Chuck Skidmore
Emerald Research
During the late 1970s and mid 1980s, many fields were discovered in the Gulf of Mexico without the benefit of the high quality 3D data that is now available. As the fields depleted, 3D data was shot to identify any bypassed zones large enough for infill drilling. In some cases, the field was divested and smaller companies with lower overhead were given an opportunity to acquire new data and extract any remaining hydrocarbons. This task was complicated by the presence of the platform obstructing the new long cable data acquisition.
Even after the new 3D was acquired, finding the missed reservoirs proved to be a most difficult task. The major obstacle is that if hydrocarbons are present in high saturations, seismic amplitude cannot distinguish them from the low-saturated already produced reservoirs. These complications greatly increase the opportunity for economic reservoirs to be missed and left behind.
New technology is now being used to address the issue of low gas saturation. Traditional seismic traces are composed of the three elastic rock properties: compressional velocity (Vp), shear velocity (Vs), and bulk density (Rhob). Of these, the Vp is extremely insensitive to the saturation levels of hydrocarbon present. In low impedance sands, Vp will be just as anomalous in reservoirs with 10% gas as it will in the economic high-saturation cases. This means that most reservoirs will produce an amplitude response that is just as strong for both economic and uneconomic saturations. Of Vs and Rhob, Rhob will decrease in reservoirs with high hydrocarbon saturation, generating a large contrast with the surrounding rocks, and increase linearly as the hydrocarbon content decreases. If only low saturation hydrocarbons are present, the Rhob contrast with the surrounding rocks is no larger than that of a wet sand reservoir.
The Vs term should increase in magnitude compared to a low gas sand reservoir, or wet sand, but the increase will only be proportional to the density decrease. Of the three elastic properties, clearly the density term is the easiest to understand and use. By decomposing the seismic trace into three separate datasets representing the elastic components, the density term can now be isolated and used to highlight new prospects in old fields.
Mixed reservoirs
A mix of old and new producing properties spread out over a five-block area, with two main levels, was chosen to demonstrate the validity of this technique. Besides the obvious structural complexity, the reservoir sands also exhibit a high degree of stratigraphic variation. The area has been active since the late 1970s, with drilling continuing into this year. The data was acquired in the mid 1990s. This provides us with a dataset that has depleted reservoirs, producing reservoirs, and new reservoirs all mixed together in the same area.
The fields (see figure) are labeled A, B, C, and D from left to right, with field B's production coming from a deeper horizon, which will be discussed later. The horizon amplitude strength around fields A, C, and D is uniformly strong even though field D was the only one fully active during time of data acquisition. At the same time, field C was only producing from two wells and field A was not producing at all. In addition to the areas already drilled, there are numerous untested amplitudes, many large enough to support a well. Clearly, amplitude alone is not robust enough to discriminate between areas that have economic saturations from those that could be costly wells drilled into already depleted reservoirs. Stacked seismic data is not the proper tool for the task of identifying bypassed pay.
The density contrast-only cube is a new type of seismic attribute where the density contrast controls the magnitude of the response. Only the areas that have a density contrast equal to, or exceeding, the current production from field D are highlighted. Field D shows both amplitude and density anomalies that track along the structural trends that form the hydrocarbon trap. Field A has strong amplitude anomalies, but no density contrast, indicating only low saturations of gas remaining. The subtle small density anomalies in field C correspond to the areas where the two wells were still producing. Their production ended within three years of the data acquisition, having been online for over 15 years.
The density anomalies around field D are not the only ones on the horizon. Circled in red on the figure are at least four other potentially economic anomalies. Field D has density anomalies along the same producing trend that have not been tested. Field A does not show any, but field C does have a number of significant anomalies that have not been drilled, two of which are circled in red. These anomalies lie outside the current field radius, but could be reached and are probably a combination of structural and stratigraphic traps. Even field B has a, small, untested anomaly having both strong amplitude and density contrast. These potential new targets that have been bypassed are in danger of being left behind.
Deeper zones
Field B produced from deeper reservoirs. The producing wells appear to cover the reservoir adequately, with a downdip amplitude termination possibly marking the gas-water contact (blue dashed line). These data were acquired about midway through the production life of this portion of the field. The complex amplitude patterns indicate that, although the reservoirs are at the same level, there may be potential stratigraphic barriers or non-uniform drainage radii due to the complexity of the deposition.
On field B's density data, it is even more apparent that the reservoir and/or the production is discontinuous. The reservoir has been broken into three main areas labeled 1, 2, and 3 on the figure. The center of the reservoir, area 2, has little or no density contrast and appears to be completely produced. The field reports confirm this with 86% of the production in this area going offline a few years before the data was shot. Area 2 has yielded the greatest amount of gas in the least amount of time, with average flow rates that were twice that of any other area. What once was a uniform gas-water contact now is shifted more up the structure on the left (area 1) than on the right (area 3).
Area 1 has produced more uniformly, possibly because the wells intersect the horizon in the stronger parts of the density anomalies. The flow rates in this area are also higher than that of area 3. This would imply cleaner and/or thicker reservoir around the wellbore, both of which would cause stronger density contrast. A comparison with the amplitude results shows great disparity and, therefore, more residual saturations present. This is the primary reason for the interpretation of a new contact farther up the structure (red dashed line). Production in this portion of the reservoir continued for another six years after the data was shot.
Area 3, on the right side of the reservoir, shows only a slight difference in the size and downdip termination on the density anomalies compared to the amplitude data. The density data is definitely more discontinuous, and only a few of the wells intersect strong density anomalies. The production in this area has been prolific, but the average production rates are significantly lower than the other two areas. The production in this area is captured midway though completion, so the differences could be due to the removal of hydrocarbons or lateral lithologic variations. The density anomaly circled in green has not been touched by any of the current wellbores and still has both strong amplitude and strong density contrast. With only one well up-structure from it, the chances are good that this feature is not being produced. The area did stay active for another seven years, so a new dataset would be needed to confirm this as a current fluid anomaly.
Conclusions
The data used for the analysis represents only one temporal snapshot of the whole area. 4D data would be a more effective tool for the field analysis, but because of the saturation issue, this data type is not often acquired. To truly evaluate any remaining hydrocarbon potential, a new dataset, to provide a 4D solution, coupled with the density technology would be the most promising. We see significant differences between the conventional amplitude data and the data extracted from the density cube. Rock property analysis and similarity of the responses to known fluid and reservoir properties of the reservoirs are very supportive that the density cube data can indeed identify potential bypassed pay targets in and around existing fields.