Joint least-squares inversion of up- and down-going signal for ocean bottom data sets

Joint Inversion Result

A joint inversion is performed in a least-squares sense with the objective goal described in equation 1. The initial guess is calculated by summing the up-image and the down-image. Panel (c) of Figure 6 shows the image after joint inversion (with 20 iterations). We can see an overall improvement from the migration images in panels (a) and (b) of Figure 6 to the inversion image. We have identified three areas of improvement with the close-up section shown in Figure 7:

1. In panel (c) of Figure 6, the near-surface reflector near z=0-400 m and x=2400-3200 m has a better relative amplitude and is more focused.

2. In panel (c) of Figure 6, the joint image has a wider illumination for the region from x=5000 m to the left and the region from x=2000 m to the right.

3. Figure 7 shows a higher signal-to-noise ratio for the joint image. In addition, the deeper reflector at z=1200-1500 m and x=3200-4000 m are better illuminated in the joint image than in the up-image or the down-image.

This example shows that joint inversion coherently combines information from primary and mirror signal to produce a better illuminated and resolved image. In the next section, we will explore the effect of a sparse geometry between shots and receivers.

Joint least-squares inversion of up- and down-going signal for ocean bottom data sets