Search and Discovery Article # 10053 (2003)
*Adapted from Transactions, AAPG Southwest Section Convention, Ruidoso, New Mexico, June 6-8, 2002. Special appreciation is extended to John G. Worrall, Program Chairman for the convention, and to Keith E. McKamey, President, Roswell Geological Society, for making the Transactions available for posting on Search and Discovery.
1Manzano LLC, Roswell, New Mexico ([email protected])
Nearly 50 wells have been drilled in Chaves County, New Mexico, searching for Siluro-Devonian reservoirs based on 3D seismic information. The success rate for this play has improved dramatically from a pre-3D success rate of less than 10% to almost 50%. Although the 3D program has doubled the number of producing wells, it appears that the 3D wells are going to recover less oil than was recovered from the pre-3D wells. High finding costs attributable to risk factors that have not yet been eliminated by the use of 3D seismic technology have recently brought this exploration play to a virtual standstill. However, the potential remains high for significant future reserve additions from the Siluro-Devonian formation in Chaves County. Future reserves will be found by using the lessons of the 1990’s to develop sound exploration techniques that will include the continued use of 3D seismic technology.
An estimated 50 wells targeting the Siluro-Devonian formations in Chaves County, New Mexico (Figure 1), have been drilled in period between 1990 and 2002 (not including Fusselman and Montoya tests). Most of these wells were drilled on the basis of 3D seismic information. Drilling depths range from under 7,000’ in the northwest to over 13,000 feet in the southeastern part of the county.
Prior to the early 1990’s approximately 260 Devonian tests had been drilled in Chaves County, all of which were drilled without the benefit of 3D seismic data. The pre-3D drilling resulted in only 21 producing well (10 new field oil discoveries and 11 development wells). The resulting combined wildcat and development success rate was only 8% (wildcat = 3% and development = 24%).
While the pre-3D success rate was poor, the production from the resulting fields was notable. The 10 Siluro-Devonian fields discovered prior to the use of 3D seismic had recovered approximately 4.5 million barrels of oil from 21 wells through the end of 1990 (2 15 MBO average per well).
During the early 1990’s the advent of more affordable 3D seismic data resulted in a renewed interest in the Chaves County Siluro-Devonian play. Affordable 3D offered a potential solution to identifying and imaging these relatively small structural traps and to limiting the number of development wells needed to drain the productive features.
Figure 2 shows the location of the estimated 225 square miles of 3D seismic data that has been acquired in Chaves County (33 surveys). The average survey size was just under seven square miles with actual survey size ranging from under four square miles to just over 20 square miles. Data acquisition and processing costs ranged between $20,000-$25,000 per square mile.
At least 47 Siluro-Devonian tests (31 wildcats and 16 development wells) were drilled based on the previously mentioned 3D seismic data (see Table 1 and Figure 3). Of the 47 wells documented, 22 were completed as Siluro-Devonian oil wells, 3 were completed in zones other than the Siluro-Devonian and 22 were plugged and abandoned.
Out of the 31 wildcats drilled, 10 were new field Siluro-Devonian discoveries, three were new field discoveries in formations other than the Siluro-Devonian, and 18 were dry holes. This equates to a Siluro-Devonian wildcat success rate of 30%.
Of the 16 development wells drilled, 12 were completed as Siluro-Devonian oil wells and 4 were plugged and abandoned, resulting in an impressive 75% development success rate. While the combined wildcat and development success rate of 47% is a dramatic improvement over the 8% pre-3D seismic success rate, it should be qualified in that seven of the 22 wells completed as Siluro-Devonian oil wells will not produce at least 50,000 barrels of oil and are not considered to be a commercial success. This brings the commercial well count down to 15 and the combined success rate down to 32%.
