1. Lee M. Liberty Research Professor Boise State University

2.  Process marine dataset using a standard processing approach segyread tape= | segyclean | supsimage perc=90

3.  Preprocessing  Clean up Shot Records  Amplitude recovery  Deconvolution  Sort to CMP  Velocity Analysis – iterative  Residual statics  NMO correction  Mutes  Stack (gains and filters often follow)  Migrate  Convert to depth

4. Delay-Time GRM (“generalised reciprocal method”) DRM (“diminishing residual matrices”) Methods to determine the residual statics: Topographic Correction (elevation statics) “Uphole”-correction using shots in borehole Refraction statics: corrections for weathered layer

5. CMP Gathers with field statics applied 1. Velocity analysis NMO-Correction Compute Residual statics NMO-Correction with new Velocities Stacking Apply Residual statics 2. Velocity analysis

6. Yilmaz, 1987 188 297

7.  Preprocessing  Clean up Shot Records  Amplitude recovery  Deconvolution  Sort to CMP  Velocity Analysis – iterative  Residual statics  NMO correction  Mutes  Stack (gains and filters often follow)  Migrate  Convert to depth

8. Yilmaz, 1987

9. Mean stack Weighted Stack Diversity Stack/Min-Max-exclude Certain traces are muted and not included in the stacking procedure

11.  Preprocessing  Clean up Shot Records  Amplitude recovery  Deconvolution  Sort to CMP  Migrate?  Velocity Analysis – iterative  Residual statics  NMO correction  Mutes  Stack (gains and filters often follow)  Migrate  Convert to depth

12.  The process by which seismic events (changes in energy) are geometrically re-located in either space or time to the location the event occurred in the subsurface rather than the location that it was recorded at the surface, thereby creating an accurate image of the subsurface.

13. -Migration is dependent on seismic velocity -seismic profile (or gather) length must include the reflected energy

14. Point source in subsurface Diffraction on zero offset seismic section with velocity of overlying medium t 2 =  0 2 +4x 2 /V 2 RMS

15. Migration actual Seismic section (unmigrated) Migration (move it back) sin(  ) = tan (  ), where  = true dip,  = apparent dip

17.  Migration steepens reflectors  Migration shortens reflectors  Migration moves reflectors updip

20. Models (perfectly migrated) Synthetic Seismic Sections (Unmigrated)

21. Schematic that shows the imaging problem for a vertical fault.

22. In an un-migrated time section reflectors do not represent the true subsurface geometry. See examples below… From Kearey et al., 2002 Seafloor Time section Geological Cross- section Dipping reflectors Bow-tie effect (A) (B) (C) (A) a syncline on the seafloor is imaged as a “bow-time section (B) The addition of diffractions from the end of reflectors results in a very complex time section (B) A dipping reflector is shallower in a time section

23.  Unmigrated image with bowtie reflectors  Migrated image

24.  Original image  Unmigrated image

25.  Wave-equation migration (computationally expensive ◦ Spread every data point along the wavefront ◦ (Huygens Principal) ◦ Or downward continue the wavefront  (usually finite Finite-difference)  the most exact pursuit of migration – assuming the use the ‘exact’ subsurface velocity field  Kirchoff migration (intermediate speed) ◦ Sum diffraction curves  F-k (Stolt) migration (fastest) ◦ Remapping in frequency-wavenumber domain ◦ assume constant velocity (or simple gradient)

26. Shot records Prestack migrated Depth section Velan; Prestack Depth migration Time migrated section Stacked section Time migrated Depth section Velan NMO stack Time migration Time-depth conversion Dip moveout shots Velan NMO DMO Inverse NMO Velan NMO stack DMO stack Prestack Migrated section Stacked section Velan NMO stack Depth migration Depth migrated section Migration; Time-depth conversion

30.  Migration after stack: ◦ Each trace on a stacked section is a “zero-offset” trace (coincident source and receiver) ◦ Migration is done by spreading each data point in a nearly spherical wavefront ◦ Midpoint “smearing” on dipping events reduces resolution ◦ Assumes no significant lateral velocity gradient  Migration before stack: ◦ Each trace has a different source and receiver location ◦ Migration is done by spreading each data point in an eliptical wavefront ◦ Must migrate “CMP fold” times as many traces ◦ Can give a “perfect” migration

32.  Also known as “partial prestack migration”  Cheap alternative to true prestack migration  Breaks prestack migration into 2 steps: [CMP sort, velocity analysis, NMO] ◦ Partial migration (DMO) on shot records [remove NMO, CMP sort, velocity analysis, NMO, stack] ◦ Migration of stacked section

34. General discussion of the term ‘Dip Moveout’ (DMO) t 2 =  0 2 +4x 2 /V 2 RMS

35. DMO is a seismic processing operation to correct for the fact that, for dipping reflections, the traces of a CMP gather do not involve a common reflecting point – e.g., line 1.

36. velocity analysis – fast velocities

37. velocity analysis – slow velocities

38. DMO corrected velocity analysis

40. Sum Diffractions (Kirchoff)

41. Diffraction Summation (Diffraction Stack) Wavefront Summation (Kirchoff Stack) unmigrated migrated

42.  Migration in the F-K domain  the quickest migration method  assume constant velocity (vertical velocity can be accommodated)  accurate to 90 o  works by transforming the data to the FK domain, applying the migration equation tan(a )=sin(b ) which shifts the data vertically on the frequency axis only to transform the data to the migrated dip.  Following inverse transformation the migrated solution is obtained.

43. 35 traces 70 traces 150 traces 256 traces

44. True V V 5% low V 10% low V 20% low

45. Random Noise + Migration Migration “smiles” esp at edges of data Unmigrated Migrated

46. Note: Steep dip migration