Several operations enhance our ability to predict the subsurface below the bottom total depth (TD) of the well when applied to zero‐offset vertical seismic profiling (VSP) data. Other key issues regarding the use of VSP data in this fashion are resolution and look‐ahead distance. An impedance log is the most useful form for presenting VSP data to look ahead of the drill bit. The VSP composite trace must first tie reliably to the surface seismic section and to the well log synthetic seismogram. The impedance log is obtained by inverting this VSP composite trace. However, before performing the inversion, we need to (1) correct the composite trace for attenuation effects below TD and (2) input velocities to provide low‐frequency information. An exponential gain function applied to the VSP data below TD adequately compensates for the loss of amplitude caused by attenuation. A calibration of the seismically derived velocities with VSP velocities yields the necessary low‐frequency information. These concepts are illustrated using a field data set and its subset truncated above TD. The output of these operations on the VSP data are compared to well log data. The question of resolution with these data was determined with a model VSP data set based on the well log data. The investigations indicate that the resolution attainable from look‐ahead data is on the order of 50–75 ft (15–23 m). This is one‐quarter seismic wavelength for the frequencies present in these data. In addition, the maximum look‐ahead distance for these data is shown to be easily 2000 ft (600) m and, perhaps, 4000 ft (1200 m5). By way of illustration, the techniques described and investigated 6were applied to an offshore VSP data set to yield an impedance log. After calibrating this curve with the well log data, the base of the target sand was correctly identified below TD. This prediction successfully yielded the thickness of the sand. Individual zones within the sand unit were identified with less confidence.
A look-ahead Monte Carlo simulation method for improving parental selection in trait introgression