Rock-Fabric/Petrophysical Classification of Carbonate Pore Space for Reservoir Characterization 1 F. Jerry Lucia 2
ABSTRACT This paper defines the important geologic parameters that can be described and mapped to allow accurate petrophysical quantification of carbonate geologic models. All pore space is divided into interparticle (intergrain and intercrystal) and vuggy pores. In nonvuggy carbonate rocks, permeability and capillary properties can be described in terms of particle size, sorting, and interparticle porosity (total porosity minus vuggy porosity). Particle size and sorting in limestones can be described using a modified Dunham approach, classifying packstone as grain dominated or mud dominated, depending on the presence or absence of intergrain pore space. To describe particle size and sorting in dolostones, dolomite crystal size must be added to the modified Dunham terminolo gy. gy. Larger dolomite crystal size improves petrophysical properties in mud-dominated fabrics, whereas variations in dolomite crystal size have little effect on the petrophysical properties of grain-dominated fabrics. A description of vuggy pore space that relates to petrophysical properties must be added to the description of interparticle pore space to complete the petrophysical characterization. Vuggy pore space is divided into separate vugs and touching vugs on the basis basi s of vug interconnection. interco nnection. Separate vugs vug s are a re fabri fab ric c sele s ele ctive cti ve and are con nec ted only on ly through the interparticle pore network. Separate vug porosity po rosity contributes contr ibutes little to permea pe rmeability bility and
©Copyright 1995. The American Association of Petroleum Geologists. All rights reserved. 1Manuscript received June 27, 1994; revised manuscript received March 21, 1995; final acceptance May 1, 1995. 2Bureau of Economic Geology, University of Texas at Austin, Austin, Texas 78713-7508. This classification of carbonate pore space is the result of research initiated while I was with Shell Oil Company and completed at the Bureau of Economic Geology’s Reservoir Characterization Research Laboratory, which is funded by industry sponsors Agip, Amoco, ARCO, British Petroleum, Chevron, Conoco, Exxon, Fina, JNOC, Marathon, Mobil, Phillips, Shell, Texaco, Total, and Unocal. I am grateful for intense discussions with colleagues Charles Kerans, Gary Vander Stoep, Fred Wang, Susan Hovorka, Mark Holtz, Steve Ruppel, and Don Bebout. An in-depth review of the manuscript by Philip W. Choquette was especially useful. Published with permission of the director, Bureau of Economic Geology, University of Texas at Austin.
AAPG Bulletin, V. 79, No. 9 (September 1995), P. 1275–1300.
should be subtracted from total porosity to obtain interparticle porosity for permeability estimation. Separate-vug pore space is generally considered to be hydrocarbon filled in reservoirs; however, intragranular microporosity is composed of small pore sizes and may contain capillary-held connate water within withi n the th e reservo res ervoir. ir. Touching vugs v ugs are a re nonfabr no nfabric ic selective and form an interconnected pore system independent of the interparticle system.
