Seismic data from the Canadian Foothills are often noisy and sparsely sampled with missing traces. We show that reverse-time depth migration (RTM) provides a robust means of imaging sparse noisy data. Traditional RTM computes finite-difference (FD) solutions to the wave equation while back propagating the reflected wavefield to the reflector location at depth. The FD stencil will provide an effective means of interpolating missing traces during the migration process since this implicit interpolation essentially allows the seismic wavefield to heal itself during propagation. The FD operator also provides an effective means of suppressing random noise. Following earlier work by Zhu and Lines (1996), we illustrate the effectiveness of RTM on model data and real data from the Canadian Foothills.
View full article as PDF (3.00 Mb)