Rock substrate and proto/cataclasites
In the NCA, with their long-ranging deformation (see Setting), the density of jointing seen in the 'intact' Hauptdolomit (point 1 in Fig. 2, Fig. 4a) is typical. Dolomite deforms in a ductile manner only above 450–500°C (Dietrich 1983); in the NCA, such a temperature was not attained during folding and thrusting (cf., Kralik et al. 1994). The proto/cataclasites that overlie the intact Hauptdolomit along a gently dipping, planar contact (point 2 in Fig. 2, Fig. 4b) may represent damage zones of faults, such as a thrust ramp or a low-angle normal fault. Within the interval of proto/cataclasites, no fault core was seen; together with the observation that the interval is underlain by intact Hauptdolomit, this suggests that the proto/cataclasites represent an erosional vestige of a fault footwall. A few kilometres south of the study area, along the southern margin of the NCA, low-angle detachment faults of Jurassic to Cretaceous age are documented (Ortner and Reiter 1999). The observation that the proto/cataclasites are onlapped and overstepped by Gosau deposits that were not subject to wholesale cataclasis underscores that the putative hangingwall (of whatever fault type inferred) was eroded before onlap.
In thin section, the proto/cataclasites show a size continuum of rock fragments generated by fracture and attrition, down to a matrix of lithified doloclastic gouge (Fig. 4c–f); fragments up from approximately fine pebble size display jigsaw to mosaic subclast fabrics. In total, from field to thin section, the proto/cataclasites are well-comparable to damage zones (not fault cores) in carbonate rocks (cf., Billi et al. 2003; Billi 2005; Hausegger et al. 2010; Ortner et al. 2018). The low dip of the proto/cataclasite interval (cf., Fig. 2) suggests that it is a vestige of a damage zone related to folding and thrusting (e.g., Tarasewicz et al. 2005) or to extensional faulting (Agosta and Aydin 2006; Demurtas et al. 2016; Ortner et al. 2022). Alternatively, the interval of proto/cataclasites may be interpreted as a rhegolith that originated from the densely-jointed 'intact' Hauptdolomit by – up-section increasing – translation/rotation of joint-delimited clasts (cf., Fig. 4a) relative to each other. In this case, the clasts should be embedded in a secondary matrix infiltrated from the exposed area above. Indeed, small 'pockets' of red-coloured internal sediment – presumably related to karstification during subaerial exposure – are locally present in the proto/cataclasite (Fig. 4d), but the overwhelming part of the matrix is of cataclastic origin (Fig. 4e, f). This supports the interpretation of the proto/cataclasites as a damage zone related to a pre-Gosau fault.
Slope breccias
These breccias differ from the cataclasites in that: (a) they consist of or contain a sizeable proportion of clasts of subangular to subrounded shape, (b) clasts are so much translated/rotated relative to each other that clast boundaries cannot be matched, (c) despite very poor sorting, there is a distinct jump in grain size between the breccia clasts and the matrix, and (d) these breccias are stratified (Fig. 5b–d). Nevertheless, both the larger clasts and the clastic matrix are derived from the local dolostone substrate (Fig. 5c, f), and no extraclasts derived from other sources was identified.
These breccias are interpreted as hillslope mantles mainly formed by slow transport processes such as downslope creep. If the breccias had formed, mainly at least, by relatively dilute and rapidly-propagating cohesive debris flows, typical features of such flows such as inverse grading and vertical size sorting of clasts, a-axis imbrication of larger clasts, and intercalated levels of fine-grained material deposited from overland flow in between debris-flow events should be present (cf., Sanders et al. 2009; Ventra et al. 2013); all this was not observed. The significance of localized presence of crackle– to jigsaw subclast fabrics (Fig. 5c, e) in these breccias is dealt with farther below.
Stream conglomerates
In transgressive situations such as the one considered herein, conglomerates may accumulate along a beachface or from streams. Beachface conglomerates in the Gosau Group commonly are associated with fan deltas and contain a matrix of mixed carbonate-lithic/siliciclastic arenite and, rarely, marine bioclasts (Wagreich 1989; Sanders 1997, 1998, 2001). The conglomerates of interest herein, in contrast, consist only of dolostone clasts within a sparse doloclastic matrix. Therefore, an interpretation in terms of stream conglomerates is preferred. Unfortunately, the outcrops are not of sufficient extent to support an unequivocal distinction between beachface or stream conglomerates. In the studied conglomerates, the good rounding and high sphericity of clasts (Fig. 6a–d) suggests a transport of a few kilometres at least. Better preserved successions of unequivocal stream deposits of similar character are exposed at the base of the Gosau Group a few kilometres farther east of the study area (Sanders and Gruber 2023), and support the possibility that such conglomerates are locally present. The question of localized in-situ clast fracture in the conglomerates (Fig. 6e, f) is treated farther below.
