of fallback ejecta that must have been deposited only moments before the resurge deposits were laid down.
Chesapeake Bay Impact Structure
The approximately 35.4 Ma old Chesapeake Bay impact structure (CBIS) is located beneath the ocean front of south east Virginia including the present Chesapeake Bay and adjacent ocean. It is today completely buried and, thus, only visible through geophysical methods and drilling (Horton et al.  and references therein). The structure is one of the best-known representatives of a marine-target impact striking a sea shelf. The near-shore setting caused the water depth to vary from a few tens of meters in the west on the shoreward side to about 300 m on its eastern side toward the ocean (Horton et al. 2005). The target sequence below the seawater consisted of Cretaceous and Paleogene sediments resting on an eastward sloping (toward the ocean) crystalline basement (Fig. 4). The thickness of the sedimentary strata increases from approximately 400 m in the west to approximately 1500 m in the east (Fig. 4) (Horton et al. 2005). The sedimentary target strata comprise three major units: Cretaceous non-marine and marine sediments, and Paleogene marine sediments. All of these sediments were poorly consolidated at the time of impact, which led to extensive collapse and inward slumping during the modification stage that significantly expanded the crater (e.g., Gohn et al.  and references therein). The impact created a 30 to 38 km wide crater in the crystalline basement that is surrounded by approximately 24 km wide outer annular trough created by the collapse and inward slumping (Collins and Wünnemann 2005), which together makes a final crater with a diameter of up to 86 km (e.g., Powars and Bruce 1999). Thus, the final concentric shape of the crater and the outer rim of the annular trough were strongly influenced by the target stratigraphy, in this case the thick pile of un-consolidated sediments resting on a crystalline basement. The stratigraphy of the infill gives an idea of the sequence of events as reconstructed by Gohn et al. (2008, 2009) and Ormö et al. (2009) based on the 1766 m deep drilling into the central crater at Eyreville and a shorter core from Langley located at the outer part of the annular trough. The deepest unit penetrated by the Eyreville core consists of crystalline basement blocks and minor amounts of a lithic breccia followed by suevite (fallback material) and lithic breccia overlain by sand with lithic blocks. This is followed by megablocks of granite originating as slumped material from the wall of the crystalline crater. This is overlain by slumped material from the sedimentary strata of the collapsed crater moat zone. Higher up in the slump sequence, there is a greater influence of resurge material (Gohn et al. 2009; Ormö et al. 2009). This polymict, sediment-dominated breccia with huge blocks of semi-consolidated sedimentary rock is followed by postimpact, secular sediments. Thus, the marine environment further enhanced the modification stage with resurge and slumping combined.
Material and Method
This study is a continuation of the line-logging and description of the resurge deposits in the Lockne crater presented by Ormö et al. (2007). The data from the upper part (158–220 m) of the LOC02 drill core were interpreted by Ormö et al. (2007) as a water transported resurge breccia. The lower part (220–310 m) of the drill core has not been studied in detail before and the sorting and mixing of lithologies differs considerably from the resurge breccia. To analyze this section of the LOC02 core, we applied the line-logging method used by Ormö et al. (2007, 2009) in their analysis of the resurge deposits in drill cores LOC01 and LOC02 from Lockne and the Eyreville core from the CBIS. This method is particularly useful at Lockne where the geology of the target and the different lithological units that formed in connection with the impact are well known. The different clasts (limestone, shale, and crystalline rock) are easily recognized in the drill cores and the source of the clasts can be determined with maximum certainty. In this study, we have also modified the method to include logging of the variations that occur in the matrix of the studied breccia. We analyzed in detail the part of the LOC02 core (Figs. 5, 6, and 7) that spans the former “matrix supported Lockne Breccia” (Lindström et al. 1996) that we herein rename as Tramsta Breccia.