It appears that reserves from the 3D wells are going to average less than the 215 MBO per well average recovery from pre-3D wells. To date, approximately 2.4 million barrels of oil have been recovered from the 22 wells drilled on 3D. It is estimated that the 15 commercial wells will produce an additional 600,000 barrels of oil over their economic life resulting in estimated recovery from the 22 3D wells of 3.0 million barrels, for an average of 143MBO per well (only about 2/3 of the recovery from pre-3D wells). This figure is somewhat skewed in that over 2.8 million barrels will be recovered from the 15 commercial wells (~185 MBO average for the 15 commercial wells).
It is important to note the accuracy with which the 3D identified the structures. The 3D seismic data correctly predicted structure 75% of the time. Seismic pitfalls, such as velocity problems, have not had a significant impact in this play. However, in a play where structure was once considered as the only risk, there are other risks to be considered, such as fault seals, reservoir quality, and migration pathways.
In at least seven of the 25 wells that were not productive in the Siluro-Devonian structure was accurately predicted and reservoir quality rock was encountered, but the structures were water-bearing, the apparent result of younger faulting breaching the reservoir causing migration of oil out of the structure.
In three of the dry holes, the structure was accurately predicted but poor quality reservoir rock was the culprit. In two of the dry holes, structure was accurately predicted, but the wells were drilled too low on the feature to be productive. Faulting below the Mississippian unconformity was identified as the culprit in two of the dry holes. In one dry-hole structure was accurately predicted, the feature had reservoir quality rock, and the feature did not appear to have been influenced by younger faulting; yet the feature was not a hydrocarbon-bearing feature. This particular feature appeared to have been blocked from basinal hydrocarbon migration by a productive fault-trend feature immediately downdip and basinward from the barren feature. Most of the remaining dry holes were the result of incorrect interpretation of the seismic data and/or poor seismic data.
In the seven Siluro-Devonian producers that look like they will be non-economic, low relief structures and small reservoir size are the most common causes of the low production. Using estimated seismic costs of $5,000,000 (225 square miles x $22,500) and a rough estimate of leasehold, drilling and completion costs of $23,000,000 (all 47 wells included), the total cost of finding and developing (including all dry holes) the approximately three million barrels of oil attributable to the 3D program was around $28,000,000. The resulting combined finding and development cost of the entire program is estimated to be $9.33 per barrel.
Three oil fields--Chisum, Northeast, Racetrack, and Racetrack, Southeast--comprising the “Racetrack Complex,” are all located along the same fault trend in the central portion of Chaves County. The Racetrack Complex offers a good case study for the 3D play Siluro-Devonian play in that the complex contains examples of 3D successes as well as offers examples of the pitfalls that have resulted in dry holes. Within the complex there are breached structures, tight structures, structures that have been blocked from hydrocarbon migration and structural closures that are located too far downdip on features to be productive.
Approximately 25 square miles of 3D seismic data has been acquired over this area and nine wells were drilled based on interpretation of the data. Drilling resulted in four producers and four dry holes. Production through April 2002 from the four wells was 430 MBO and estimated recoverable reserves should be around 770 MBO (average 193 MBO per well).
Figure 4 is a structural contour map drawn on the top of the Mississippian unconformity. The most obvious feature is the high-angle reverse fault that bounds the west side of the complex. Throw on the fault is 100-200 feet (down-to-the-west) and initial movement on the fault was Late Mississippian to Early Pennsylvanian time with continued flexure into Wolfcampian time. The fault “scissored” with the eastern limb moving northward and in front of the western limb. The timing of the scissor faulting is not clear, although it is probably younger than Late Pennsylvanian time. This is probably the best example of “scissor” faulting in Chaves County. The faulting apparently resulted in that portion of the western limb in the “shadow” of the eastern limb being blocked from hydrocarbon migration, as evidenced by the barren structures along the southern end of the western limb. The presence of oil-bearing structures in the structurally lower eastern limb also seems to support hydrocarbon migration from the basin as opposed to upward along the fault planes.
