Seismic diffraction imaging to better characterise mass-transport complexes: examples from the Gulf of Cadiz, south west Iberian Margin

15 Mass-transport complexes are characterised by complex, laterally discontinuous inter16 nal structure, such as pressure ridges, local shear zones and intact translated blocks. Their 17 internal structure is often poorly imaged by seismic reflection techniques, which are fun18 damentally limited in lateral resolution by the bandwidth of the seismic source. Diffrac19 tion imaging, instead, directly images subsurface heterogeneity by primarily imaging the 20 diffracted part of the seismic wavefield. We apply seismic diffraction imaging to two ma21 rine multi-channel seismic profiles containing mass-transport complexes from the Gulf 22 of Cadiz (south west Iberian Margin). We observe that mass-transport complexes gen23 erate a large amount of diffracted energy relative to the un-failed sediments. We demon24 strate that, in combination with conventional seismic reflection images, diffraction im25 ages can be used to better discriminate the lateral extent (runout) of mass-transport com26 plexes, particularly for thin bodies that are not well-resolved using conventional imag27 ing. We suggest that diffraction imaging may have similar applications for marine geo28 hazard assessment to seismic discontinuity attributes, such as the similarity attribute, 29 with the advantage of being closer to a true image of the heterogeneous subsurface. Ap30 plying diffraction imaging to image mass-transport complexes from 2-D seismic data is 31 challenging, but may provide some unique insights that are not available from conven32 tional reflection images. 33

In recent years there has been an increased focus on using seismic reflection data to image mass-transport complexes to better understand the parameters that control their  The seismic acquisition and processing flow were designed to maximise the tem-163 poral and spatial resolution of the resulting seismic images. The shot interval was cho-164 sen to ensure a nominal coverage of at least 12-fold with a midpoint interval of 3.125 m. 165 A relatively small seismic source (an airgun array with total volume 930 cu. in.) was used 166 to maximise the dominant source frequency. The source array and streamer were towed 167 at a relatively shallow depth (approximately 3 m) to ensure that the frequency of the first 168 source and receiver ghost notches was as high as possible. Detailed acquisition param-169 eters for the two profiles are given in Table 1. Broadband pre-processing was performed 170 onboard using RadExPro seismic processing software. Traditional pre-processing focuses 171 on imaging specular reflections, meaning that diffractions are often ignored or removed. 172 Preserving diffractions through the pre-processing flow requires care as diffraction tails 173 are generally lower amplitude, higher frequency and dip more steeply compared to re-174 flections. The broadband pre-processing flow consisted of i) swell noise removal (to en-175 hance the signal-to-noise ratio at low frequencies); ii) deghosting (to correct for the source 176 and receiver ghost effect); iii) designature (to transform the data to zero-phase and re-177 move the bubble pulse, boosting the low frequency content) and iv) shot domain τ − 178 p muting (to remove steeply dipping noise). For most of the survey area, the signal pen-179 etration depth was similar to, or less than, the two-way travel time (TWTT) of the first 180 waterbottom multiple, therefore no multiple attenuation was performed. Instead, a bottom-181 mute was applied from above the first waterbottom multiple before imaging to prevent 182 -7-manuscript submitted to JGR: Solid Earth   de-migrated using the migration velocities (Appendix A), giving a local dip field that 210 approximates the unmigrated dip.

212
For diffraction separation we treat the recorded seismic wavefield as being composed 213 of i) specular reflections, ii) diffracted energy and iii) noise (including other seismic ar-214 rivals). If the noise is small, we can retrieve the diffracted wavefield by eliminating the 215 specular reflections. We perform the separation using a dip-guided PWD approach on 216 common-offset gathers (Claerbout, 1992;Fomel, 2002;Fomel et al., 2007). This approach 217 assumes that for unmigrated, common-offset seismic data specular reflections are later-218 ally coherent events with continuously varying slope (i.e., smooth). PWD filters can pre-219 dict smooth, laterally continuous energy that is close to the estimated local dip. This 220 approximates the specular reflected wavefield. We subtract this from the pre-processed, 221 unmigrated, common-offset data to eliminate the reflections. The remaining data con-222 tains the diffracted wavefield, noise and some residual reflection energy.   The seismic profiles analysed for this study were acquired using a short streamer 232 (approximately 500 m far offset) compared to the water depth (>1 km), so there is no 233 significant differential moveout of reflection events in common-midpoint domain to per-    281 Assuming that diffractors are evenly distributed throughout the mass-transport com-282 plex, some of the diffraction energy will always come from outside the vertical plane of 283 the profile (i.e., |x| > 0 in Fig. 3). If the body is wider than it is thick, the apparent thickness of the slide from diffraction images will be greater than the apparent thickness 285 of the slide from reflection images. This results in a "shadow" of diffraction energy be-286 low the true basal surface of the mass-transport complex in 2-D seismic data. 287 We can use this "diffraction shadow" to quantify the width, perpendicular to the 288 profile, of the zone of potential diffractors that contribute to the image. For a mass-transport 289 complex exposed at the seafloor we can make the simplifying assumption that potential 290 internal diffractors are at, or near, the seafloor, so v rms ≈ v water . We consider that the 291 seafloor is equivalent to the potential top surface of the mass-transport complex. The 292 seafloor depth is known independently from multi-beam swath bathymetry.

