Thursday, 19 December 2013

Marine Processing - Part 8 | Picking Velocities II

This sequence of blog posts will build up into a complete description of a 2D marine processing sequence and how it is derived.
No processing sequence is definitive and techniques vary with   time (and software), however the idea is to provide a practical guide for applying seismic processing theory.

The seismic line used is TRV-434 from the Taranaki Basin offshore New Zealand. It is the same marine line used in our basic tutorial dataset.

The data are available from New Zealand Petroleum and Minerals, under the Ministry of Economic Development under the “Open File” System.
The processing sequence developed so far:

  • Reformat from field data (in this case, SEGY)
  • Apply a combined minimum-phase conversion and anti-alias filter
  • Resample to a 4ms sample interval

  • Assign 2D marine geometry; 12.5m CDP spacing and 60 fold gathers
  • QC shots and near trace plots

  • Amplitude recovery using a T2 Spherical divergence correction
  • Amplitude QC and Edit using peak and RMS amplitude displays

  • Swell noise suppression using projective filtering
  • Interpolation to 6.25m group interval, 480 channels to unwrap spatially aliased dips
  • Tau-p transform 500ms robust AGC "wrap", 960 p-values and transform range from -1400ms -1 to +1400 ms-1
  • Tail mute to remove linear arrivals and linear noise
  • Predictive deconvolution: 32ms gap, 480 ms operator
  • Rho filter for low frequency compensation
  • Inverse Tau-p transfrom
  • Interpolation back to 25m group interval, 240 channels

In the last post we looked at some of the things that you need to bear in mind when you are performing velocity analysis.  

This post continues with that theme and looks at how to identify some key issues as part of our quality control checks.

The most important thing to remember is that whilst we are trying to pick the velocities of seismic reflections, there are other signals that we can pick by mistake.
The main problems we’ll see are:
  • Multiples - low velocity signals that are a delayed version of the primary. The delay is caused by the reflection energy reverberating between two strong reflectors, typically the sea surface and seafloor. There’s a nice diagram of this here.
  • Diffractions - where the energy is scattered (off an object or discontinuity); the tails of the diffraction have anomalously high velocities
The velocity we are picking is the “stacking velocity” – that is to say, it is the velocity which in conjunction with the hyperbolic NMO equation is going to flatten the CDP gathers and allow us to stack the data.

The “stacking velocity” has no other meaning apart from this, however in the situation where the source-receiver offset is small (which is generally less than 45 degrees) and the rock layers are horizontal and isotropic, the “stacking velocity” approximates the “root mean square velocity” (RMS velocity). 

This in turn is the RMS of the (P-wave) “interval velocity” within each of the isotropic layers.  There’s a useful page on these relationships here.

Most velocity analysis software will display the interval velocities while you work, which can act as a guide if you know the interval velocities you might expect in some common rock types.

Interval velocities (blue) plotted alongside the stacking velocities (black) on a semblance display

While the P-wave velocities for different lithologies vary significantly, the majority of multiples are caused by the water column. This leads to a corresponding interval velocity which matches that of seawater. Some typical ranges and values you might find are given below:

Expected Interval Velocities in ms-1
Around 1480ms-1 in general, although there is some variation due to salinity and temperature. I’ve seen velocities as high as 1540ms-1 on data from Antarctica, for example
Seafloor sediment
The soft, seafloor sediments typically ramp up with compaction; from seawater velocity (1480ms-1) very gradually up to 1750-2000ms-1  - depending on thickness and age 
Partially compacted Sandstone and Shale
Partially compacted sediments still have fluids trapped within the pores and so the velocities can have a large range. 2500ms-1 is a typical value but it can push up to 5000ms-1 in older rocks such as the “Old Red Sandstone” from the Devonian Period, where the pores have been closed by pressure and diagenesis 
Thin, reefy, limestones tend to have velocities close to 3700ms-1, while bulk limestones (or chalks) can be higher
Salt (Halite)
Usually a constant 4500ms-1, although you have to be aware of things like anhydrite that can “raft” in the salt if it becomes ductile
Basalt and Granite
“Bulk” igneous rocks tend to have high interval velocities of around 5500-6500ms-1
Anhydrite layers typically have very high velocities, around 7000ms-1

As well as incorrectly identified signals, we also have to watch out for simple mistakes and errors.  We can do this by looking both at the velocities (usually termed an isovel plot) and the stacked section.

Anomalously high or low values, relative to the overall trend, indicate a miss-pick or problem with the velocity field and show up as “bulls eyes”. These must be edited or removed.

A single velocity miss-pick (2500ms-1 rather than ~1700ms-1) on the isovels plot (left), and the corresponding impact on the stack (right, circled). The result is dimmed/faded reflectors in that region

While it is easy to see a single miss-picked velocity on this occasion, when dealing with a missed trend it can be much more challenging to spot. It is important to carefully compare the current stack to the previous one, looking for areas where there are improvements and places where the stack quality has degraded.

It’s also important to look at the stack while you are picking; a sudden lateral change in velocities can dramatically change the overall primary trend, but the multiples may be close to the primary trend you have been following.

The brute section (top), an incorrect velocity trend (middle) and improved picks (bottom); centred on the “tricky” area of the line i.e. the main thrust fault. The incorrect trend has followed slow multiples and fast diffractions, and so misses the high-velocity overthrust (around CDP 1320).

Of course, the stack is only one guide. We also need to look at NMO corrected gathers to make sure that the primaries are flat, the diffractions are up-dip, and the multiples are down-dip. It can be hard to judge “flatness” with a single velocity so comparing different velocity fields, or looking at 95%, 100%, and 105% of the stacking velocities, side by side is worthwhile.

Two CDP gathers (900 and 1800): NMO corrected with 95% (left), 100% (centre), and 105% (right) of the stacking velocities. CDP 900 looks flattened with 100%, but above 1000ms TWT, CDP 1800 looks to have been picked too slowly; some upward dip remains. CDP 1800 looks to be best flattened with 105% of the stacking velocity, above 1000ms TWT.

At this stage in the processing the limitations we have on picking are the presence of multiples, diffractions, and dip. The velocities in and around the over-thrust at the high CDP end of the line vary rapidly, and it is difficult to identify the correct trend.

To improve further, we need to use the existing velocity field to remove the multiples, and then move on to pre-stack imaging of the data.

By: Guy Maslen

No comments:

Post a Comment

Total Pageviews