Where is SEP going? Alumni have been known to comment that SEP looks unfocused. I think it was my good friend Rick Ottolini who said we look like an oil-company research lab from the 1980's. They see me playing with radar images of volcanos, bottom-sounding surveys from deep ocean ridges, and some badly recorded depth soundings of the Sea of Galilee. Is Claerbout oblivious to the industry and the changes in the world? Here is why we are going off in these seemingly irrelevant directions.
Oil men think porosity and permeability are everything. Academics believe they can solve all problems with ``inversion''. We academics are way ahead of the oil men in our high count of failures, both reported and unreported. I have had many more failures than successes. Successes are rare. Likewise, all the people I most respect have reported many failures. Still we persist, despite grave frustrations. ``Inversion'' (I prefer to call it ``estimation'') is a systematic approach to problems. It is not merely a black art. There are theories, claims, counter claims, etc. And many of these claims can be tested. Because of my experience with a high percentage of failures I decided to start young people off with simple problems where the meaning of failure is self evident (almost). I chose topographic analysis. Putting data on a regular mesh. Filling in gaps. Segregating signal and noise. That's why we have Galilee, Vesuvius, Madagasgar, Fernandina, and other nonpetroliferous data everywhere on display around here. We are learning. And the tools we are building carry straight over to billion dollar industrial problems.
The first billion dollar problem I have in mind is insufficiently dense 3-D seismic data acquisition. Spatially aliased multiples stack into primaries in marine data. This problem is getting worse as more streamers are added to multi-streamer acquisitions. Sean Crawley and I are successfully interpolating multi-streamer data by estimating and inverting prediction error filters. Poor sampling lowers resolution and creates imaging artifacts in land data. Nizar Chemingui and Biondo Biondi are generating high-resolution images from under-sampled land data by inverting an imaging operator called Azimuth Moveout. These two applications to a billion dollar seismic problems arose from estimation methodologies and tools developed using non-seismic data!
The past two years have brought me the magical helix. This discovery (brought to us by our topographic toys) has spun off projects in three different areas: (1) revitalizing wave-equation migration in 3-D, (2) preconditioning estimations (big speed up), and (3) regularizing velocity estimation (blending measured with prior information). The amazing thing about the helix (as embodied in Sergey Fomel's F90 library) is that any program that does something nifty with two-dimensional data, immediately with no change does the same nifty thing on three- or four-dimensional data.
SEP got started thanks to wave-equation migration. When 3-D came along, for a while SEP lost its competitive edge in migration, as Kirchhoff methods seemed the only way to go. But now the advantages of wave-equation imaging are luring the industry again. We made important progress in that direction; both in the basic technology for wavefield extrapolation and in methods to migrate 3-D marine data. James Rickett and Sergey Fomel worked with me on implicit solutions of the 3-D one-way wave equation that exploit the power of the helix. Biondo Biondi developed common-azimuth migration that drastically reduces the cost of 3-D wave-equation prestack migration and yields encouraging subsalt images when compared with Kirchhoff images.
The dream of estimating velocity by wave-equation methods is getting closer thanks to the recent progress in migration-velocity analysis achieved by Biondo Biondi in collaboration with Paul Sava. Bob Clapp cleared the path by sharpening our tools to handle the nonlinear and undetermined nature of velocity estimation. My title, ``Everything depends on V(x,y,z)'' says the world is three dimensional. This is where SEP is making its mark, uniquely among university consortia. This is our path.