INTRODUCTION The goal of reservoir characterization is to describe the spatial distribution of petrophysical parameters, such as porosity, permeability, and saturation. Wireline logs, core analyses, production data, pressure-buildup data, and tracer tests pro vide vid e quan q uantit titati ative ve meas m easurem urement entss of o f petro p etro physica phys icall parameters in the vicinity of the well bore. These well bore data must be b e integrat i ntegrated ed with w ith a geologic geol ogic model to display the petrophysical properties in three-dimensional space. Studies that relate rock fabric to pore-size distribution, and thus to petrophysical properties, are key to quantifying geologic models in numerical terms for input into computer simulators (Figure 1). Geologic models are generally based on observations interpreted in terms of depositional environments and sequences. In the subsurface, cores and wireline logs are a re the main source of o f data dat a for these interpretations. interpretations. Engineering models are based on wireline wirelin e log calculations calculat ions and average rock ro ck properpro perties from core analyses. Numerical engineering data and interpretive geologic geologic data are joined at the rock-fabric level because the pore structure is fundamental to petrophysical properties and the pore structure is the result of spatially distributed depositional and diagenetic processes. The purpose of this paper is to define the important geologic parameters that can be described and mapped to allow accurate petrophysical quantification of carbonate geologic models by (1) describing the relationship between carbonate rock fabrics and petrophysical properties, (2) presenting a generic petrophysical classification of carbonate 1275
1276 276
Class assification of Carbonate Pore ore Spa Space
Table Table 1. PorePore-typ type e Termino Terminology logy and Abbre Abbreviat viations ions Used in This Paper Compared Compare d to Abbreviati Abbre viations ons Used in Lucia (1983) and Choquette and Pray (1970)
A RI C O CK F B R
R
Sedimentation
Wireline Logs
Term
Abbreviations Choquette and Lucia (1983) Pray (1970)
Core Analysis
Interparticle Intergrain Intercrystal Vug Pressure Separate vug Moldic Tecto ectonics nics Tracer Tracer Tests Tests Tests Intraparticle Intragrain Intracrystal Figure 1—Integration of spatial geologic data with Intrafossil numerical engineering data through rock-fabric studies. Intragrain microporosity Shelter Touching Vug pore space, and (3) suggesting that rock-fabric Fracture units are systematically stacked within high-freSolution-enlarged fracture quency cycles. Cavernous Breccia Fenestral Diagenesis
Porosity Permeability Saturation
PORE-SPACE PORE-SPACE TERMINOLOGY AND CLASSIFICATION
Production
IP IG IX VUG SV MO WP WG WX WF
BP – BC VUG – MO WP – – –
µG SH TV FR
– SH – FR
SF CV BR FE
CH* CV BR FE
*Channel.
Pore space must be defined and classified i n terms of rock fabrics and petrophysical properties fabrics to petrophysical rock properties in carbonto integrate geological and engineering information. ate rocks. The Archie classification focuses on esti Archie (1952) made the first attempt attempt at relating relating rock mating porosity, but is also useful for approximating permeability and capillary properties. Archie (1952) recognized that not all pore space can be observed using a 10-power microscope and that the surface PORE TYPES texture of the broken rock reflects the amount of matrix porosity. Therefore, Therefore, pore space is divided CAVERNOUS INTERGRAIN MOLDIC FRACTURE into matrix and visible porosity (Figure 2). Chalky INTERCRYSTAL INTRAFOSSIL SHELTER SOLUTION-ENLARGED texture indicates a matrix porosity of about 15%, FRACTURE sucrosic texture texture indicates a matrix porosity of about 7%, and compact texture indicates matrix porosity ARCHIE (1952) of about 2%. Visible pore space is described according to pore size, and ranges from no visible pore MATRIX space to larger than cutting size. Porosity and perVISIBLE (A,B,C, and D) meability trends and capillary pressure characteristics are also related to these textures. LUCIA (1983) Although Although the Archie Archie method is still useful useful for estiestiVUGGY INTERPARTICLE mating petrophysical properties, relating these SEPARATE TOUCHING descriptions to geologic models is difficult because the descriptions cannot be defined in depositional CHOQUETTE and PRAY (1970) or diagenetic terms. A principal difficulty is that no provision is made for distinguishing between visible NONFABRIC FABRIC SELECTIVE SELECTIVE interparticle pore space and other types of visible pore space, such as moldic pores. Research on carFigure 2—Petrophysical classification of carbonate pore bonate pore space (Murray, (Murray, 1960; Choquette and types used in this paper (Lucia, 1983) compared with Pray, 1970; Lucia, 1983) has shown the importance Arch ie’ s or iginal igi nal clas sifica sif icatio tion n (1952) (1 952) and the fabricfab ricof relating pore space to depositional and diagenetic selectivity concept of Choquette and Pray (1970). A = no fabrics and to distinguishing between interparticle visibl vis ibl e po re space s pace , an d B, C, a nd D = in crea sing sin g po re (intergrain and intercrystal) and other types of pore sizes from pinpoint to larger than cutting size.