Section with event bed
Intervals 1 and 3: In interval 1 (Fig. 7), the bioclasts of corals and rudists, small wood fragments, and angular clasts of dolostone come from different environments: fragments of wood and dolostone clasts from areas onshore, and corals and rudists from shallow-marine settings. The lime-muddy bioturbated matrix (Fig. 8b–d) records a setting of overall low energy, yet the marine- and land-derived components suggests mixing during episodic high-energy events (cf., Mount 1984). To carry land-derived components into a shallow-marine setting requires offshore-directed transport such as upon coastal submergence by storms or tsunami (see below for discussion). The angular shape and the 'rugged' outline of the dolostone clasts (Fig. 8d, e) suggests that these clasts had lodged within soil or in other sediments (scree?) before reworking and embedding in the matrix of interval 1. Relative to the record in the overlying interval 2, the postulated high-energy events that led to component mixing were of limited impact. Notably, all of the characteristics of interval 3 are closely similar to those of interval 1, and record an overall quiet shallow subtidal setting ('bay', 'lagoon') punctuated by high-energy events of limited outreach with respect to component mixing.
Interval 2: As described, the sharp-based interval 2a (Fig. 7) is an orthobreccia of fine-pebbly to small-cobbly dolostone clasts in a sparse matrix of lithic arenite (Figs. 9b–f, 10a). The shapes – from angular to embayed to well-rounded (Fig. 10b) – and surfaces from smooth to rugged, of the dolostone clasts, suggest that they came from different environments: Well-rounded, smooth clasts may from a stream bed and/or a beachface; the prevalent fraction of angular clasts with rugged and embayed surfaces, in turn, might have rested on land in soil, scree slopes, or debris flows. No lithoclast infested by marine macroborers was observed. With the potential exception of serpulids and cyanoids that may tolerate schizohaline/brackish conditions, the bioclast spectrum is dominated by taxa indicative of normal-marine environments (Table 1). The portions of interval 2 that, in their matrix, display an equigranular coarse calcite spar with 'amoeboid' crystal boundaries, and with components apparently floating within, represent pseudosparite (Fig. 10d–f). The Hjulstrøm curve suggests that to transport small cobbles ~ 10 cm in size even if only by traction requires a flow ~ 2–3 m/s in velocity, which provides a crude impression of minimum current velocities required to keep clasts in motion. Because at least most of the lithoclasts of interval 2 were mobilized from terrestrial areas, this implies that coastal submergence was followed by an offshore-directed flow strong enough to transport small cobbles.
Deposition of interval 2 is envisioned as follows: During an event of exceptional coastal flooding, a wide onshore strip was submerged by a surge with shallow-marine bioclasts. During or close to the end of submergence, seaward backflow transported clasts up to at least cobble size from the submerged onshore area to the sea. At the site of observation (Fig. 7), this backflow deposited part of its coarse load. The lack of organization with respect to clast size across the bed, and the matrix of winnowed arenite all suggest that deposition was rapid, such that the clasts collided with each other in a traction carpet (e.g., Todd 1989; Sohn et al. 1997, Sohn 2000; Moore et al. 2011). The subtle bipartite division of interval 2 may result from a twofold backflow, or from a transformation of a single backflow by deposition mainly of lithoclasts (cf., Sakuna et al. 2012).
