A GEOLOGIC MODEL OF FRACTURE OCCURRENCE IN CARBONATE ROCKS IN YATES FIELD Juan Florez1, Amgad Younes2, Richard Steevers2 and Diana Sava1 1 School of Earth Sciences, Stanford University 2 Co.

Abstract located in the Guadalupe Mountains, about 100 miles The structural style observed at Yates Field, east of Yates (figure 2). By this methodology we combined with the analysis of fracture systems within were able to characterize the structural style of the a reservoir analog observed at outcrops, as well as a field area in addition to the brittle deformation style rock physics analysis of well log and core data, of the San Andres Formation at an outcrop analog. provide the basis to construct a conceptual model of fracture-occurrence at Yates Field. Two main factors control the density and spatial distribution of fractures: location with respect to faults, and lithofacies. Our results indicate that, in this particular case, the location with respect to normal faults is the dominant factor controlling fracture distribution. In addition to this, a relationship between fracture density and lithofacies was observed both at outcrops and in the subsurface data. Finally, fracture density was different for each particular set, as a consequence of the impact of the coeval stress field on fracture-occurrence.

Introduction Yates field is located on the Central Basin Platform (CBP) of the Permian Basin in western Figure 1. Location Map of Yates field showing the (figure 1). It was discovered by surface paleogeography of the Permian Basin (After, geology in 1926. Oil originally in place exceeds 5 Craig, 1988). MMB. After producing more than 1.3 MMB during 74 years, the field is still delivering about 25,000 At surface, Yates Field is a broad, subtle bopd. The geological model for the field is based on anticline depicted by the Lower Cretaceous. It is well logs and production data. Seismic information is surrounded by normal faults arranged in an limited to four 2-D lines, very few for the 91 km2 orthogonal pattern. Immediately north and south of area. In spite of the high density of wells in the field Yates, the system of NW-trending normal faults is (average 10-acre spacing), there is still significant predominant and cut pre-existing NE-trending faults. uncertainty regarding the existence of faults, spatial Immediately west of Yates the NE-trending normal distribution of fracture density, and orientation of the fault system is more conspicuous. In addition to these conductive features. faults, there is a system of sub-vertical, mutually Reservoir heterogeneities can be summarized in orthogonal fracture sets. The younger set of fractures three main groups: (1) stratigraphic or is N50-70°W and abuts against the older, N40°E, set. sedimentologic, (2) structural or geo-mechanical, and Fracture densities are different for the different (3) diagenetic or geochemical. The present study is sets. In addition to this, two main factors were found focused in understanding the factors controlling to control fracture density within the San Andres structural heterogeneities at Yates Field. Formation: (1) The location of the outcrop in respect In order to understand the distribution of to major faults, and (2) the lithofacies. fractures in carbonate rocks of Yates Field, we Slip along the fault planes produces stress employed a twofold analogue study. First, we studied concentrations at the tips and alters the stress field in the structural style in the field area, analyzing the the surrounding area. Therefore, fracture density patterns of faults and fractures within the cretaceous could not only by a function of the position in respect limestone cropping out at Yates. Afterwards, we to the faults, but also of the amount of shear on those studied the fractures and faults within the San Andres faults. On the other hand, the conspicuous difference Formation, the main reservoir at Yates, at outcrops in fracture density between the lithofacies observed at

Stanford Rock Fracture Project Vol. 12, 2001 C-1 outcrops, indicates a strong control of facies on the (Adams, 1940) proposed that Yates anticline was elastic properties of these rocks, as should be developed by differential compaction, probably expected. These observations were confirmed by rock related to the underlying structure. physics analysis of subsurface data. Finally, sets of fractures with different orientations showed different fracture densities and apertures. Since the fracture- saturation and aperture is a function of the magnitude of the stress field, this observation indicates that stress field that originated the dominant fracture system, N 40° E, was higher in magnitude than the stress field that generated the N 50° W set.

