Reliance on prestack time-migrated seismic data to define structural highs without incorporating all subsurface data and without taking into account the regional and local lateral depositional trends may result in dry holes or poorly positioned production wells due to local velocity changes, which are usually caused by some depositional or structural phenomenon. Tying check-shot control to depositional units may reveal those phenomena and permit assumptions to be made about velocities in areas beyond check-shot control points. We discovered a significant gas accumulation in an area surrounded by dry holes and marginal wells in the Vicksburg Formation in McAllen Ranch Field, Hidalgo County, Texas, by treating a seismic velocity anomaly as a geologic problem and by simple application of arithmetic and geometry to a 3D velocity model. Due to the effects of the anomaly, seismic data displayed in time gave no indication of the existence of a 325 ha (800 ac), 150 BCFG anticlinal structure. A subsurface model that accounted for the velocity anomaly was able to predict its extent and severity by readily identifiable thickness changes in the anomalous units. The resulting discovery yielded a sevenfold increase in field production within a two-year time span.
Seismic velocity anomalies, also referred to as fault shadows when associated with velocity differences on opposite sides of a fault, have frequently been documented in the literature. Recent examples from the Texas Gulf Coast (Allen and Bruso, 1989; Lowry et al., 1995; Fagin, 1996; Meyerhoff and Braddock, 1998) demonstrate that fault shadows can mask the presence of structural accumulations of hydrocarbons. These examples, from the Wilcox and Frio trends of south Texas, demonstrate the successful positioning of one or two additional wells. These well locations would have been off structure and wet using only time data without appropriate depth correction.
The example presented in this study, from McAllen Ranch Field in the Vicksburg trend in Hidalgo County, Texas, resulted in the identification of 25 well locations on a 325 ha (800 ac) structure that has produced more than 150 billion cubic feet (BCF) of gas. The method used to predict the size and orientation of the structure did not involve reprocessing of seismic data or complex modeling, but instead used computer-based implementation of a simple mathematical process and an understanding of the 3D geometry of the unit that was causing the velocity anomaly.
McAllen Ranch Field is located in northwestern Hidalgo County, Texas (Figure 1), and it produces from the Oligocene-aged Vicksburg Formation from approximately 4250 m (14,000 ft) true vertical depth (TVD). The field is overlain by sands and shales of the Frio Formation and various Miocene-age sediments (Figure 2). The Vicksburg Formation at McAllen Ranch Field developed as a series of coarsening-upward shelf delta wedges downthrown to a major expansion fault (Figure 3). The basinward-thinning sedimentary wedges form traps when draped across an east–west-oriented, deep-seated high.
The part of McAllen Ranch Field examined in this study lies at the distal end of the zone of Vicksburg expansion (Figure 3), on a lease operated by Chevron USA Inc. The reservoir referenced in this study is the Guerra Sand, a 150–300-m (500–1000-ft)-thick sand that was deposited in a proximal delta-front environment. Overlying the Guerra Sand is a series of thinner lower Vicksburg sands which in turn are overlain by approximately 900 m (3000 ft) of upper Vicksburg shale. Overlying the Vicksburg are sands and shales of the Frio Formation (1800-m [6000-ft] thick) and Miocene formations (900-m [3000-ft] thick). Basinward, the Vicksburg section is cut off by a major fault, locally known as the Monte Christo Fault, which expands the Frio section (Figure 3).
The Guerra Sand produces from a southwest–northeast-oriented anticline. The structural configuration as interpreted prior to 2005 is shown in Figure 4a. At that time, the reservoir was believed to be fully delineated due to the presence of several dry holes and marginal wells along the southern edge of the structure. Wells developed in the Guerra Sand exhibit a common subsea gas-water contact of (), and in general, any well that encountered the sand top below a subsea depth of () did not have sufficient pay thickness to be considered economic.
It is not unusual to use time mapping of seismic data followed by depth conversion for structural positioning of a well location. This generally results in the well encountering the objective reservoir at or near the expected subsurface elevation relative to nearby wells unless there are strong lateral velocity variations. An example of a potential strong velocity variation would be in an area of growth faulting. Care should be taken to understand the differences in average interval velocity that occurs due to thickening of the section on the downthrown side of the fault and the effect those differences have on the dip of seismic reflectors in the footwall of the fault. Examples cited in the literature demonstrate wells that were drilled on apparent structural highs on time data that turned out to be structurally low due to those velocity differences.
An interpreter may incorrectly believe that using prestack time-migrated data with a nearby seismic check-shot survey available for depth control will reduce any uncertainties due to lateral velocity variations. Such an example is shown in Figure 5a, in which a proposed development well location (#99) appears to lie more than 30 m (approximately 100 ft ) updip in the same reservoir from a well (#72) in which a seismic check-shot survey was recorded.
The original proposed location of the #99 well in Figure 5a was based solely on the apparent northwest dip observed on the prestack time-migrated data that appeared to place it updip from the #72 well. The check-shot survey provided only a single point of depth control and was not sufficient to address the issue of lateral velocity changes. The pitfall of relying only upon the appearance of the seismic data and not taking into consideration any subsurface data that could have added to the structural understanding of the area of interest resulted in a proposed location that would have been a dry hole. Fortunately, the pitfall was recognized during the planning process for the well, and an alternate location was selected. Had the well actually been drilled in such a position, it would have encountered the reservoir low to the #72, at a location below a known gas-water contact, as shown in Figure 5b.
