Tape 21 results: 

Velocity Analysis: 
The filtered data for Tape 21 was sorted into CMP gathers in order to perform a velocity analysis (VELDEF). The general topography of the preliminary stack was flat (Fig. \ref{577647}).  At around 10 s we can see a reflection multiple at twice the time interval in which the primaries arrived. We decided to mute the data below 9 s so that the multiples wouldn't interfere with the velocity analysis and migration results. We picked out the moveout velocities in an interval of 50 CMP increments (700, 750, 800, 850...) because a larger interval would yield an under-migrated image. Fig \ref{430222} shows an example of the velocity picking for CMP 2050. The first pick on the top is the seafloor reflection. Velocities tends to increase as the depth increases, so the rest of the picks are selected based on this principle. The NMO is turned on to observe how flat the reflections look after our picks, and then we correct accordingly to flatten the ones that were not. After the velocities were picked for every CMP, the result was the final velocity model shown in Fig \ref{680582}. This model was used for the NMO correction and the Kirchhoff migration (MIGTX). Other models were used before with a larger interval of velocity picks, but the stacked image had unclear boundaries between traces, indicating that not all CMP gathers had been flattened after the NMO correction (Fig \ref{164699}).