Geologic setting Figure 2 shows the regional tectonic setting of the Central Basin Platform (CBP), where Yates is located, and its relationship to the outcrops at Guadalupe Mountains. Basically, the CBP is a major Figure 2. Regional W-E cross section illustrating the pre-Permian horst bounded by normal faults. The San structural setting of Yates and the visited Andres Formation is mostly a dolomitized limestone outcrops. (modified from Renfo et al, 1984). with different facies, deposited in a carbonate shelf platform during the Permian. The CBP apparently Stratigraphy controlled the facies distribution of San Andres and The stratigraphy at the Central Basin Platform the other units of the Guadalupian series. Carbonates can be divided into four main sequences separated by were deposited within the platform, while sands and three main unconformities: the pre-Permian, Permian- shale were deposited in the Delaware and Midland Triassic, and the Triassic-Cretaceous unconformities. basins, which were deeper depositional centers The four sequences are pre-Permian, Permian, adjacent to the platform. Triassic and Cretaceous. The Permian sequence is The time gap between the Triassic and the composed of the Wolfcamp, Leonardian, Cretaceous suggests that Yates area was above the Guadalupian and Ochoan series. Relevant for Yates is depositional base level during a long period of time, the Guadalupian series, which in this area is and therefore was exposed to erosion, or at least not composed of dolomitized limestones, mixed deposition. Considering the low deformation observed siliciclastic-carbonate rocks and evaporites in the Cretaceous rocks, it is inferred that Yates has (anhydrite-gypsum and salt). At Yates, this series is not suffered major tectonic deformation since the composed of the San Andres, Grayburg, Queen, Early Cretaceous. It is also inferred that this region Seven Rivers, Yates and Tansill formations (Donogue was part of the foreland basin during the Laramide and Gupton, 1956). The Ochoan series is reduced to a Orogeny. Probably it has been exposed to erosion thin interval (25’-200’) of shale, anhydrite and since the early Tertiary. dolomite (Figure 3). About 200 km west of Midland and Yates is the The San Andres Formation is mainly composed rift basin, which is part of the Basin and of dolomitic limestone deposited in a shallow marine Range extensional province, a tectonic environment platform, close to a ramp (Kerans et al, 1994; that has been active during the Cenozoic (Russel and Sonnenfeld, 1991). The location of the ramp was Snelson, 1994). Field observations suggest that Yates controlled by normal faults. There is an eustatic is within a normal-fault stress state, with maximum unconformity between Grayburg and upper San compressive stress in WNW direction. The same Andres rocks. Above the unconformity a relatively stress state is suggested by drilling-induced tensile thin package of shale, dolomitic mudstones and fractures interpreted from FMI logs. The stress sandstones are found, corresponding to the Grayburg regime suggested by these two independent sources is and Queen Formations (Donoghue and Gupton, in agreement with the results derived from regional 1956). studies (Zoback and Zoback, 1991). The Seven Rivers is a thick section (about 400’) The CBP structure does not seem to have had a of anhydrite, followed by 55’-90’ of sandstone. The major impact in the facies distribution of the Top of the Guadalupian series is a thin interval of cretaceous sediments. Nevertheless, the fact that the shale, sandstone and anhydrite, that laterally subtle anticline resembles the structure of the Permian correlates with a gross package of salt, prominent in rocks at subsurface (Hennen and Metcalf, 1929) the surrounding areas of the field, but virtually absent suggests that this subsurface structure is still at Yates (Donogue and Gupton, 1956). impacting the deformation in the area. Earlier studies

Stanford Rock Fracture Project Vol. 12, 2001 C-2

The Triassic system is reduced, in this area, to a stratigraphy (Tinker et al, 1995). This facies thin unit (0-200’) of sandstones and claystones. It is distribution can also be controlling the occurrence of bounded by unconformities: The Triassic-Permian caves, which are associated with high productivity unconformity (below), and the Cretaceous-Triassic areas (Craig, 1988). unconformity (above). The big time gap represented Kerans et al (1994) and Sonnenfeld (1991) by the Cretaceous-Triassic unconformity indicates studied the sequence stratigraphy of the San Andres that this area was above the stratigraphic base level Formation in the areas where its depositional setting during a long period of time (about 100 m.y.). was quite similar to Yates. Tickler et al (1995) The Cretaceous is composed by a transgressive present a summary of the stratigraphy of the upper depositional cycle that starts with sandstones and San Andres at Yates. Basically they divide the section continues with shale and limestones. The remarkable in three genetic cycles within a major stratigraphic lateral continuity of the limestones suggests sequence. Each cycle is composed of dolomitic deposition in the inner shelf. These cretaceous rocks wackestone and mudstone to the west, and dolomitic are flat and virtually not deformed, but there are few fusulinid packstone and grainstone to the east (Figure normal faults and some large and subtle folds. 4). The development of caves to the top of every cycle is interpreted to be the consequence of LITHOLOGY syndepositional subaerial exposure of the packestone

Main to grainstone facies, probably controlled by a system

SYSTEM

SEQUENCE

GROUP FORMATION Thickness(m) Unconformities of early regional fractures (Tinker et al, 1995; Craig, 0 1988).

