A predominantly chaotic interval, interpreted as a well-preserved MTD, is present across the eastern portion of the dataset (Fig. 1; Table 1), the base of which sits at ~400 m subsurface. The deposit is continuous along-channel for 68 km, covers >340 km2 and has a preserved volume of ~19.3 km3 (Fig. 1E). The MTD is thickest around the second major meander bend, tapering both up- and downstream (Fig. 1E).
The MTD spans the ~2.5 km channel width, and extends onto channel overbank areas on both sides of the channel for up to 5 km (Fig. 2A-B). In overbank areas corresponding to the external levees (sensu Tek et al., 2021), the MTD overlies: (a) Flat, <10 km2 platforms dipping in accordance with the underlying overbank stratigraphy, typically located along peaks (black reflectors); they are interpreted as mechanically weak layers (Fig. 2D). (b) Steep (up to ~80°) steps that link adjacent flat platforms, against which adjacent overbank reflectors terminate; in plan view, most are gently curved, up to 5.7 km long, and are oriented subparallel to the channel (Fig. 1C, 2D); some shorter (4 km long), relatively straight, channel-subperpendicular steps join laterally offset channel-subparallel stretches (Fig. 1C, 2D). Where the MTD overlies terrace deposits (sensu Tek et al., 2021), its base conforms to the terrace-bounding surfaces, i.e., dipping up to ~20° toward the channel (Fig. 2B).
The MTD is imaged as chaotic to transparent packages, folded and faulted reflectors, and as blocks of coherent reflectors (Fig. 2A-B). The dominant MTD seismic facies comprises chaotic reflectors with little internal reflectivity (Fig. 2A-B & E). Undeformed or weakly-deformed blocks of reflectors separated from the overbank reflectors by chaotic packages but with similar seismic character as the adjacent in-situ overbank are interpreted as megaclasts (Fig. 2). Two of these are adjacent to the left bank of the channel (Fig. 2A, C-E), with a third adjacent to the right bank further downstream (Fig. 2B). They occur within 1 km of scallop-shaped planform indentations in the channel-wall, interpreted as headwall scarps in the MTD source area (Fig. 2).
The megaclasts are encased in chaotic material, can be up to 120 m thick, 4.1 km long, and have surface areas up to 3.9 km2 (Table 1, Fig. 2C); they exhibit similar dip and reflector sequences as the nearby in-situ overbank (Fig. 2A & D). In plan-view their long edges are sub-parallel to the strike of steps in the failure surface, and their corners match to kinks in these steps (Fig. 2D), suggesting that the blocks were locally sourced by retrogressive failure into the channel-wall and slid along decollement surfaces parallel to the orientations of overbank strata. Between and on top of the rotated blocks, wedges of chaotic material are thickest where the top of the megaclasts meet the MTD base and thin towards and away from the channel. These are interpreted as slump and debris-flow material sourced from shallower overbank stratigraphy that filled depressions around and on top of the megaclasts (Fig. 2).
For most of its length, the MTD overlies a ~40 m thick, up to 2.5 km wide sequence, imaged as high amplitude reflectors (HARs) nested in a concave-up surface, interpreted as sand-rich deposits of the pre-existing channelform8 (Fig. 2B). However, these deposits are absent beneath the thickest parts of the MTD (Fig. 2A, 3). Instead, sigmoidally-stacked arrays of HARs are seen laterally, encased by chaotic deposits. The HAR stacks are up to 40 m thick, 1.5–7 km2 in area and dip up to 50° toward the thickest part of the MTD (Fig. 2A-B, 3B-C). These HARs are interpreted as imbricated thrusts of channel-floor deposits, ripped up from their original locations and forced onto the channel margin opposite to the failure source (Fig. 4). These occur in two distinct clusters: (1) adjacent to the maximum thickness of the MTD, where the thrust geometries suggest a transport direction to the SE toward the oceanward channel margin (Fig. 3A-B), and (2) adjacent to the downstream landward channel margin, where thrusts suggest NE transport (Fig. 3A & C). At the upstream end of cluster 1, there is a 700 m long stretch where no HARs are present under or within the MTD (Fig. 3A). Here, the underlying HARs appear to have been ripped up and transported down-channel.
The erosional nature of the MTD indicates that the volume of displaced materials exceeded that of the initial failure. The entire MTD top is incised by a post-failure channelform locally scouring to the MTD base (Fig. 1E-F), indicating that the original size of the MTD exceeded the preserved volume. Loci of subsequent erosion corresponds to topographic lows on the original MTD surface (Fig. 5).
Table 1. Volumes and areas of the MTD and three largest megaclasts. See supplementary material 1 for information on other channel-wall collapse MTDs.
|
Maximum thickness (m)
|
Median thickness (m)
|
Maximum long axis length (m)
|
Area (m2)
|
Volume (m3)
|
Maximum MTD length or long axis length for megaclasts (km)
|
Area (km2)
|
Volume (km3)
|
Wall-collapse MTD
|
265
|
50
|
|
340828165
|
19267539760
|
68
|
340.828165
|
19.26753976
|
Megaclast 1
|
141
|
120
|
2400
|
1933116
|
169916285
|
2.4
|
1.933116
|
0.169916285
|
Megaclast 2
|
136
|
115
|
4080
|
3902347
|
341396797
|
4.08
|
3.902347
|
0.341396797
|
Megaclast 3
|
130
|
105
|
2760
|
2767878
|
257922000
|
2.76
|
2.767878
|
0.257922000
|