The clast line-logging was performed along a thin line drawn along the visible “center” of the core (cf. Ormö et al. 2007, 2009). The visible length-axis and the lithology were determined for every clast (≥5 mm length axis) that touched the line in accordance with the method described by Ormö et al. (2007, 2009). We visualize the clast size variation in ½ phi intervals for each meter (Fig. 8a). As the logged part of the core is more than 150 m, spanning both Lockne Breccia (resurge deposits) and Tramsta Breccia, only a few representative histograms were selected (Fig. 8b), but all the data are used in a three-axis plot (depth, phi, and counts in each ½ phi category), which shows the histograms on a countered surface plot (Fig. 8a). The ratio of crystalline versus limestone clasts is calculated by dividing the number of crystalline clasts with the total number of clasts (crystalline + limestone clasts). In the case when a few, but large, clasts are presented as clasts per meter, the ratio based on numbers of clasts can be skewed.
The line-logging provides a mathematical quantification of the composition and sedimentology of the breccia and its variation caused by the target geology. From line-log data, the number of clasts (Fig. 6), average grain size and standard deviation (Fig. 7), the clast size distribution (Figs. 8a and 8b), and type of clasts are calculated and presented per meter core (Fig. 9).
The matrix line-logging is performed much in the same way as the clast line-logging along a central line drawn along the core. Any material made up of fragments smaller than the cut-off size of the clast log (here 5 mm) is said to represent matrix (Fig. 10). In this particular case, there are two main matrix categories: dark mud and polymict sand. The top and bottom level of each matrix category where it crosses the line is noted with the thickness of the matrix having to be equal or thicker than the clast cut-off size (i.e., 5 mm). Two distinct successions can be observed, clast abundant and clast poor (Fig. 7). Each matrix type continues until the visible beginning of the next, independent of the occurrence of clasts larger than the cut-off size. That is, the matrix occurring just above a logged clast is said to continue until the lower end of the clast even if that end is in direct contact with another matrix type. The plotting of the results from the matrix log produces an occurrence graph much resembling a barcode. To know if a certain “bar” in this “barcode plot” (Fig. 10) is completely formed by matrix or partially from a large clast, it is necessary to complement the matrix log with the occurrence of a large clast, here ≥7 phi (128 mm), and the clast lithology to evaluate any particular patterns that may appear in the plot (Fig 10).
In addition, smaller (<7 phi) claystone lenses (Fig. 5) in the resurge deposits are displayed as dots (Fig. 10). Although mostly completely disintegrated by the forceful water resurge, the cohesive force of the clay seems to have caused some of the clay to be lumped together today occurring as flattened lenses in the core. We find it unlikely that flat chunks of clay would survive the transport. Thus, we suggest that they formed clay balls that after the deposition by the resurge became flattened to lenses by the overload of postimpact sediments. This enables us to compare the amount of clay in the resurge section of the core with that of the matrix-supported part of the core (i.e., the Tramsta Breccia).
In addition to the sedimentological study, we have performed numerical simulation to investigate (1) the fate of the target sediments during the cratering process and (2) if sediments can remain as emplaced masses within the transient crater and in this way provide a viable source for the observed Tramsta Breccia. The numerical simulation uses iSALE based on the SALE (simplified arbitrary Lagrangian Eulerian) hydrocode (Amsden et al. 1980) that has undergone several improvements (e.g., Collins et al. 2004). For methodological reasons, the impact is assumed to have been vertical. The granitic projectile is set to be spherical with 600 m diameter and an impact velocity of 15 km s−1 (cf. Shuvalov et al. 2005; Lindström et al. 2005b). We model a three-layered target with 500 m water covering an 80 m thick limestone layer on top of a granitic basement. In an iterative process, we perform simulations with several spatial resolutions in the range 2–5 m of each cell to validate the numerical results. The cell number of the mesh is adjusted to cover 12 km in the radial direction and 7 km in the vertical direction. Due to radial symmetry, we simulate half of the physical domain. Twelve thousand sediment tracers have been assigned in the model (Fig. 11). We also analyzed the maximum temperature experienced by the sediment tracers during their mobilization within the growing transient cavity for time frames 1.0–4.0 s. In this way, it is possible to compare the modeled results with direct observation of the current appearance of sediments in the analyzed breccia.