FREDER.

ACEOUS ~2 miles

CRETACEOUS 200

WACHITA/ FREDER. WACHITA/

CRETACEOUS Stanford Seismic Project

LOWER CRET NO CAVES CAVES TRINITY Cretaceous-Triassic Tectonic 400 Timegap ~80 m.y. Stratigraphic cross section A-A’

TRIASSIC TRIASSIC Triassic-Permian Tectonic(?) Time gap < 45 m.y.

OCHOA

OCHOA

600

YATES Western boundary of caves Cycle 3 Cycle 2 Cycle 1

RIVERS SERIES) 800

SEVEN SEVEN RIVERS STRATIGRAPHIC CROSS SECTION A-A’

WHITEHORSE Vertical exageration 1.8

PERMIAN

(GUADALUPE LEGEND 200 ft 1000 Limestone

PERMIAN PERMIAN (GUADALUPE SERIES) 3000 ft

QUEEN Interbedded Sandstoneand shale Dolomitic Mudstones and Shale Intra-Permian Eustatic(?) Mixedshale, sand and Grainstones and Packstones with caves Time gap < 1 m.y. dolomite Packstones and Wackestones

GRAYBURG 1200 Anhydrite Wackestones Sandstone Figure 4. Outline of Yates field showing the western

ANDRES DolomiticLimestone limit of cave occurrence, and the stratigraphy of

SAN ANDRES

SAN and Dolomites the upper San Andres at Yates (after Tinker et al, 1995). Figure 3. General stratigraphy of the overburden and reservoir sections at Yates field (modified from Donogue and Gupton (1956). Faults and Fractures at Yates The surface geology suggests that faults and Stratigraphic heterogeneities fractures are the main structural heterogeneities at Kerans et al (1994) grouped the carbonates of Yates. It could be argued that there is not direct San Andres in four main rock types: mudstones, evidence faults at surface can be extended to the wackestones, packstones and grainstones. The reservoir. In fact, previous studies interpreted these distribution of these rock types is controlled by their normal faults as a consequence of salt and anhydrite genesis, and is intimately related to the sequence mobilization (Adams, 1940). However, there is a

Stanford Rock Fracture Project Vol. 12, 2001 C-3 consistent pattern between the orientation of the faults unaltered limestone. Both the calcite veins and the and fractures observed at different localities and the change in color evidence that this fault was a fluid orientation of fractures within the San Andres conduit in the past. Formation at Yates. There is also a systematic pattern of normal faults that is coincident with the orientation of major lineaments at surface, productivity trends at subsurface (Tinker and Mruk, 1995), and the current state of stress in the field.

Figure 6. Detailed view of a fault zone showing the gouge at the fault-core, the breccia zone and the adjacent fractured rock. The throw in this fault is about 6 m; it is a branch of the normal fault Figure 5. An example of the normal faults observed shown in figure 5. north of Yates field. The estimated throw is 20 m.

The azimuth of the older set of faults is 040° (N40°E). These faults have less continuity because are cut by the younger ones. The amount of slip is locally smaller than the WNW faults. This set is more systematic and continuous further east of Yates, where the other WNW set is not well developed. Two main sets of normal faults were observed within and around Yates field: WNW (290°) and NE (040°). The strike of the younger set is WNW (290°). These faults are 2 to 5 km long, and have spacing from 0.2 to 1.0 km. Observed vertical slip ranges from 2 to 25 meters (figure 5). Direct observation of a fault exposure shows that the fault zone is composed of fault rock, a very well cemented gouge and breccia Figure 7. Orientation of fractures in the San Andres (figure 6), whereas the surrounding damage zone and Formation at Yates, sector NW. the tips of the faults are areas of high-density of fractures. Within the fault zone there are thick calcite Four main sets of joints were measured in the veins, and the breccia fragments with a light pink cretaceous limestone at outcrops in Yates (figure 9). color that contrasts against the light bluish gray of the

Stanford Rock Fracture Project Vol. 12, 2001 C-4

The orientations of these sets are 290° (WNW-ENE), (figure 10). Age relationships were not sorted out for 040° (NE-SW), 340° (NNW) and 70° (ENE). The the other sets. Both WNW and NE sets are consistent density and distribution of each set is not uniform and present not only at Yates but also further west, at across the field, but in general the NE-SW and Guadalupe Mountains. WNW-ENE sets are more consistent. The same patterns of fracture orientation, density and distribution are observed in the San Andres Formation in the subsurface (figures 7 and 8). The dominant fracture pattern varies from well to well, but three main groups can be distinguished: (1) wells were the N 40° E set is dominant; (2) wells were the N 70° W set is dominant; and (3) wells were the NNW and ENE sets are dominant.

