Re: How to estimate Sonic velocity above first logged depth?
As Petrel-User mentions, checkshots are the only reliable measure of the vertical seismic velocity from the borehole.
You only have accurate information about sub-surface velocities in the part of the well where you have both a sonic log and a checkshot survey. Outside of this, you are either missing the fine detail (checkshots only), or the information is not accurate (sonic only)
The sonic log has a number of drawbacks. Firstly its higher frequency - sonic our sound frequencies - than seismic vibrations, which means that the wavelength is smaller. This matter because the measurement will be more influenced by small-scale variations (and fractures) in the rock layer that seismic waves will not "see". The second issue is that the tool has a finite length; this means that it won't measure the response of thin beds (which are less than the tool length) properly. This smooths the response of the log, and of course its measuring transit times, not the actual velocity. If you smooth out the transit times, they will be slightly incorrect.
This is why we usually calibrate the sonic log with checkshot information (the "drift correction"), since as you progress deeper and deeper the cumulative effect of these errors can lead to significantly different travel-time measurements overall, and hence the a "drift" from the true velocity values.
In projecting back to the near surface you have a number of options, depending on he accuracy needed
- use the first checkshot to give a constant
The time and depth from the first checkshot gives you the average velocity to the surface. You could use this as an estimate, which of course would preserve the time-depth relationship accurately to the top layer. This replaces the entire 800m layer with a single velocity, so its likely not to correspond to any actual lithology.
- estimate a gradient
Depending on the degree of uplift, the top 800m of so of the section often show a strong compaction velocity in the seismic velocities. This means that a constant velocity is a poor estimate in many cases - which may or may not be important in your context.
Usually the velocity within a layer is expressed as V(z) = V(0) +Kz; where (z) is the depth, and K is the gradient. K is measured in "metres/second per metre". This can be used in two ways - Z can be relative to the top of layer (useful for compacted lithology) or to the actual surface. The gradient is the same in both cases.
A typical soft-sediment gradient for the near surface in the marine case is 0.6 (m/s per m)
You can estimate the gradient based on the information from other wells, measuring the velocity gradient from the calibrated sonic logs, or at a push from the seismic velocity information (converted to interval velocity and depth)
On land surveys you may also have an estimate for the velocity in the very near surface from the seismic data. They will have used a "replacement velocity" as part of the elevation and refraction statics solutions. This will normally be listed in the processing sequence and in the processing report. This can be used as a reliable V(0) at surface value as part of you modelling.
Hope this helps!