Comparing the results from the clast log with the published results on the resurge deposits (i.e., above 221.69 m depth) in the same core (Ormö et al. 2007) shows that the value of the average clast sizes together with the standard deviation changes along the core and reflects the degree of sorting (Fig. 7). Where there is a small variation (average and standard deviation) in succeeding meters, and the clast sizes form a bell-shaped distribution, it indicates that the material is relatively well sorted.
With the lithological information, the ratio of crystalline versus limestone clasts can be calculated, ranging from 0 to 1. If the ratio has small variations (about ±15%) for consecutively one meter intervals, the lithologies are considered to be well mixed as in the water transported Lockne breccia (see Ormö et al. 2007). However, if the ratio has a large variation spanning much of the scale, this indicates a lithologically heterogeneous material. In the 150 m long logged section (Fig. 9), two distinct areas stand out. These are the upper part (down to 220 m depth) with relatively stable ratio in succeeding meters and the lower part, where the ratio shows strong fluctuations. For the upper part, the water resurge mixed the two lithologies well, but in the lower part, no or little mixing took place, suggesting a different process of transport and deposition.
The core logging presented by Ormö et al. (2007) shows that above approximately 220 m depth in the LOC02 core, the clast size distribution displays a general fining upward sequence. Ormö et al. (2007) suggest that the sequence down to the 220 m level represents the resurge deposits, i.e., water-transported sediments. The transition to the underlying Tramsta Breccia is chaotic over a few meters, possibly because of rework by the resurge. The Tramsta Breccia has many fewer clasts per meter (Fig. 6), and in sections only a clay matrix. At about the 295 m level, large clasts start to be more frequent (Fig. 7), and at this level, the material in the Tramsta Breccia becomes more basement rock dominated (Fig. 9). This transition is gradual over several meters in the core.
Likewise, the clast composition below 220 m level is more heterogeneous, and the matrix content can be completely dominant or even absent for some intervals of the core. Similar to the sorting (see above and Fig. 7), the transition from the upper resurge to the lower slumped deposit is gradual showing evidence of mixing (reworking) in the interval 218–222 m (Fig. 10). In the section of the core below the approximately 220 m boundary, the dominant matrix type is a dark gray or black claystone, which clearly originated from the target alum shale. In the section 245–268 m, the black clay constitutes more or less the whole core with exception of a few large clasts. These are mainly stink stones from the same stratigraphical target unit. Clasts originating from the sedimentary target sequence are dominant down to approximately 295 m depth, below which crystalline basement clasts become more frequent, indicating an inherited stratigraphic stacking order that mimics the target stratigraphy, suggesting large-scale slumping.
From the line logging of the polymict breccia in the LOC02 drillcore, two distinct lithological sections can be distinguished. The upper section, which reaches down to about the −220 m level, is (Figs. 6-10) sorted and well mixed (crystalline and limestone fragments) with small amounts of clay and a high number of clasts per meter. In contrast, the section below shows all the opposite characteristics (Figs. 6-10). Sorting is absent and it is poorly mixed. This section of the core contains plenty of matrix (clay) and has a lower number of clasts per meter. We suggest that two different depositional regimes are at hand; the upper section is water transported and sorted resurge and the lower section is formed in a slump-dominated regime.
The selected time frames from the numerical simulation in Fig. 11 show both the maximum temperature experienced by the sediment tracers introduced in the model (Fig. 11 a1–3) and the absolute temperature at each location in the target at the specific time frame. It is obvious that sedimentary material remains spread along the floor of the growing cavity, but also that all of this material is subject to very high temperatures. Only outside the crater rim, the temperatures are low enough not to cause visible alteration of the dark mud of interest here. Nevertheless, no obvious thermally altered sediments have yet been encountered in the core.
The drill core LOC02, retrieved in the moat of the inner crater, contains about 90 m (220–310 m) of a rock, which is unsorted, poorly mixed, and contains plenty of matrix (clay) with a lower number of clasts per meter. The overlying resurge deposits are better sorted, show greater mixing, and higher number of clasts per meter (Ormö et al. 2007). We can, after logging the lower part of the LOC02 core and with the knowledge from modeling and impacts formed in marine condition (with different water depth), present an interpretation of the deposits. Due to the obvious differences between the resurge deposits and the underlying sequence, we suggest it to be a separate unit with a different genesis. We give this unit the name Tramsta Breccia.