Figure 10. Abutting relationship between the N 40° E joint set (parallel to the pen) and the N 70° W joint set (younger).

Fracture systems in the San Andres Formation at surface Outcrops of the San Andres Formation were Figure 8. Orientation of fractures in the San Andres studied at three different localities in the Guadalupe Formation at Yates, central part of the field. Mountains, west of Carlsbad (New Mexico). The area is located west of the Permian Basin, and east of the N N Rio Grande graben. Large NW-trending normal faults, parallel to the graben trend, are found at the west of the Guadalupe Mountains. Three localities were visited: Wilson Canyon, Algerita escarpment at Lawyer Canyon and two exposures along the highway 83 (figure 11). Considering their position with respect to the large normal faults at Algerita, the structural Well 3502Key WellKey 2363 Rose Diagram Uni-Directional Rose Diagram Uni-Directional position of the localities can be divided into proximal Total Number of Points = 26 Total Number of Points = 48 Bucket Size=5degrees ErrorSize=0degrees Bucket Size = 5 degrees Error Size = 0 degrees (Algerita escarpment), intermediate (Wilson canyon) 0 6 0 5 and distal (outcrops at highway 83, near N N Alamogordo). The area surrounding the outcrops at highway 83 is affected by the presence of two large NE-trending lineaments that correspond to faults with some dextral strike-slip component (Russell, 1982). In spite of the proximity of these faults, and probably due to the fact that the slip along these faults is relatively small, the WellKey 4061 WellKey 8810 Rose Diagram Uni-Directional Rose Diagram Uni-Directional rocks at that outcrops display a lower degree of Total Number of Points = 85 Total Number of Points = 18 Bucket Size=5degrees ErrorSize=0degrees Bucket Size=5degrees ErrorSize=0degrees fracturing and deformation than those at the Algerita 0 6 0 3 escarpment. At Wilson canyon, the western border of the Figure 9. Fracture orientation at surface outcrops, Guadalupe Mountains is part of the Huapache cretaceous limestone in Yates. Well number is monocline, where the beds are gently dipping to the given for location reference. east due to the presence of a fault (the Huapache fault) in the subsurface. The structural position and The abutting relationship observed in the field the facies of the San Andres Formation in these areas shows that the WNW set is younger than the NE set

Stanford Rock Fracture Project Vol. 12, 2001 C-5 are somewhat similar to those of the reservoir in the found within the N 40° E set just at the intersections Yates field. with the NW set, suggesting again that the NE set is At each locality, but Wilson Canyon, scanlines older. Fracture density of the N 40° W set is also along beds of different lithologies were done in order higher than the N 40°E set, as is discussed below. to evaluate fracture geometry. Fracture orientation (Figure 12), spacing and height were recorded along every scanline. Bed thickness and lithology were recorded as well. A fracture set is defined as a group of fractures with approximately the same orientation. It is assumed that all the fractures of a given set are genetically related, even though this is not necessarily true since the presence of subordinate fractures like cross-joints or splay cracks may be misleading. The spacing for every fracture set has been calculated from the scanline data using simple trigonometric relations.

Figure 12. Filling and age relationship of the two main joint sets. Calcite is filling the NW set (parallel to the red pen). Joints of the NE set are closed and just locally filled with calcite at the intersections with the NW set. It indicates that the NE set is older.

