Using a case study from the Mungo Field in the Central North Sea, we investigate the relative impact of acquisition and processing improvements on 4D seismic repeatability. The results show that, while advancements in both have helped to reduce 4D noise, significant noise reduction can be attributed to processing alone. 4D noise can be thought of as any non-production-related amplitudes that are observable on a 4D difference section, and has both random and coherent components. Both are undesirable as they can mask any real 4D signal. A great deal of effort is employed to reduce noise levels by optimizing acquisition and processing between the 3D surveys, which are differenced to highlight the 4D signal. This paper studies the changes introduced by acquisition and processing improvements using the calibrated difference in reflectivity measure.4D noise levels were calculated between three acquisition vintages, upon which two processing flows were applied. An additional application of 'STAR', a 4D post-stack noise reduction process was tested (Zamorouev et al. 2006). Comparisons of repeatability were made using calibrated difference reflectivity (CDR) to highlight improvements through processing, acquisition and combinations of both. For each result a CDR value, map and amplitude spectrum were calculated to represent the 4D noise in that 4D difference volume.For the Mungo 4D Noise Case Study, the total contribution of all improvement is to drive down the 4D noise level by a factor of 3. Acquisition efforts to better match source and receiver locations between baseline and monitor surveys account for approximately one-third of this uplift. The application of STAR accounts for approximately one-third of the total noise reduction, while improvements to the 4D processing flow provide the remaining uplift.A typical method for the analysis of 4D anomalies is the study of 3D seismic difference volumes. The energy in a difference volume is a sum of the 4D signal, random noise which does not match between datasets, and coherent noise caused by the imperfect matching of events. Both forms of noise are undesirable as they can mask any 4D signal that would otherwise be interpretable, as shown in Figure 1. A great deal of effort is employed to reduce 4D noise levels and improve repeatability by optimizing acquisition and processing parameters between 3D surveys, the final objective being to remove both coherent and random noise from the difference volume and improve our ability to interpret the 4D signal. This paper studies the changes introduced by each acquisition and processing step of a typical flow, and attempts to identify where the greatest reduction in 4D noise is achieved.
Repeatability measurementsIt is common to use an attribute to measure the level of noise in a 4D difference volume. This can be recalculated as new processing or acquisition is introduced to determine the improvement in the seismic difference, or reduction in the noise level. One measure is the CDR (Dyce et al. 2004).This attribute computes the RMS a...