Most features considered typical of storm-induced deposition were equally plead to support an origin from tsunami (cf., Bahlburg et al. 2010; Shanmugam 2012; Paris et al. 2018). Along active margins, seismic faulting and submarine mass-wasting can generate tsunami runup > 10 m in amplitude (e.g., Borrero et al. 2001; Bohannon and Gardner 2004; Locat et al. 2004; Spiske and Bahlburg 2011; Spiske et al. 2014; Ryan et al. 2015; Takashimizu et al. 2020). Tsunami typically run up with 5 to ~ 20m/s (e.g., Jaffe and Gelfenbaum 2007; Yeh and Li 2008; Spiske and Bahlburg 2011), sufficient to mobilize boulders. Tsunami impact typically comprises several runup/backwash pulses, but the first pulse usually is highest. Offshore 'tsunamites' thus may comprise a single or several layers, respectively (e.g., Massari et al. 2009; Sakuna et al. 2012). For geological preservation, the highly erosive tsunami backwash (up to > 10 m/s) of hyperconcentrated flows or grain flows is more relevant than uprush (Fujiwara et al. 2000; Le Roux and Vargas 2005; Dawson and Stewart 2007; Yeh and Li 2008; MacInnes et al. 2009; Paris et al. 2010; Spiske and Bahlburg 2011; Varela et al. 2011; Feldens et al. 2012; Spiske et al. 2014; Gusman et al. 2018; Slootman et al. 2018). In the present example, flooding of a hilly coastscape was followed by a backflow that carried the mix of land-derived lithoclasts and marine bioclasts into a gently-dipping shallow subtidal environment (cf., Sakuna-Schwartz et al. 2015). The thickness of interval 2, and its composition of clasts up to cobble size, are well-comparable to historical offshore tsunamites (cf., Sakuna et al. 2012, their Table 2). The sharp contact between interval 2 and 3 perhaps results from partial reworking of the original top by storms or swell (cf., Weiss and Bahlburg 2006).
Storm-induced flooding also comprises flood and ebb surges with amplitudes similar to tsunami (e.g., Coch 1994, Borrero et al. 2004; Weisberg and Zheng 2006; Robertson et al. 2007). Pebbles to cobbles from land may initially have been mobilized during storm flood (velocity up to a few m/s; cf., Robertson et al. 2007), then carried offshore by ebb surge. In contrast to a wave period of ~ 10 min to > 1 h of tsunami (Weiss and Bahlburg 2006), storm surges wane over 5 to 10 h, implying a relatively gentle ebb flow (e.g., Weisberg and Zheng 2006). Storm ebb surges may cut channels in the shoreface, and can transport sediment out to a few hundreds of metres off shore into waters a few metres deep (Coch 1994). 'Tsunamites' typically comprise 1 to 4 layers, but the same was also observed in historical storm layers (Phantuwongraj and Choowong 2012). Similarly, mixing of fossils from different environments and angular bioclasts are considered typical of bay/lagoonal 'tsunamites', but the same can also be achieved by storm surges (Fujiwara et al. 2000; Dawson and Stewart 2007; Varela et al. 2011). Field data on the bottom velocity of storm-induced offshore flows are unavailable; most data are from measurements 1 m or more above sea bed, where flow might be swifter than directly above ground. On an open shelf, offshore storm flows and surge ebbs may attain rare peak velocities of 1-1.5 m/s at 3 m above sea bed (Morton 1981). In macrotidal estuaries on wide gently sloping shelves, spring tides superposed with typhoons may induce ebb surges up to > 2 m/s at sea surface (Zhang et al. 2012). Numerical modeling of storm surge ebb on open shelves indicates velocities of up to 1.6 m/s sustained over less than an hour (Suter et al. 1982). In any case, even if transferred 1:1 to the sediment surface, these flow velocities are clearly on the lower verge required for transport cobble-sized sediment.
Alternatively, interval 2 might be interpreted as a debris flow that originated on a hillslope on land (without flooding) and intruded into the shallow marine area. This seems precluded because: (1) The arenite matrix with marine bioclasts is inconsistent with the unfossiliferous terra rossa-like matrix as it is widespread in terrestrial deposits of the Gosau Group; (2) If it is postulated that the debris flow diluted by uptake of sea water and mixed in marine-derived material (sand, bioclasts) to the extent that its original matrix was completely exchanged, it should have passed a fully turbulent state including suspension while propagating under water, i.e., it was no longer a debris flow. It further seems doubtful whether the low gradient of the shallow marine area the flow intruded into provided enough slope energy to allow for such a process. As discussed, the debris-flow like features of interval 2 can be explained by rapid deposition from a grain flow within a traction carpet along the base of a flow of unknown total depth. In summary, the event bed is tentatively interpreted in terms of a 'tsunamite' because it seems most straightforward to derive from processes associated with tsunami runup and backflow.
Cairns of breccio-conglomerates
The dolostone clasts in the cairns of breccio-conglomerates (Fig. 12) may stem from the debouch of small streams or fan deltas, or from a transgressive beachface. The clast- to matrix-supported fabric, and the poorly-defined clast fabric aside of a crude parallel alignment of clast [a,b] planes suggest rapid deposition. If the breccio-conglomerates formed by tsunami backflow from flooded coastal areas, a much more diversified mixture of bioclasts and lithoclasts may be expected. Redistribution of beachface to upper shoreface sediments into final deposits typically takes place during beach reformation upon waning storms or swells (Sanders 2000). In summary, these deposits are preferably interpreted as shore-zone deposits accumulated during waning of storms or swell.