N N

Figure 11. Geologic map showing the location of the outcrops of San Andres Formation where fracture densities were measured: 1) Algerita –Lawyer LawyerKey Canyon WilsonKey Canyon Rose Diagram Uni-Directional Rose Diagram Uni-Directional Total Number of Points = 61 Total Number of Points = 30 Canyon-; 2) Wilson Canyon; 3) Highway 83 Bucket Size = 5 degrees Error Size = 0 degrees Bucket Size = 5 degrees Error Size = 0 degrees

(Artesia-Alamogordo). 0 8 0 8

There are three main fracture sets in the area: N N N40°E, N40°W and N70°W. The age relationships of these sets are similar to those observed at Yates; the NE set is older than the NW sets. The N40° W set is younger than the N70°W, which is not well developed in this area, in comparison to Yates. A distinction between the two NW sets is complicated by the fact that the orientation of a given set may vary slightly due to differences in lithology. The N40°E set is HWY 83 -Alamogordo-Key St. 1 HWY 83 -Alamogordo-Key St. 2 Rose Diagram Uni-Directional Rose Diagram Uni-Directional Total Number of Points = 45 Total Number of Points = 100 locally affected by a dextral strike-slip shear Bucket Size=5degrees ErrorSize=0degrees Bucket Size = 5 degrees Error Size = 0 degrees component Locally there are other sets of fractures 0 8 0 18 and small faults, such as those associated with second order thrust faults that were found close to a larger feature with incipient strike-slip displacement. Figure 13. Fracture orientation in the San Andres Formation at the different outcrops visited in the There is a conspicuous difference between Guadalupe Mountains. aperture of the NE and the NW sets (figure 13). Joints of the NW sets have apertures of about 1 cm, Within a given stratigraphic interval, joints occur while apertures of the N 40 E set are hairline to 1 at different scales. Some joints are large and cut mm. White calcite is filling the NW set, and it is through all the layers of the interval, whereas other

Stanford Rock Fracture Project Vol. 12, 2001 C-6 are smaller and confined to just to one or two single πh3 beds. The spacing increases accordingly to the height Φf = αJ (4), where J = 2 (5); of the joints. The spacing of confined joints is in the 6ST order of 1 m to 3 m, whereas the spacing of larger the crack density parameter can be also expressed in joints is in the order of 10 to 30 m. terms of the ratio h/T as: 1 h3 Fracture density from outcrop data ε = 2 (6); therefore J ≈ 4ε (7). In geological and engineering literature different 8 ST techniques and terms are used to quantify the distribution of fracture intensity within a rock body. 0.30 Measurements of fracture intensity are made either in Mudstone one or two dimensions (Dershowitz and Herda, Wackestone 0.20 1992), and they cannot be directly extrapolated as Packstone three-dimensional quantities. In sedimentary rocks, the distribution of fracture intensity is frequently 0.10 described in terms of the fracture-spacing index density Crack (Narr, 1991). Fracture index (I) is defined as the ratio of bed thickness (T) to average spacing (S). The 0.00 relationship between fracture spacing and thickness 0123 has been studied by and Wu and Pollard (1995) and Location Bai and Pollard (2000). Spacing can be calculated either from the Figure 14. Crack density parameter (ε) at the different scanline method or the area method (Wu and Pollard, localities, split by rock type. 1=Algerita 1995; Younes, 1996). According to these authors, the escarpment, 2= outcrops at highway 83. results from the scanline method will under-estimate the fracture density unless the fracture set is well developed, but area method is limited to pavement 1.50 exposures. In order to evaluate the fracture density Mudstone from the 2D outcrops, in addition to the fracture index Wackestone we measured the average fracture height. For those 1.00 Packstone cases when the fracture height is larger than the bed under study, a height equal to the bed thickness has been assigned. 0.50 In addition to the fracture index (I) we have Specific Joint Density Joint Specific calculated other two parameters to estimate fracture 0.00 density. The crack density parameter (ε), established 0.00 1.00 2.00 3.00 by Hudson (1981) for penny-shape cracks, which is Location expressed as: Na3 Figure 15. Specific joint density (SJD) by locality and ε = (1) split by facies. As in the previous figure, V 1=Algerita escarpment and 2= outcrops at where N is the number of fractures in the volume V, highway 83. and 2a is the crack length. If we assume T = 2a and V = T3, then: The crack density parameter (figure 14) and the T I specific joint density (figure 15) give higher values of N = (2); and ε = (3); relative fracture density for the outcrops located near S 8 the large normal fault (Lawyer Canyon at Algerita the other parameter is the specific joint density (J), a Escarpment). Both parameters also indicate that new quantity. The assumption that T=2a implies that fracture intensity is higher in fine-grained carbonates fracture height is equal to bed thickness for all the (mudstones) than in coarse-grained (wackestones). fractures, but that was not always the case in the The same plot for the fracture index does not show rocks we studied. In order to incorporate the any apparent trend (Figure 16). Since the obtained variability of fracture height (h) in the estimation of results from Figures 14 and 15 are consistent with the the fracture density, we derived the expression to field observations, it is considered that both the crack calculate the fracture porosity (φf) for penny-shaped density parameter and the specific joint density are fractures of height h and aspect ratio more appropriate to quantify fracture density than the fracture spacing index. α (aperture/crack height):