The lack of sorting, poor mixing, and large amount of matrix indicate that a relatively low amount of water was available during transport. The matrix originates from the Cambrian part of the target and is present in disproportional amounts relative to the preimpact sedimentary succession. This characteristics and lower sequence's stratigraphical position indicate that the Tramsta Breccia was formed in the time interval after the initial collapse of the transient crater (i.e., crystalline breccia lens formation) and before the arrival of the resurge. Lindström et al. (2005b) saw in their numerical simulation that it took about 40–60 s between impact and when the water movements turned from an outward to an inward direction at a location set to 10 km from the touchdown point (i.e., the initiation of the resurge). Ormö and Miyamoto (2002) modeled the time for the resurge to reach the basement crater rim with an assumed location for the water cavity wall 12 km from touchdown. In their simulation, it took the water 100 s to reach the crater rim and an additional 50 s to reach the crater center. This would give a time window of about 100 s for the slumping and the formation of the Tramsta Breccia. However, simulations by Shuvalov et al. (2005) show that the water cavity wall most likely was positioned only about 3 km outside the basement crater rim on the downrange side and 2 km outside the rim on the uprange side. This significantly reduces the time available for the slumping, especially from the uprange side (see discussion below). Ormö et al. (2010b) made a simple 2-D simulation of the Lockne event assuming a vertical impact. This simulation gives a time frame of the initiation of the resurge at 34 s, and at 90 s, the water has almost reached the basement crater center. Unpublished time frames from this simulation show that the resurge reaches the basement crater rim at the time frame of about 75 s. This gives the time interval available for the displacement of the Tramsta Breccia material. The occurrence of blocks of the para-autochthonous impact breccias, i.e., the Tandsbyn Breccia and Ynntjärnen Breccia within the Tramsta Breccia, shows that these lithologies were already formed at the onset of the slumping.
At the CBIS, numerical modeling shows that the collapse and slumping of the poorly consolidated target sediments surrounding the nested basement crater took about 5–6 times longer time than the excavation and early stage modification of the basement crater (Collins and Wünnemann 2005). Slumping was initiated somewhat before the water resurge. Sedimentological evidence shows that the resurge started during the slumping phase and occurred simultaneously with the ongoing slumping. The resurge phase peaked slightly after that of the slumping (Gohn et al. 2009; Ormö et al. 2009). The modeling suggests that the main collapse and resurge modification of the CBIS were over after just about 10 min. During this time, the crater had grown as a result of headward slumping far beyond the reach of the basement crater (Fig. 4), which has a diameter that more closely corresponds to the magnitude of the event as calculated from the reconstructed size of the transient crater (Collins and Wünnemann 2005).
Similar to the situation at Lockne, the target stratigraphy for CBIS is well known and can be divided into lithologically distinct units that can be recognized in the resurge deposits (Ormö et al. 2009). The circumstance that the target water depth in CIBS varied from a few tens of meters in the west to a couple of hundred meters deep on the ocean-facing side in the east most likely generated a nonradial symmetry of the resurge. Ormö et al. (2010b) used projectile impact experiments to analyze the effects of such a varied target water depth on the resurge dynamics. They describe it as a consequence of the size relation between the projectile diameter and the water depth. This theory for the development of concentric craters in layered targets states that if the projectile diameter is larger than the water depth (the shallow water side of the crater), the excavation of the water layer will follow that of the seafloor, and the water ejecta curtain will be attached to the seafloor crater ejecta curtain (e.g., Ormö et al. 2002; Shuvalov and Trubestkaya 2002). For the opposite situation, with a projectile diameter equal to or less than the water depth (the deep water side of the crater), a shallow excavation flow develops along the interface between the weaker water layer and the rigid substrate. The water cavity wall and ejecta curtain will form detached from the seafloor crater ejecta curtain, i.e., forming a concentric crater although only on this side. Thus, the water cavity wall will stand farther out from the seafloor crater rim on the deep water side than on the shallow water side. The experiments indicate that the resurge from the shallow water side of the crater would be the first to reach the seafloor crater rim as the water cavity wall is standing closer to the crater on this side. However, the resurge from the deep-water side of the crater is more forceful and moves faster than the shallow-water resurge. The deep-water resurge may even flood the inner crater before the shallow-water resurge. With enough difference in force between the resurge flows, the deep-water resurge can stop the shallow-water resurge and push it back out over the rim (Ormö et al. 2010b). At present, any potential radial variation in the resurge deposits at CBIS is not well known as only two drillings have been analyzed in this respect. These are the Eyreville and the Langley cores (Ormö et al. 2009). The expansion of the CBIS by slumping and resurge erosion stopped as the central crater filled up with sediments. The farther out in the annular trough, the more of the preimpact sedimentary sequence is preserved, i.e., higher portion of material from the upper part of the target sequence was observed in the upper resurge deposits in the Langley core (Ormö et al. 2009).
