Time-lapse seismic - 3-D Seismic Interpretation - page 183

so to what extent are surveys repeatable? What can be done to match one survey to
another if acquisition parameters are not precisely identical?
Field management issues: when do the time-lapse data need to be available to have
an impact on real-world decisions such as where to place an infill well?
An excellent survey of the topic has been given by Jack (1997 ) .

8.1 Rock physics

Modelling of the effect on rock properties of changes in fluid fill, pressure and temper-
ature is an important first step. The issues have been well summarised by Wang (1997) .
In general, time-lapse is most likely to work when there are large changes in fluid com-
pressibility with production, and in rocks with low elastic moduli (poorly consolidated
or with open fractures). For example, time-lapse has been very effective in following
steam injection for enhancing recovery of high-viscosity oil from high-porosity sands
(Jenkins et al. , 1997) . In other cases, it may not be certain whether a change in seismic
response is great enough to be measured until after a time-lapse survey has been carried
out. Effects of changes in fluid saturation are modelled using the Gassmann methodol-
ogy explained in chapter 5 . An empirical observation is that fluid effects are often larger
than the Gassmann theory would predict. This may be because clay content is often
not sufficiently taken into account when assessing the properties of the rock frame, or
because hydrocarbon saturation is patchy rather than uniform. For example, injected
water will move preferentially along high-permeability layers within a reservoir and
produce a patchy mixture of hydrocarbon and water at the seismic scale.
Understanding the effects of pressure change is more difficult. If the reservoir pres-
sure is allowed to drop, then the net overburden pressure on the reservoir rock increases.
For unconsolidated sands or some high-porosity limestones, this may lead to a large re-
duction in pore volume and thus a significant increase in impedance. For well-cemented
sands, the pressure effect will be small. To estimate the likely effect in any particu-
lar case, the main source of information is laboratory measurement of rock velocity
under controlled pressure conditions. Most laboratory measurements are made at ultra-
sonic frequencies (around 1 MHz) so that results can be obtained from small samples.
Unfortunately, the Gassmann theory is limited to low frequencies; real rocks show con-
siderable change of velocity with frequency (usually referred to as dispersion ) between
the seismic and ultrasonic cases. At high frequencies, there is not enough time for pres-
sure to equilibrate between pores as the seismic wave passes, and rocks appear stiffer
than at low frequencies. A method of predicting the size of this effect by modifying the
dry rock moduli has been described by Mavko et al. (1998). An additional complication
of laboratory measurements on core samples is that they may not replicate in situ be-
haviour, mainly because clay minerals can be modified by interaction with pore fluids
or by drying.