Stanford Rock Fracture Project Vol. 12, 2001 C-7

carbonates, as well as fracture distribution in the rocks. 5.00 Mudstone 4.00 Wackestone densely cemented interparticle porosity 7 Packstone intraframe porosity microporosity 3.00 moldic porosity well log data 6

2.00 Hashin−Shtrikman Bounds 1.00 5 Vp [Km/s] Fracture-Spacing Index 0.00 4 0123Location spherical pores 3 Figure 16. Fracture Spacing Index (T/S) by locality and split by facies. Locality numbers are the 2 crack−like pores same that the previous figures. Note that in some 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 cases the fracture index does not show the Porosity relationship between facies and fracture density Figure 17. Vp - porosity scatter plot for one well observed in figures 14 and 15. (black), with superimposed data from other dolomites from literature (gray). Different curves represent theoretical predictions for different pore Rock Physics Analysis of Well Logs types using self –consistent effective medium and Core Data model. The elastic properties of the reservoir can be investigated using sonic logs. Figure 18 shows the P Three main depositional environments have been wave velocity – porosity scatter plot for San Andres identified for the San Andres formation in this well: formation derived from well logs. The velocities shoal, shoal-margin and low energy subtidal. We correspond to the brine saturated state. On the same infer than they can be traduced in terms of lithofacies figure we superimposed data from other carbonates, as dominantly mud-supported (subtidal), mixed whose pore types are known (Anselmetti and Eberli, (shoal-margin) and dominantly grain-supported 1977). We also superimposed theoretical predictions (shoal). Figure 20 illustrates that mudstones for different pore types, calculated using the self- predominate in the subtidal environment and consistent effective medium model (Berryman, 1980). packstones are dominant in the shoal environment. Different curves correspond to different pore shapes: from crack like pores (aspect ratio=0.1), to spherical 7 Castagna relation pores (aspect ratio=1). We conclude that there is a 6.5 Vs=−0.05508 Vp2+1.0168Vp−1.0305 [Km/s] significant variability in pore shape within the San 6 Andres Formation. However, we can distinguish two main clusters, one corresponding mainly to 5.5 microporosity (lower cluster) and the other one to 5 intraframe and moldic porosity (upper cluster). Pickett relation

Vp [Km/s] 4.5 Figure 19 presents the Vp - Vs scatter plot for Vs=Vp/1.9 the San Andres reservoir for one of the wells. For 4 comparison, we superimposed the empirical Castagna 3.5

(1993) relation for limestones, as well as Pickett 3 (1963) line for limestones. We can see that the two 2.5 published relations fit the data reasonably well. This 1.5 2 2.5 3 3.5 4 Vs [Km/s] suggests that the dolomitization process of the Figure 18. Vp - Vs scatter plot for the data in one well. original limestone is not complete, and also that the Superimposed are the Castagna (1993) and calcite reprecipitation might be significant. Pickett (1963) empirical predictions for Next we use the information about the limestones. depositional environments, rock texture and fracture distribution available from core data from one of the Figure 20 also shows that mud-supported wells. We combine this information with the sonic log carbonates are stiffer than grain-supported rocks due from the same well, in order to determine possible to the lower porosity of the former. These conclusions relations between elastic properties and lithofacies, are in agreement with the outcrop observations and derived from the depositional textures of the the results obtained from fracture-density analysis of core data (Table 1). We can conclude that the