The pitfall was resolved just prior to the well staking by a thorough search for subsurface well data that revealed a dipmeter from the #72 (Figure 5b) that was not known to exist at the time the #99 was originally proposed. The dipmeter indicated 20° southeast dip at the sand top, in direct conflict with the seismic data. Fortunately, the team that proposed the well chose to honor the well data, specifically the dipmeter, over the seismic data, even though a reasonable explanation for the apparent northwest dip on the seismic data could not be provided at the time.
One should assume that a seismic interpretation in time accurately portrays geometry in depth unless one has an understanding of the local seismic velocities. In this example, a fault shadow effect was suspected as the cause of the anomaly because of the presence of the Monte Christo fault that expanded the Frio section in the area. The method used to quantify the effect of the fault shadow began with creation of arbitrary 2D lines through the 3D seismic volume to tie all wells with check-shot surveys to each other. It quickly became clear that a 450-m (1500-ft)-thick interval in the lower Frio Formation known locally as the Massive Frio had interval velocities that were as much as 50% faster than the overlying or underlying units. The Massive Frio is the zone that was involved in the expansion across the Monte Christo Fault, reaching a thickness of greater than 2100 m (7000 ft) on the downthrown side. This had the effect of a giant lens that distorted all of the data beneath the fault plane (Figure 6; see also Figure 5a and 5b).
The average measured log value of bulk density of the shale units within the Massive Frio is , compared to for the shale units in the overlying Upper Frio or the underlying Upper Vicksburg sections. The higher shale density in the Massive Frio may be the cause of the faster velocity within the unit. A determination of the depositional source of the difference in densities is beyond the scope of this study.
The top and base of the Massive Frio are readily correlated on well logs and can be easily tied to strong, continuous events on the 3D seismic data. This permitted a 3D velocity model to be created involving four simple steps (Figure 7):
Using the 3D volume, map the time surfaces bounding the top and base of the three units with distinct velocity differences: (1) ground surface to top of Massive Frio (average interval velocity of  to ), (2) top of Massive Frio to base of Massive Frio (  to ), and (3) base of Massive Frio to top of Guerra Sand (  to ).
Using the 3D volume, create isochron volumes between these surfaces.
Create contoured average interval velocity maps for each isochron volume. Because the number of wells with check-shot surveys was very limited, key wells for which an isochron value and interval thickness could be determined with confidence were used to derive an average interval velocity. The velocity maps were hand-contoured to impart a “geologic bias” that conformed to the local depositional and structural strike in areas of sparse control.
The surfaces, interval thicknesses, and velocity maps were exported to GOCAD for efficiency in performing the following calculations. For each unit, the isochron value was multiplied by the contoured average interval velocity to yield a predicted thickness for the unit. Next, the units were summed and the result was a prediction of the shape and orientation of the top of the objective Guerra Sand (Figure 4c).
The resulting prediction of the structural configuration of the Guerra Sand top showed a structural high extending almost 2 mi southwest of the successful #99 well (Figure 4c). Based on this interpretation and the favorable results of the #99, Chevron undertook a two-year development drilling program in the newly identified structural field extension, named the South Block.
The final structural configuration after 25 wells were drilled very closely matched the model’s prediction of the size and shape of the structure (Figure 4d). More than 150 BCF of new gas reserves were discovered, and field production increased from 22 million cubic feet (MMCF) per day to more than 170 MMCF per day.
Several valuable lessons can be learned from the success shown in this example. The most important is that one should always honor well data, even if it directly conflicts with what the seismic data are suggesting. In the example presented, a dipmeter provided a more accurate resolution of the subsurface than did the seismic data. Second, simple solutions can work quite well. The velocity model used in this study did not require any seismic reprocessing but only involved simple mathematical calculations. The postdrilling result closely matched the model’s prediction. Finally, the success of this project shows that there are still substantial reserves to be found in a seemingly mature province. McAllen Ranch Field is now 50 years old and has been thought to have been fully delineated for the last 25 years, but new discoveries can still be made.
The author would like to thank the reviewers for their thoughtful evaluation of the paper and for their many suggestions for improvement. Thanks are also extended to Seismic Exchange Inc. (SEI), for their permission to publish the seismic data shown here. The author would also like to give special thanks to M. Allen, M. Hugele, D. Lucidi, and T. Maciejewski of Chevron, without whose support the success of this project would not have been possible.
Richard C. Bain obtained a B.S. in geology from Waynesburg College and an M.S. in geology from Ohio State University. He is a development geologist for the Delaware Basin with Chevron’s MidContinent Business Unit in Houston. During his 35 years with Chevron, he has worked at a variety of assignments in south Louisiana, the Gulf of Mexico, the Permian Basin, east Texas, and south Texas. For 15 years, he was a development geologist for Chevron’s Lobo Trend properties in Webb and Zapata Counties, Texas, and Chevron’s Vicksburg Trend properties in Hidalgo County, Texas.