Geostatistical Processes for 4D Monitoring

In seismic acquisition, several campaigns can be carried out on the same area. This is the case when testing different acquisitions systems or for 4D monitoring. When computing a second acquisition on a same area, one question may be to decide reshooting less lines than the previous campaign. The first point of this paper is to look at the possibility to complete by means of geostatistical interpolation methods under-sampled reshoot campaigns. To perform this completion, several options are available: interpolation by kriging of the under-sampled campaign using the existing lines and interpolation of the under-sampled campaign integrating the full first seismic campaign. The used example is a 4D monitoring case. There are two full campaigns shot at the same locations but different times. A selection selecting one line over two has been applied on the second seismic campaign in order to “simulate” an under-sampled survey. In order to fill-in this incomplete campaign, two geostatistical methods have been applied: kriging and collocated cokriging. It shows the benefit of multivariate techniques in order to improve the estimation of under-sampled campaign. The second point of this paper is to extract by geostatistical method the common geological part of the two seismic campaigns