The slumped material within the CBIS rests upon a suevitic breccia deposited before the resurge in a similar manner as in Kärdla (cf. Suuroja et al. 2002, p. 1132). Seismic profiles passing the CBIS indicate a poorly developed basement crater rim. The centro-symmetrical outline of the collapsed annular trough shows that slumping of the exterior target sedimentary succession could pass the basement crater rim from all directions. The transition from slump deposits into deposits with interlayering of resurge deposits finally terminating with resurge deposits sensu stricto suggests that this slumping was initiated at the time the water was expelled from the target area, but continued simultaneously with the resurge (Gohn et al. 2009; Ormö et al. 2009). On the other hand, at Kärdla, the incomplete collapse of the basement crater rim seemed not to have allowed any extensive exterior slumping, although it opened access to the entry of the resurge.
In Lockne, most of the sediments in the outer crater were removed by the shallow excavation flow that formed the outer crater during the crater excavation stage. On this eroded surface, overturned flaps of Proterozoic crystalline basement rocks were emplaced by the excavation of the nested basement crater. The near complete removal of the sedimentary strata of the outer crater left no steep wall of sedimentary rocks facing the nested crater. In places, Cambrian clay remained on the floor of the outer crater and became covered by the flaps, but we estimate that only minor amounts of the clay may have been squeezed out and back into the nested crater from the hinge zone of the flaps. This clay (Fig. 3b) cannot contribute to the large volume of Cambrian clay observed in the crater infill. In addition, the well-developed overturned flaps on the downrange side of the Lockne basement crater prevented any large-scale slumps of the target sediments into the crater from that direction.
If the sediments constituting the Tramsta Breccia did not slump back from the downrange side of the crater, then what alternatives remain?
Our numerical modeling shows that some of the target sediments remain inside the crater distributed along the floor of the growing cavity (Fig. 11). This is a consequence of the material flow lines during the crater excavation stage. Much material is moved downwards, outwards, and some also eventually upwards to end up in the near-field ejecta layer. In a case such as the very plastic material as the dark target mud at Lockne, it would smear along the transient crater floor much like the lining of melt in the standard crater model (cf. Melosh 1989 and references therein). However, the modeling does also show that the sediments will experience temperatures of more than 1500K. The dark mud matrix and other material occurring in the investigated Tramsta Breccia do not show any trace of thermal alteration. We therefore conclude that the origin of the bulk of material in the Tramsta Breccia must have originated from outside of the basement crater rim. As this is not possible from the downrange side, then the slumping must have occurred from the uprange direction. In Fig. 12, we illustrate our interpretation of the most likely scenario for the formation of the Tramsta Breccia. The process is a consequence of the concentric crater formation in the layered target, and the uprange/downrange asymmetry of the crater caused by the oblique impact. For simplicity and clarity of the diagram (Fig. 12), we have neglected the water resurge that occurred after the slumping of the Tramsta Breccia material, and likewise, the postimpact water layer that came to cover the whole final crater. The locations of the drill cores are projected on the profile. At the LOC01 drilling, the 40 m large block of crystalline Tandsbyn Breccia rests upon a meter thick layer of black shale (the dark clay). It is interpreted as a slab of Tandsbyn Breccia that slumped into the crater at about the same time as the Tramsta Breccia was emplaced, i.e., just before the onset of the resurge.
