AN: T72A-1226
TI: Effect of Seafloor Topography on MCS Reflection Images and OBS Locations in the Mariana Subduction Factory Seismic Experiment
AU: * Gunther, R H
EM: rgunther@stanford.edu
AF: Stanford University, Department of Geophysics, Stanford, CA 94305 United States
AU: Klemperer, S L
AF: Stanford University, Department of Geophysics, Stanford, CA 94305 United States
AU: Goodliffe, A M
AF: University of Hawaii, SOEST, Honolulu, HI 96822 United States
AU: Kerr, B C
AF: Stanford University, Department of Geophysics, Stanford, CA 94305 United States
AB: During a wide-angle seismic experiment over the Mariana arc we deployed 53 OBSs along three lines parallel to the arc, with spacings of 10 to 20 km. During airgun shooting to the OBSs with 200 or 250 m shot spacing, we acquired low-fold MCS data aboard the R/V Ewing, both in-lines directly along the lines of OBSs, and cross-lines perpendicular to the arc. To prevent misinterpretation of sideswipe from rough seafloor topography as real geology, we used bathymetry data to distinguish between seafloor and subsurface reflectors. The Hydrosweep multibeam sonar measured traveltimes to the seafloor on lines perpendicular to the ship at c. 25 m intervals. Points on the seafloor with traveltime minima or maxima were identified as reflectors. The predicted reflections were compared to reflections in the MCS data. We found sideswipe occurring up to 400 ms before the vertical seafloor reflection. By using the bathymetry data we can determine the origin of first arrivals despite having only 2D seismic data. On a profile passing beside active volcanoes, we found cases where the main seafloor reflection arrives at 35 degrees from the vertical. Swift currents caused the OBSs to drift significantly on their way to the seafloor. Most OBSs were retrieved between 200 and 600 m from where they were dropped. Instrument locations uncertain by this amount lead to c. 100 ms uncertainty in modeling seismic arrivals, so we calculated instrument locations from direct water-wave arrivals on the OBS records. The in-line position components were constrained with an accuracy of 5 to 25 m, or c. 5% of the shot spacing for most stations. Cross-line position components are also required for the data to be inverted in a 3D model. Success in determining off-line positions depended on the distance of the nearest cross-line from the OBS, and on the seafloor conditions. Over the deep, flat seafloor of the fore-arc, crosslines between 10 and 20 km distant led to off-line location accuracies of 20 to 50 m. Along the volcanic arc, crosslines more than 10 km distant generally did not provide data useful for ranging because extreme topographic variations distorted the first-arrivals.