Stanford Rock Fracture Project Vol. 12, 2001 C-8 different lithofacies, derived from original the data obtained from the Algerita escarpment and depositional textures, have different elastic the exposures along the highway 83 shows that properties. Fracture toughness depends on the elastic fracture density is higher nearby fault zones. Fracture properties of the rock, therefore the observed clustering and high-productivity trends identified at relationship between fracture density and lithofacies Yates (Tinker and Mruk, 1995) may be related to the should be always expected. Another consequence of fractured damage zone near to one of these small this observation is that these different facies might be faults. seismically distinguishable, which could be an Our results indicate that structural position and indirect method to predict fracture-occurrence. lithofacies are the two main factors controlling fracture-occurrence. For this particular case, the 6500 proximity to the faults rather than the fold curvature BLUE − Subtidal environment GREEN− Shoal margin environment controls fracture density. On the other hand, different RED − Shoal environment 6000 lithofacies, derived from the original depositional SQUARE − Mudstone TRIANGLE −Wackstone textures, give rise to diverse elastic properties. In this CIRCLE − Packstone 5500 case it was found that mud-supported rocks are stiffer than grain-supported carbonates due mainly to the 5000 lower porosity of the former. Vp [m/s] Spacing between the faults varies from 200 to 4500 500 meters. The direct observation of one of these normal faults shows that the core zone is filled with 4000 cement, and may therefore constitute a barrier for fluid flow, but the damage zone has the higher density

3500 0 5 10 15 20 25 30 35 of open fractures. If those open fractures are not Porosity completely cemented, they will constitute important Figure 19. Vp - porosity scatter plot from core end log fluid conduits parallel to the fault zone. The data, which emphasizes the different depositional orientation of the maximum horizontal stress (WNW) environments and facies. may also help to increase the fracture permeability

along the WNW-trending set. The result of this would Table 1: Fracture distribution with respect to the be zones of high permeability separated by a fluid- depositional textures and lithofacies. Depositional Shoal Shoal Subtidal flow barrier. environment margin The reservoir model can be improved by a more Depositional Grain- Mixed Mud- detailed analysis of the relationship between fracture texture supported supported density and the other factors controlling its Dominant Packstone Wackestone Mudstone distribution like faulting and lithology. A more lithofacies detailed study of fault architecture may also be Fracture necessary, since the permeability along and across a distribution 16% 34% 50% fault may vary according to the slip distribution and fault segmentation.

Conceptual Model of Faults and Conclusions The dominant orientations of faults and fractures Fractures at Yates at Yates are WNW (290°) and NE (040°). Field Based on the data available and the field observations indicate that Yates was affected by a observations, a preliminary conceptual model is post-cretaceous normal-fault stress regime that proposed for the Yates field. First of all, field generate a series of systematic, relatively small, observations show that Yates area has been under a normal faults. It is proposed that these faults are normal-fault stress regime at least since the early controlling the distribution of fractures in the field Tertiary. Although the normal faults observed at and therefore the fluid flow in the reservoir. surface do not necessarily extend to the Permian Scanlines measures on the San Andres outcrops rocks at subsurface, enough evidence has been seen to at different localities around the Guadalupe postulate that similar faults can be present in the Mountains provide a quantitative estimation of reservoir. These faults may strongly control the fracture density within this unit. For a particular set, movement of fluids within the reservoir. two main factors have been found to control fracture The observed faults are relatively small. Fault set density: structural position in respect to the faults and length varies from 200 m to about 5 km (figure 21) lithofacies. and slip ranges from 2 to 25 meters. These types of Three main sets of fractures were found in the faults can hardly be identified even with a dense Guadalupe Mountains area: NW (320°), WNW network of wells like that of Yates. A comparison of