Process reconstruction for the downrange (west) side (Fig. 12):
Excavation—Concentric crater growth with shallow excavation of outer crater in water and sediments. Outgoing water flow is strong enough to remove limestone also outside the reach of the water cavity (Lindström et al. 2008). Some sediment, mainly dark mud, remains smeared along the floor of the expanding basement cavity, however heated to high temperature.
Modification—Flap deposition pushes the remaining ductile sediments (i.e., dark mud) outward while some get trapped in pockets. The downrange ejecta flaps prevent slumping of sediments, but fragments of mainly limestone and crystalline ejecta will be transported by the water resurge. Some of the minor amounts of strongly heated sediments remaining inside the basement crater (mainly dark mud) slump back to its deeper parts.
Seafloor crater at the end of modification—Resurge deposits and secular sediments cover parts of the crater exterior and all of the inner crater.
Process reconstruction for the uprange (east) side (Fig. 12):
Excavation—The water cavity and the basement crater ejecta flap are relatively smaller on the eastern side of the crater (cf. Shuvalov et al. 2005). This causes the target sediments to remain closer to the rim, and the rim to be relatively poorer developed than on the downrange side.
Modification—Dark mud is mobilized (white arrow) and flows past the crystalline rim (compare with the western rim where extensive flap prevents slumping). The process is possibly aided by the near vertical collapse of the approximately 2 km high water cavity wall that presses down on the sediments.
Seafloor crater at the end of modification—Dark mud slumped from outside the basement crater rim is covered by resurge deposits and secular sediments.
Our interpretation of the line-logging data, experiments, and modeling is that the Tramsta Breccia formed by material (mainly dark mud from the target sediments and fragments from the crystalline rim) that slumped down into the Lockne basement crater shortly before the water resurge entered the crater. Similar slumping of the target sediments across the annular trough and the central basement crater's rim occurred at the CBIS but much more extensively. At the Kärdla crater, the few breaches in the basement crater rim were not sufficient to allow extensive slumping of exterior sediments, although permitting the entry of the water resurge.
Because of the occurrence of crystalline flaps that rest on progressively higher stratigraphic levels of the target sedimentary sequence out to 5 km downrange from the basement crater rim at Lockne, only minor amounts of clay could enter from this side. The matrix-dominated breccia can be described from the logging result in terms of clasts per meter, sorting (standard deviation), ratios of different components, and the occurrence of the clay (shale) as main constituent to the matrix in the “barcode” plot, which is introduced in this study. Based on numerical simulation, we suggest that some sediments did not leave the crater, during the cratering phase, but that they were heated to over 1500 K. However, the clay retrieved in the drilling in the crater moat shows no sign of having been exposed to high temperatures and, thus, cannot originate from the center of the crater, but from the higher parts of the crater wall and the close vicinity outside the basement crater rim and the rim itself. It is possible that the heated material occurs at deeper levels in the crater infill that have not been reached by the drilling.
Most likely, the slumping of the material of the Tramsta Breccia occurred from the uprange (east) side of the crater where, as a consequence of the oblique impact into the layered target, the nested, basement crater rim was poorly developed and the necessary amount of target sediments (mainly dark mud) remained in close proximity to the basement crater. The deposition of the Tramsta Breccia is bracketed between the excavation phase and the entry of the resurge. This gives about a 75 s long time window for the formation of Tramsta Breccia.
The authors thank their late friend and colleague Maurits Lindström (formerly at Stockholm University) for help, inspiration, and fruitful discussions. This study received support from the Swedish Research Council (Vetenskapsrådet)-funded project (90449201), the “Swedish Deep Drilling Program” (SDDP) in the frame of the “International Continental Scientific Drilling Program” (ICDP). The work by J. Ormö is supported by grants AYA2008-03467⁄ESP and AYA2011-24780/ESP from the Spanish Ministry of Science and Innovation. The numerical simulation presented in this work was carried out by Alain Lepinette. We also express our gratitude to Alasdair Skelton for improving the English language, and to the reviewers David S. Powars and Paul C. Buchanan for valuable comments on the manuscript.