Processing Pitfalls

Module 14
Processing Pitfalls and Quality Assurance
Learning Objectives

Awareness: Pitfalls associated with processing problems Quality control steps in seismic processing

Processing Pitfalls

Topics

Pitfalls of seismic exploration

Processing Pitfalls Interpreting Pitfalls

Seismic Processing Quality Assurance

Geometry setup Amplitude scaling, trace editing Pre-stack parameter selection Velocity analysis Post-stack processing

Processing Pitfalls

FX Decon Example
Before FX Decon

After FX Decon

After Yilmaz, Seismic Data Analysis Volume 1, 2001, p. 975

3

Processing Pitfalls

FX Decon

Benefit
Enhances coherent signal Improves S/N

Detriment
False alignments Mixed wormy appearance Distorts static effects Phase and amplitude effects

Processing Pitfalls

Processing Pitfalls

Processing Pitfalls

Why Avoiding Pitfalls is Important

Save time and money More reliable seismic interpretations Obtain best quality data Translates to more E&P business success Avoid drilling based on artifacts and processing errors Application of vigorous quality assurance focused on travel time, amplitude and phase accuracy, is a major competitive advantage

Processing Pitfalls

Pitfall: No data processor on exploration team

Data Processor Geophysicist Structural Geologist

The addition of a data processor to exploration team will help prevent many of the current pitfalls
7

Processing Pitfalls

Pitfall: Assuming that marine data never require static corrections.

Water Surface

Normal Velocity Uncorrected Reflection

Water Bottom

Low Velocity Fill

Buried Erosional Surface

Corrected Reflection

The subwater layer may be just as complex as onshore low velocity layer
8

Processing Pitfalls

Pitfall: Assuming that conventional well surveys really are continuous and sonic log velocities are always true velocities
V V V

Conventional Well Sonic + CWS Survey (CWS) A combination of conventional well surveys and sonic logs can produce more accurate velocities Well

Sonic

Processing Pitfalls

Pitfall: Doing velocity analyses at regular intervals without regard to structure.

Velocity analysis locations that disclose analysis structurally-related velocity variations

Regularlyspaced velocity locations

Interpolation from points of structural discontinuities will give better values than across them.
10

Processing Pitfalls

Pitfall: Failing to take into account the effect of the Fresnel zone in mapping bed terminations.
Constant Velocity = V Z0 Wavelength = l = V / f Spherical Wavefront

Fresnel Zone

Fresnel Zone Velocity

11

2-way time Frequency

Processing Pitfalls

Pitfall: Routine application of spiking deconvolution

Wavelet Input Seismic Trace

Earths Reflectivity

White Earth Reflectivity: Good Spiking Deconvolution

Decon whitens input trace to match earths reflectivity

*
*

=
Output Wavelet Output Seismic Trace

Non-White Earth Reflectivity: Bad Spiking Deconvolution

Earths Reflectivity Input Wavelet Input Seismic Trace

=
Output Wavelet

Output Seismic Trace

Decon whitens input trace which distorts earths reflectivity

12

Processing Pitfalls

Pitfall: Overstacking.

Data quality improves dramatically as the initial traces are summed into the stack, but eventually a point of diminishing return is reached. Stacking additional traces beyond this point produces an overstacked result that may be less suitable for interpretation and more expensive to produce than lower fold data.

Overstacking may change or eliminate curved or dipping reflections and diffractions that are important to fault mapping.
Stratigraphic interpretation may be hampered because overstacking may attenuate high frequencies resulting in poor temporal resolution.

Compare well-processed single-fold sections or near-trace gathers to help judge the effectiveness of the stack and reveal a different view of the geology.
The fold of the stack should be determined by exploration objectives, not by how homogenous or pretty it looks.
13

Processing Pitfalls

Pitfall: Making changes in shooting and processing parameters without documenting these changes.

Filter 1

Filter 2

Phase change produces apparent fault.

14

Processing Pitfalls

Pitfall: Misusing seismic derived velocities

Seismic velocity calculations have been used to:

Determine abrupt changes in subsurface velocity

Estimate lithology, sand percent and fluid content

Determine burial history and depositional environment Locate over-pressured areas

Seismic velocities are truly an important geologic tool. Stacking velocity is rarely the acoustic velocity of the rocks, and may not have geologic significance. Stacking velocities are used to optimally align primary reflections on CMP traces They are really correction numbers used in the stacking process and may vary as much as twice the acoustic velocities because of dip or other irregularities. If DMO is not applied, their geologic value is questionable. Dr. Hewitt Dix showed (1954) how to determine interval velocities between two reflecting horizons. Interval velocities have true geologic significance.
15

Processing Pitfalls

And MORE

Interpretation of Migration artifact as a true seismic signal

Failure to recognize out-of-plane reflections in 2-D processing

Failure to recognize wave images that do not meet 3-D P-wave acquisition and processing assumptions (e.g. reflected refractions, shear waves)

16

Processing Pitfalls

Discussion

What other pitfalls have you encountered?