293
The workflow to calculate the zone of diffractors that contribute to the image is   (2).

302
The distance from the profile to the projected base of the "diffraction shadow" tells 303 us the minimum extent of the zone of diffractors from the profile, in the direction of max-304 imum extent. If we assume that the majority of diffraction energy is generated by the   1 Figure 4. Seismic profile INS2-Line1 from the Portimão Bank area (Fig. 1) Fig. 6d shows a stack of the separated diffractions. This section is comparable to 367 the unmigrated stack (Fig. 6b). Diffraction tails are clearly seen throughout the section,  complex exposed at the seafloor (Fig. 4). It shows the conventional seismic reflection im-379 age (Fig. 5a), the corresponding diffraction image (Fig. 5b) and the energy of the diffrac- of mass-transport complexes in the Infante Don Henrique Basin. Fig. 7a shows the con-395 ventional seismic reflection image and Fig. 7c shows the similarity attribute of the con-396 ventional image (hereafter referred to as "conventional products"). Fig. 7b shows the diffrac-397 tion image and Fig. 7d shows the logarithm of the energy attribute of the diffraction im-398 age (hereafter referred to as "diffraction products").

399
Nine mass-transport complexes are interpreted from a combination of the conven-400 tional and diffraction products (labelled in order of decreasing depth from 1 to 9). In-401 terpretation of a mass-transport complex is guided by one or more of the following fea- Potential  Fig. 5b supports an interpretation of MTC A as 465 a "structured" rather than "structureless" deposit, even if the morphology of such struc-466 ture is not well-resolved by seismic methods. 467 We also clearly resolve a normal fault plane below MTC A in the diffraction im- for seismic profile MP06b) compared to the conventional seismic image (Fig. 7f). The 490 diffraction image, and corresponding energy attribute, clearly highlights these events. 491 We also observe this effect on seismic profile INS2-Line1 (Fig. 5). Here, there is a  The record of buried mass-transport complexes identified from marine geophysi-513 cal data is biased toward events that can be clearly resolved in multi-channel seismic re-514 flection images (i.e., relatively thick and laterally extensive). This means that catalogues 515 of mass-transport complexes are biased towards larger events, or younger events that are 516 -25-manuscript submitted to JGR: Solid Earth still preserved in the bathymetry (e.g., Urgeles & Camerlenghi, 2013). Screening for mass-517 transport complexes using diffraction imaging will allow for a more complete catalogue 518 of smaller, deeper events. to the energy attribute of the diffraction image (Fig. 7). There are high-level similari-525 ties between the two: relatively large events (MTC3, MTC4 and MTC8) are clearly im-526 aged by both methods. Many smaller events, however, are not clearly delineated from 527 the background geology by the similarity attribute. Moreover, the similarity attribute 528 seems to be sensitive to features other than geological heterogeneity -we see noise from 529 high-amplitude laterally continuous horizons at a similar amplitude to the genuinely dis-530 ordered geology of the mass-transport complexes. It is difficult to discriminate a high-531 amplitude, horizontal un-failed horizon from a thin mass-transport complex using the 532 similarity attribute. 533 We argue that when screening for mass-transport complexes, diffraction images and 534 derived attributes may be more useful than discontinuity attributes of reflection images, 535 because they are more sensitive to the target (i.e., heterogeneous geology) and they con-536 tain lower "noise" generated by high-amplitude, coherent reflections. The diffraction im-537 age suffers less from interference from high-amplitude reflections, or edge effects and smooth-538 ing that may be introduced by window-based attributes. Diffraction images and derived 539 attributes are a more "physically correct" alternative to conventional attributes because 540 diffraction images directly image subsurface heterogeneity (i.e., fine scale disordered ge-541 ology) at the scale of the seismic wavelength or below. tors associated with a strongly heterogeneous body. In other words, it offers a constraint 548 on the minimum width of a mass-transport complex imaged by a 2-D seismic profile. 549 We demonstrate the method by applying it to INS2-Line1, where there is a well 550 defined "diffraction shadow" beneath MTC A (Fig. 8b). The presence of such diffrac- to be approximately horizontal, for most mass-transport complexes) and assumes that 560 the body is thin compared to the water depth. It also relies on being able to separate 561 diffractions generated by the body (the "diffraction shadow") from diffractions gener-562 ated by the background geology surrounding the body, which may not always be straight-563 forward.

564
The method is simple but nevertheless could be a useful way to estimate a lower 565 bound on the extent of mass-transport complexes from a single 2-D seismic profile, where 566 other geophysical information is not available. This is a common scenario when screen-567 ing for mass-transport complexes for marine geohazard studies in frontier areas; for aca-568 demic and vintage datasets; and in polar areas, where acquiring 3-D towed-streamer seis-569 mic data may be impossible due to year-round ice cover. It is trivial to extend the method 570 to deal with buried mass-transport complexes, so long as i) the velocity model to the top 571 of the body is known; ii) the slide is thin relative to its depth; and iii) the topography 572 of the top surface is small, relative to its depth. Whilst we have shown that diffraction images clearly offer better imaging of het-576 erogeneous geology compared to reflection images, there remain some limitations, par-577 -27-manuscript submitted to JGR: Solid Earth ticularly regarding the data used for this study and the specific application to charac-578 terise mass-transport complexes.
We first solve for the un-migrated local dip value, α(x , t ). Then we calculate the hor-677 izontal and vertical (time) shift (x −x and t −t). The de-migrated dip field α(x, t) is 678 estimated by applying image warping (with the horizontal and vertical shifts) to α(x , t ).

679
The effect is to reverse the effect of migration on the dip field, to "de-migrate" the dip