Reconstruction of depositional setting
In attempting an environmental reconstruction, the stratigraphic contacts in the mapped area are reiterated (Fig. 13). The most complete synthetic section (column A, Fig. 13) comprises all lithologies up to a first interval of shoreface conglomerate. Choosing the boundary between intact Hauptdolomit and overlying rocks as a correlation datum, it is seen that the proto/cataclasites are not everywhere present below the basal Gosau deposits (columns B and C, Fig. 13); yet, slope breccias also are not present everywhere in contact with the Hauptdolomit, and so are the stream deposits (Fig. 13). These relations may be explained by: (a) partial subaerial erosion of the protocataclasite interval, leading to (b) redeposition of protocataclasite fragments into slope breccias, succeeded by (c) incision of a stream that, upon (d) channel backfilling due to base-level rise onlapped and overstepped the slope mantles (Fig. 13). Presumably later, and stratigraphically higher up, a very shallow marine-subtidal setting ('lagoon') established that contains the section with the interpreted 'tsunamite' (Fig. 7). Still higher up, shoreface conglomerates and marine hybrid arenites accumulated (location 4, Fig. 2; Figs. 13a, 14; cf., Sanders 1998).
Significance of in situ clast fragmentation
Letting aside the proto/cataclasites, which per se are rocks formed by brittle shear, in all of the described deposits there is evidence for in situ clast disintegration (= while embedded in sediment) ranging from crackle to mosaic fabrics, with fracture pores filled with the same primary matrix as seen in the rest of the deposit (Figs. 5c, e, 6e, f, 9e, f, 12). If mere overburden had caused clast fracture, this should be seen in all clast-supported fabrics. Instead, also in conglomerates, overburden induced pressure solution along convexo-concave contacts while the majority of clasts remained unfractured (Fig. 6c, d). The observed clast/matrix fabrics and the deformation bands that cross-cut clasts (Fig. 12), are inconsistent with fracture merely by sediment load. A hallmark of deposits of rapid, catastrophic mass-wasting (CMW = rockslides, rock avalanches; Evans et al. 2006) is pervasive in-situ clast fracture by dynamic disintegration (e.g., Davies and MacSaveney 1999, 2002, 2009; Dufresne et al. 2016; Schilirò et al. 2019; Sanders et al. 2022; Sanders and Gruber 2023). The matrix of CMW deposits is a cataclastic gouge generated by clast fracture and attrition such that a 'self-similar' grain-size distribution results (cf., Perinotto et al. 2015; Dufresne et al. 2016). In the studied Gosau succession, thus, the clasts fractured in situ most probably were produced during seismic events. A characteristic of coseismically fractured coarse-clastic deposits is that fracturation is confined to ensembles of clasts and/or to subvertical belts of clasts, whereas the clasts in the adjacent sediment are intact (cf., Meyer et al. 2006; Tuitz et al. 2010; Kübler et al. 2017; Sanders et al. 2018; Tokarski and Strzelecki 2020). This feature fits with the patchy distribution of clasts fractured in situ observed in the studied Gosau deposits. Furthermore, non-stylolithic planar contacts between clasts (Fig. 11), be they fractured or not, are characteristic of coarse-clastic sediments compacted by coseismic rocking (Sanders et al. 2018). Soft-sediment deformation structures that presumably are produced by early post-depositional coseismic displacement are present also higher up in the succession at site (see Fig. 3 and Supplementary Fig. S3). Because the arenitic sediment affected by coseismic deformation was not fully lithified, these structures lack features of brittle deformation such as cataclasis, slickensides, and Riedel shears (Supplementary Fig. S3).
A few kilometres south of the study area, along the southern margin of the NCA, Ortner and Reiter (1999) had identified Jurassic-Cretaceous, top-SE low-angle detachments that cut from Triassic carbonate rocks down into older (anchi-) metamorphic rocks. Such detachments, and associated secondary faults, produce accomodation space for wedge-top basins and may explain the intervals of cataclasites directly below the base of Gosau successions. Similar cataclasites are locally present along the base of the Gosau at other locations, but so far were not mapped or otherwise studied in detail. Coseismic deformation structures in successions of wedge-top basins seem hardly surprising, yet, in the Gosau Group, it seems that such features to date were nearly totally overgone (Sanders 2001). Together, (i) fragmentites in the rock substrate directly below, (ii) coseismic deformation structures within Gosau deposits, and (iii) paralic event beds that result from coastal inundation may be viewed as a 'trinity' related to the synorogenic tectonic setting.