17

Processing Pitfalls

Processing Quality Assurance

18

Processing Pitfalls

Data Initialization
Parameter tests dynamite hole depth, charge size, etc. Instrument tests Prime shooting and in-fill shooting Seismic acquisition reports by QC representatives (bird-dogs) Results from any previous processing

Make sure all relevant data are in hand

Seismic data tapes

Survey/Navigation
Processed results Shell SPS format, UKOOA, etc. Navigation processing and QC reports

Maps
Topographic maps to tie data to elevation or cultural features Geologic maps that indicate surface lithology which may affect data quality Aerial photographs may show vegetation changes that may correspond to surface conditions that affect data quality

Documentation
Geologic setting, well logs especially sonic and check shots Exploration targets Project parameters costs, deadlines, etc.
19

Processing Pitfalls

Data Initialization

Display shot records, looking for problems

Impractical to display all shots except for small 2-D surveys Use brute stack and amplitude measurements
Map surface consistent amplitudes to find potential problem areas Source amplitudes locate bad shots Receiver amplitudes locate bad traces and surface noise sources

Time-slice preliminary 3-D stack or near trace cube to find geometry or navigation problems
Analyze first arrivals with linear moveout to detect bad trace distances geometry errors Some refraction statics and tomographic statics programs help detect and correct geometry problems during model building Be careful accepting data-derived geometry corrections because trace delays can come from many sources other than bad geometry

Plot source and receiver locations for land surveys Plot boat track and cable feather for marine surveys Plot CMP fold with and without in-fill shooting
20

Processing Pitfalls

Amplitude Recovery and Processing

Make sure bad traces (excessively noisy and very low amplitude) are eliminated from data volume
Use histograms of trace amplitude to set automatic editing parameters Evaluate the effect of trace editing on CMP fold Use maps of amplitude to identify areas of high noise

Make sure all reversed polarity traces are corrected and all traces are despiked
Common receiver stacks help locate reversed traces Some statics programs find reversed traces Find the spikes with amplitude analysis

Use amplitude attribute displays to check effect of amplitude processing (geometric spreading correction, gain)
Use graphs for 2-D and maps for 3-D Also use average frequency, peak frequency, etc.

Display some shot records after each stage of amplitude processing

21

Processing Pitfalls

Pre-Stack Parameter Selection

Determine optimum method of attenuating source-generated noise, if present, by applying processes to selected shot records and comparing displays Display brute stacks with noise attenuation applied; compare before and after Evaluate deconvolution parameters by comparing filtered displays of selected CMPs with one parameter at a time changed Display brute stacks and time slices with deconvolution applied; compare before and after; evaluate need for further phase correction to zero-phase

22

Processing Pitfalls

Velocity and Residual Statics Analysis

Evaluate velocities via iso-velocity and iso-time plots and maps Velocities should be consistent with structure and stratigraphy Evaluate effect of revised velocities by comparing before and after displays of: Inline and crossline CMP stacks Time slices in target zone Every 20th, or so, CMP gather corrected for NMO Evaluate effect of revised statics by comparing before and after displays of: Inline and crossline CMP stacks Time slices in target zone Every 20th, or so, CMP gather corrected for NMO Do these things after each velocity and residual statics analysis stage

23

Processing Pitfalls

Post-Stack Processing

Perform filter analysis via filter scans and select time variant filter (TVF) parameters Apply TVF to all QC outputs but not to data passed from one process to another Evaluate effect of each process by comparing before and after displays of: Inline and crossline CMP stacks Time slices in target zone

Run suites of test migration velocities (e.g 90%-110%) and evaluate over and under migration effects

24

Processing Pitfalls

Summary

Pitfalls abound in seismic processing and interpretation -- so be careful Most pitfalls result from accepting data at face value
Know the assumptions of processing steps and avoid misuse Look behind the scenes at tests and preliminary results to judge the quality and faithfulness of the final result

Add an experienced processor to your exploration team to obtain expert advice

Analyzing data plots at every step is too impractical, inefficient and ineffective for all but the smallest survey
Look for ways to map diagnostic parameters so that problem areas get the detailed analysis they need Relate data problems to surface conditions, cultural activity, topography, geology, etc. whenever possible because this builds your knowledge base for planning future acquisition and processing work Document your work for later reference

25