Figures 5 and 6 are seismic lines across the complex, demonstrating many of the attributes that are common throughout the play. Productive features are always fault-bounded (usually by down-to-the-west faults), with varying amount of rollover into the fault. Thinning in the interval from the Lower Pennsylvanian clastics (Atoka) to the Siluro-Devonian is evident and supports the timing of the primary movement as Late Mississippian to Early Pennsylvanian. Additional thinning is common in the Permian Abo to Lower Pennsylvanian section, although it is normally relatively minor when compared to the thinning of the deeper beds. As seen in Figures 5 and 6, the Siluro-Devonian seismic event is often very difficult to pick. Mapping on the Mississippian unconformity seismic event has proven to be reliable with only a few exceptions.
Productive features found in the Racetrack Complex are low-relief features having only 25 feet to 125 feet of closure. Area under closure for these structures ranges from less than 40 acres to more than 300 acres; however, the productive area of these reservoirs is much less than the area under closure. Throughout the play, the productive area is normally only 25% to 35% of the area under closure. Why the structures are only l/3 filled is uncertain. Thin oil columns (10-50 feet) “sitting” on top of a very strong water drive combine to limit the economically recoverable reserves from fields across the play.
The following is a brief discussion of things to look for and things to avoid when exploring for Siluro-Devonian reservoirs in Chaves County, New Mexico. These observations and recommendations are far from “all encompassing” and are intended to be an aid for future exploration. There will certainly be exceptions to almost every point
Mentioned; so “buyer beware.”
1. Look along prominent fault trends, and when possible, look along established productive fault trends. Every productive feature that I have seen in Chaves County is fault-bound (usually on west side), and most are located on fault systems with established production along the fault trend. There are several prominent fault systems in the eastern part of the county that, to date, have yielded little to no Siluro-Devonian oil. In addition, very little oil has been found in structures located along minor splays off the more prominent systems.
2. Mapping the Mississippian seismic event is recommended. However, as one moves south and east into the deeper parts of the play, where there is a thicker Woodford Shale section, it is practical to map the Woodford event. Beware if the Woodford structure differs significantly from the Mississippian structure.
3. Lower Pennsylvanian (Atoka)-to-Mississippian isochron mapping has proven to be very effective; however, it is not always easy to tie into a reliable and continuous Lower Pennsylvanian seismic event. Abo-to-Mississippian isochron maps should be examined on every prospect. If no thinning is evident from the Abo to the top of the Mississippian, there probably is not a prospect.
4. Avoid small structures. With these reservoirs only being l/3 filled, small structures are usually uneconomic. Structures having less than 80 acres under closure generally have a productive area of less than 25 acres and recoverable oil of less than 30 MBO.
5. While these reservoirs are usually associated with low-relief structures, it is probably wise to avoid structures having less than 50 feet of relief.
6. Use caution when dealing with structures where the faulting cuts into the San Andres section. This becomes more common as one moves into the northern and western parts of the county; these faults seem to be more common on the east side of the structure. There are productive fields associated with this type of structure, but there are numerous structures that appear to have been breached by this type of faulting. At a minimum, when dealing with structures affected by younger faulting, make sure that there is good rollover into the fault(s). Rollover into the fault is not necessary when dealing with structures unaltered by younger faulting.
7. Where there is strong evidence of a scissor fault, avoid drilling in the “shadow” of the basinward limb of the fault. While the jury is out as to whether these features are sourced upward along the fault plane or by migration from the basin (or both); the Racetrack example provides evidence that a structure in the “shadow” can be barren.
8. Use 3D technology to explore for these reservoirs and do not wait to acquire 3D data for development. One or two properly placed wells will drain the vast majority of these features.
9. Try to use existing 2D seismic and/or other relatively cheap exploration methods to establish a good reason for shooting a 3D survey. While there have not been many large 3D surveys shot in Chaves County, it is interesting to note that the two largest surveys shot to date have resulted in no production.
10. Go to Church because you’re gonna need all the help you can get.