Auxiliary Material for Paper 2012GC004454 Three-dimensional seismic structure of a Mid-Atlantic Ridge segment characterized by active detachment faulting (Trans-Atlantic Geotraverse, 25{degree sign}55'N-26{degree sign}20'N) Minghui Zhao Key Laboratory of Marginal Sea Geology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou, China Juan Pablo Canales and Robert A. Sohn Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA Zhao, M., J. P. Canales, and R. A. Sohn (2012), Three-dimensional seismic structure of a Mid-Atlantic Ridge segment characterized by active detachment faulting (Trans-Atlantic Geotraverse, 25{degree sign}55'N- 26{degree sign}20'N),Geochem. Geophys. Geosyst., 13, Q0AG13, doi:10.1029/2012GC004454. Introduction Auxiliary material for this article contains a Microsoft Word version of the readme file and three figures that are described below. 1. 2012gc004454-txts01.doc Text S1. Microsoft Word version of the readme file that contains the following information. S1. Random Initial Models The 100 random 1D initial models were built by randomizing five parameters defining a two-layer media: P-wave velocities at the top of the upper layer (seafloor), at the interface between both layers, and at the bottom of the lower layer, and the thicknesses of both layers. The five parameters were allowed to vary randomly within certain predefined limits to avoid constructing unrealistic models (Auxiliary Figure S1). The velocities at the seafloor, middle interface, and at the bottom of the model were limited to 2?4 km/s, 5.3?7.3 km/s and 7.3?8.5 km/s, respectively. The thicknesses of both layers were limited to be within 1?3 km and 7?9 km for the upper and lower layers, respectively. For each Monte Carlo ensemble a 3D starting model was constructed by draping the randomized 1D velocity model from the seafloor bathymetry. S2. Model Uncertainty and Resolution Tests We define statistical model uncertainty (Auxiliary Figure S2) as the standard deviation of the 100 solutions. The central part of the preferred 3D model up to a depth of 4 km bsf (~8 km below sea level, bsl) is well constrained; velocities there have standard deviation <0.2 km/s (Auxiliary Figure S2), smaller than the variance of the initial models used in the Monte Carlo approach (Auxiliary Figure S1b). As expected, the edges of the model have large uncertainties (>0.3 km/s) because of the lack of ray coverage in these areas (Auxiliary Figure S2). Uncertainties in the well-constrained parts of the model are generally <0.2 km/s, although the central part of the final model between 6 to 8 km bsl has somewhat larger uncertainties (0.2?0.3 km/s, Auxiliary Figure S2) due to a decreased number of crossing rays. We estimate model resolution by conducting checkerboard tests, which provide information on the lateral and vertical resolution of a velocity model by examining the ability to recover a synthetic velocity anomaly pattern superimposed on a background velocity model [Zelt, 1998]. We built a perturbed velocity model by superimposing a 3D sinusoidal velocity anomaly on a 1D model (Auxiliary Figure S1) with the magnitude of 10% background velocity. Synthetic travel times were generated by tracing rays through the checkerboard model and then adding Gaussian random noise with a standard deviation of 50 ms. The synthetic travel times were then used in an inversion using the same parameters used in the true data inversion. We tested two checkerboard patterns with sinusoidal functions with wavelengths of 20 and 10 km in the horizontal directions (both with vertical wavelength of 4 km). Thus these checkerboard patterns allow us to assess the inversion?s ability to resolve velocity anomalies of sizes 10?10?2 km and 5?5?2 km, respectively (Auxiliary Figure S3). The tests show that our inversion cannot constrain velocity variations at depths greater than ~4 km bsf (8 km, bsl); thus we limit our interpretation of the velocity model to structure above that level. Above a depth of 4 km bsf, features of 10 km in size are well resolved along most of the central part of segment (Auxiliary Figure S3a, S3e, S3f), but along the eastern and western sides of the model they can only be resolved in the upper 2 km (<6 km bsl) and from -15?Y?5 km (Auxiliary Figure S3b, S3e, S3f). The inversion can resolve features as small as 5 km in size at depths shallower than 2 km (<6 km bsl) along the central part of segment (-5?X?5 km and -25?Y?25 km, Auxiliary Figure S3c, S3d, S3g). At depths between 2?4 km bsf (6?8 km bsl), however, the inversion can only resolve these smaller anomalies between -5?X?5 km and -15?Y?15 km, Auxiliary Figure S3c, S3d, S3h). 2. 2012gc004454-fs01.jpg Auxiliary Figure S1. 100 randomized starting models. (a) 100 Monte Carlo ensembles (blue lines) together with ensemble mean (red line). (b) Corresponding standard deviation for the 100 starting models. 3. 2012gc004454-fs02.pdf Auxiliary Figure S2. P-wave velocity uncertainties of the final model shown at cross-sections through X=0 km and Y=0 km, and at depth slices at Z=0.5, 1.0, 2.0, and 4.0 km bsf. Contours are drawn at 0.1 km/s interval. 4. 2012gc004454-fs03.pdf Auxiliary Figure S3. Checker board tests. Results from two tests are shown: anomalies with wavelengths of:(a, b, e, f) 20 km (lateral)?20 km (lateral)?4 km (vertical);(c, d, g, h) with 10 km?10 km ?4 km. Recovered anomalies are shown for (a, c) X=0 km, (b, d) Y=0 km, (e, g) Z=5, and (f, h) Z=7 km. Horizontal sections for the case of smaller anomalies in panels g and h show only the central part of the model with hypocenters from the catalog of deMartin et al. [2007] (dots). Grid lines and yellow +/- symbols indicate the sinusoidal synthetic velocity anomaly patterns, which are up to 10% in magnitude relative to background velocity (mean of the 100 initial ensembles, Auxiliary Figure S1). References deMartin, B. J., R. Reves-Sohn, J. P. Canales, and S. E. Humphris (2007), Kinematics and geometry of active detachment faulting beneath the Trans- Atlantic Geotraverse (TAG) hydrothermal field on the Mid-Atlantic Ridge, Geology, 35(8), 711-714.