Auxiliary material for Paper 2010GL044412 Small-scale structure of the Kane oceanic core complex, Mid-Atlantic Ridge 23 deg 30'N, from waveform tomography of multichannel seismic data J. Pablo Canales Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA Canales, J. P. (2010), Small-scale structure of the Kane oceanic core complex, Mid-Atlantic Ridge 23 deg 30'N, from waveform tomography of multichannel seismic data, Geophys. Res. Lett., 37, L21305, doi:10.1029/2010GL044412. Introduction This auxiliary material presents additional comparisons between observed predicted data (Figure S1) and tests of how particular features of the models are constrained by the data (Figures S2 and S3). 1. 2010gl044412-fs01a.pdf, 2010gl044412-fs01b.pdf Figure S1. Observed (left) and synthetics seismograms predicted by the TT model (middle) and the final WT model (right) for selected shot gathers of profile K1. Vertical axes are two-way traveltime with velocity reduction of 5 km/s. Blue arrowheads point to the seafloor reflection. Yellow marks in the observed and WT shot gathers highlight secondary arrivals that are well predicted by the WT models and present in the data, but that were absent in the TT seismograms. These highlighted arrivals show the clear improvement in the match between observed and WT seismograms with respect to the TT seismograms, providing confidence on the WT results. 2. 2010gl044412-fs02.pdf Figure S2. Top panels show the preferred WT model for profile K1 (left), as in Figure 2b, and a modified version (right) in which the low velocity channel enclosed in the black polygon has been removed. This modification of velocities inside of the black rectangle was done by linear interpolation between velocity values at the top and the bottom of the rectangle. Outside of the black polygon both models are identical. To illustrate which part of the dataset requires the low velocity channel, we compare data from one shot gather shot # 4689 with the corresponding synthetic shot gathers predicted by the preferred and the modified WT models (bottom panels). Yellow star and bar plotted in the preferred WT model show the location of the shot and receivers, respectively, shown in the bottom panels (note that they are not plotted at their true depth, which is at sea level). Blue arrowheads in bottom panels point to the seafloor reflection. The presence of the low velocity channel is detected in the data by the abrupt amplitude change observed along the first refracted arrival near 5,000 m offset. This arrival is absent or very weak (indistinguishable from the background noise) in the observed data for offsets larger than 5,000 m, as well as in the shot gather predicted by the preferred WT model. In contrast this arrival, although still weak, can be clearly seen in the gather predicted by the modified WT model. This exercise demonstrates that the low velocity channel obtained from the inversion is required by the data and is not an artifact of the procedure. 3. 2010gl044412-fs03.pdf Figure S3. Top panels show the preferred WT model for profile K1 (left), as in Figure 2b, and a modified version (right) in which the high velocity lens enclosed in the black rectangle has been removed. This modification of velocities inside of the black polygon was done by linear interpolation between velocity values at the top and the bottom of the polygon. Outside of the black rectangle both models are identical. To illustrate which part of the dataset requires the high velocity lens, we compare data from one shot gather shot # 4781 with the corresponding synthetic shot gathers predicted by the preferred and the modified WT models (bottom panels). Yellow star and bar plotted in the preferred WT model show the location of the shot and receivers, respectively, shown in the bottom panels (note that they are not plotted at their true depth, which is at sea level). Blue arrowheads in bottom panels point to the seafloor reflection. Yellow arrowheads point to arrivals observed in the data that are well predicted by the preferred WT model, but that are not predicted by the modified WT model. Red arrowheads point to arrivals predicted by the modified WT model but that are absent in the observed and the synthetic shot gather of the preferred WT model. This exercise demonstrates that the high velocity lens obtained from the inversion is required by the data and is not an artifact of the procedure.