1. Seismic Imaging in GLOBE Claritas

2. VELSECT : Automatic Velocities
High Density Velocity Analysis
Picks made at every CDP, approx every 60ms
Optionally constrained by top and base horizons
Automatic Analysis
Picked on high fidelity semblance spectra
Spectra optimised in pre-processing
Runs efficiently in parallel
Editing and Smoothing
Geological constraints used for edits
Statistically robust (300,000+ VT pairs)

3. VELSECT : Stacking Results
Manual Velocities : 2km spacing
VELSECT Velocities
Very similar results – but improved (eg under channels)
3000km of data picked over a weekend, automatically
Fast, accurate, repeatable and reliable

4. VELSECT : Interval Velocity Results
Manual Velocities, Dix Inverted
VELSECT Velocities, Dix Inverted
VELSECT results show more detail, better resolution
High velocity limestones resolved to two bands
Tied into wells and show good match with sonic-derived functions
Resolve coals from carbonates through velocity profile
Identify possible overpressure zones
Excellent for curved-angle calculations : AVA and Imaging
VELSECT : Automatically create VELocity SECTions

5. Solution : Imaging Under Channel, original
Original post stack migration
The channel creates a low velocity zone with steeply dipping sides that defocuses seismic energy.
Imaging is severely disrupted under the channels.
Channels are all across the prospect area.

6. Solution : Imaging Under Channel, PreSTM
Pre-stack Time Migration using VELSECT velocities
Solution is improved, but not complete.
Ray path bending is not fully accounted for by the preSTM alone, and additional imaging is needed

7. Solution : Imaging Under Channel, PreSDM
Pre-stack Depth Migration
Complete solution.
Modelled channel and near surface velocities successfully correct for the ray-path bending at the sea-floor, as well as the bright limestone event (approx 1500ms)
Channel shape is unchanged in all cases – but the velocity variation is correctly modelled.

8. Solution : Imaging Under Channel, Velocity Model
VELSECT technique employed after preSTM
VRMS values converted to interval velocities
test lateral smoothing and use to depth convert
use smoothing which produces simplest depth image
secondary smoothing in depth
Run PreSDM as second imaging phase
Interpretation free preSDM modelling methodology

9. TRV-434 : structure West East
Shooting Direction
27.2km Total Length
3.2km Cable Length
Overthrust
Tikorangi Limestone
Approximate Depth in metres
Schematic of TRV-434 taken from previous depth imaging study
Note the location and depth of the overthrust relative to the cable length

10. TRV-434 : Original Time Processing
Imaging using a conventional late-1980’s sequence, with DMO and post-stack time migration.
Sub-thrust imaging is poor; shot-receiver ray paths are complex and the simple DMO-Stack-Migration approach cannot resolve the structure. Sub-thrust imaging is confused, with broken, crossing events (circle)

11. TRV-434 : Modern Time Processing
Imaging using a modern sequence that addresses spatial aliasing and employs two passes of pre-stack time migration
Overall image is much cleaner, and imaging has improved considerably. A layered structure starts to appear, but is still smeared (circle). Pre-stack time migration still assumes the shot-receiver ray-path is symmetrical about the trace midpoint, however.

12. VELSECT : Raw Velocity Results
RAW VELSECT velocities
trends can be seen
data is still noisy
cannot be used for stacking

13. VELSECT : Edited Velocities
Edited VELSECT velocities
around 60% edited out
still 100,000+ picks
interval velocity editing
iterative approach

14. VELSECT : Smoothed Velocities
Final VELSECT velocities
spatial frequency filter
extract low pass component
1-2km radius filter
10% spatial nyquist limit

15. VELSECT : Interval Velocities
VELSECT Interval velocities (left) and preSTM data (right). The velocity field shows structure that matches the seismic image, and geological expectations

16. TRV-434 : ADMIRE Depth Imaging
GNS Science’s Automatic Depth Modelling Iteration via RMS velocity Estimation (ADMIRE) approach creates a grid-based depth model that is ray-traced to produce an image
Imaging is considerably improved with asymmetric ray-paths being managed correctly. Layer structure beneath the overthrust is now imaged sufficiently to resolve faulting, enabling detailed interpretation and analysis.

17. TRV-434 : Layer Based Imaging Vs ADMIRE
Layer based pre-stack depth migration approaches use a layered earth model
ASSUMES : layered earth represents the velocity structure accurately
REQUIRES : detailed structural interpretation of each layer with each iteration
- Time consuming, expensive and can result in model-driven solutions
ADMIRE pre-stack depth migration uses a grid based model, created from the data
ALLOWS : velocities to be independent of structure, and extremely complex
REQUIRES : no structural model or interpretation, just careful quality control
- Computer intensive, automatic and data driven
1480m/s
3750m/s
6000m/s
Layer-Based Model
ADMIRE Model

18. TRV-434 : Layer Based Imaging Vs ADMIRE
Conventional Layer-Based Imaging
ADMIRE Grid-Based Imaging
Even after a large (13+) number of layer-based model updates the conventional depth imaging approach lacks the clarity and resolution of the ADMIRE image (with 5 model updates)
Where seismic velocities are independent of sub-surface structure the ADMIRE approach produces a more accurate image, with less iterations, and no interpretation