Stanford Rock Fracture Project Vol. 12, 2001 C-9

(290°) and NE (040°). Enough data to compare Hudson, J. A., 1981, Wave speeds and attenuation of elastic fracture density between different localities was waves in material containing cracks: Geophys. J. Roy. collected for the NW and NE sets. This data shows Astr. Soc., 64, p. 133-150. that fracture density of the NW set is higher than the Kerans, C., F. J. Lucia, and R.K. Senger, 1994, Integrated NE set. characterization of carbonate ramp reservoirs using Permian San Andres Formation outcrop analogs: American Association of Geologists Bulletin, 78, p. 181-216. Narr, W., 1991, Fracture density in deep subsurface: Techniques with application to Point Arguello oil field. American Association of Petroleum Geologists Bulletin, 75, p. 1300-1323. Pickett, G.R., 1963, Acoustic character logs and their applications in formation evaluation, J. Petr. Tech., 15, 650-667. Renfro, H.B. D. Feray, R. Dott and A. Bennison, 1984, Geological highway map of Texas. American Figure 20. Distribution of normal faults in the area just Association of Petroleum Geologists. north of Yates field. A similar distribution of Ruger,A., 1998, Variation of P-wave reflectivity with offset normal faults is proposed to exist in Yates at and azimuth in anisotropic media, Geophysics, 63, 935-947. subsurface. Russel, L.R., and S. Nelson, 1994, Structural style and Acknowledgements tectonic evolution of Albuquerque basin segment of We would like to acknowledge Stanford Rock the Rio Grande rift, New Mexico, USA. In: Landon, S., ed., Interior Rift Basins. American Association of Physics Project, and DOE, contract no. DE-AC26- Petroleum Geologists, Memoir 59. 99FT40692, Marathon Oil Co and Professors G. Russell, E., R. Kelley, F. Kottlowsky and J. Robertson, Mavko and A. Aydin. 1982, New Mexico highway geologic map. New Mexico Geological Society. References Sonnenfeld, M.D., 1991, High frequency cyclicity within Adams, J.E., 1940, Structural development, Yates area, shelf-margin and slope strata of the upper San Andres Texas. American Association of Petroleum Geologists sequence, Last Chance Canyon, in Meader-Roberts, Bulletin, 24, p. 134-142. Sally, M. Candelaria, and G.E. Moore, eds.: Sequence Anselmetti, F.S. and Eberly, G.P., Sonic velocity in stratigraphy, facies and reservoir geometries of the carbonate sediments and rocks, Carbonate San Andres, Grayburg and Queen formations, Seismology, ed. Palaz, I. and Marfurt, K.J., SEG. Guadalupe Mountains, New Mexico and Texas: Bai, T., and D.D. Pollard, 2000, Fracture spacing in layered Permian Basin section, Society of Economic rocks: a new explanation base on the stress transition, Paleontologists and Mineralogists Publication 91-32, Journal of Structural Geology, 29, p.43-57. p.11-51. Berryman, J.G., 1980, Long-wavelength propagation in Tinker, S., and D. Mruk, 1995, Reservoir characterization composite elastic media, I and II, J. Acoustic. Soc. of a Permian giant: Yates field, west Texas. Internal Am., 68, 1809-1831. report: Marathon Oil Company. Castagna, J.P., Batzle, M.L., and Kan,T.K., 1993, Rock- Tinker, S.W., J.R. Ehrets and M.D. Brondos, 1995, Physics- the link between the rock properties and Multiple karst events related to stratigraphic cyclicity: AVO response, Offset-Dependent Reflectivity-Theory San Andres Formation, Yates field, west Texas. In: and Practice of AVO Analysis, eds. Castagna J.P. and Budd, A., H.A. Saller and P.M. Harris eds., Backus, M.M., SEG, Tulsa. Unconformities and porosity in carbonate strata. Craig, D.H., 1988, Caves and other features of Permian Am.Ass Pet.Geol. Memoir No. 63, p. 213-237. karst in San Andres dolomite, Yates field reservoir, Thomsen, L., 1986, Weak elastic anisotropy, Geophysics, west Texas, in Choquette, P.W, and N.P., James eds, 51, 1954-1966. Paleokarst: New York, Springer-Verlag, p. 342-363. Wu, A., and D. Pollard, 1995, An experimental study of the Dershowitz, W., and H. Herda, 1992, Interpretation of relationships between joint spacing and layer fracture spacing and intensity. Rock Mechanics and thickness. Journal of Structural Geology, 17, No. 6, p. Rock Engineering, 25, p. 757-766. 887-905. Donogue, D, and W.L. Gupton, 1956, Yates field, Pecos Younes, A.I., 1996, Fracture distribution in faulted and Crockett counties, Texas. In: Herald, F.A., editor: basement blocks, Gulf of Suez, Egypt: Reservoir Occurrence of oil and gas in western Texas. Bureau of characterization and tectonic implications. Ph. D. Economic Geology, University of Texas at Austin, p. Dissertation, The Pennsylvania State University. 169 426-433. pp. Hennen, R.V., and Metcalf, R.J., 1929, Yates oil field, Zoback, M., and M. L. Zoback, 1991, Tectonic stress field Pecos County, Texas. American Association of of North America and relative plate motions, in Petroleum Geologist Bulletin, 13, pp 1509-1556. Simmons, D.B., Engdahl, E.R., Zoback, M.D., and Blackwell, D.D., eds, Neotectonics of North America:

Stanford Rock Fracture Project Vol. 12, 2001 C-10

Boulder Colorado, Geological Society of America, Decade Map Volume 1.

Stanford Rock Fracture Project Vol. 12, 2001 C-11