Lucia
1277
Figure 3—Geological and petrophysical classification of carbonate interparticle pore space based on size and sorting of grains and crystals. The volume of i nterparticle pore space is important for characterizing the petrophysical properties.
space. Recognition of the importance of these factors prompted modification of Archie’s classification. The petrophysical classification of carbonate porosity presented by Lucia (1983) emphasized petrophysical aspects of carbonate pore space, as does the Archie classification . However, by comparing rock-fabric descriptions with laboratory measurements of porosity, permeability, capillarity, and Archie m values, Lucia (1983) showed that the most useful division of pore types for petrophysical purposes was of pore space between grains or crystals, called interparticle porosity, and all other pore space, called vuggy porosity (Figure 2). Vuggy pore space was further subdivided by Lucia (1983) into two groups depending on how the vugs are interconnected: (1) vugs interconnected only through the interparticle pore network are separate vugs, and (2) vugs that form an interconnected pore system are touching vugs.
Choquette and Pray (1970) discussed the geologic concepts surrounding carbonate pore space and presented a classification that is widely used. They emphasized the importance of pore-space genesis and used genetic, not petrophysical, divisions in their classification. They divided all carbonate pore space into two classes: fabric selective and nonfabric selective (Figure 2). Moldic and intraparticle pore types were classified as fabric-selective porosity by Choquette and Pray (1970) and grouped with interparticle and intercrystal porosity. However, Lucia (1983) demonstrated that moldic and intraparticle pores have a different effect on petrophysical properties than do interparticle and intercrystal pores and thus should be grouped separately. Poretype terms used in Lucia’s (1983) classification are listed in Table 1, which compares his terms with those suggested by Choquette and Pray (1970). Although most of the terms defined by Choquette
1278
Classification of Carbonate Pore Space
180 ) a i s p160 ( e r u s s140 e r p t n120 e m e c a100 l p s i d d 80 e t a l o p 60 a r t x e r 40 i a / y r u 20 c r e M 0
22 21 Percent porosity
19
10 14 15 27 27 26 22 22 12 21 16
16
18
0
20
40
18
12
60
18
80
18
100
16 21
120
140
Figure 5—Porosity-air permeability relationship for various particle-size groups in nonvuggy carbonate rocks (Lucia, 1983).
Average particle size (µm)
Figure 4—Relationship between mercury displacement pressure and average particle size for nonvuggy carbonate rocks with greater than 0.1 md permeability (Lucia, 1983). The displacement pressure is determined by extrapolating the data to a mercury saturation of zero.
pores, but it is consistent with the Archie terminology and with the widespread, less restrictive use in the oil industry of the term “vuggy porosity” in referring to visible pore space in carbonate rocks.
and Pray are also used here, interparticle and vug porosity have different definitions. Lucia (1983) demonstrated that pore space located both between grains (intergrain porosity) and between crystals (intercrystal porosity) are petrophysically similar. A term identifies these petrophysically similar pore types: interparticle. “Interparticle” was selected because of its broad connotation. The classification of Choquette and Pray (1970) does not have a term that encompasses these two petrophysically similar pore types. In their classification, the term “interparticle” is used instead of “intergrain.” [See Choquette and Pray (1970, page 247) for their justification for using “interparticle” vs. “intergrain.”] Vuggy p orosi ty, as d efi ned by Luc ia (19 83) , is pore space that is within grains or crystals or that is significantly larger than grains or crystals; that is, pore space that is not interparticle. Vugs are commonly present as leached grains, fossil chambers, fractures, and large, irregular cavities. Fracture porosity is categorized as a type of vuggy porosity to be inclusive. This definition deviates from the restrictive definition of vugs used by Choquette and Pray (1970) as nondescript, nonfabric-selective
ROCK-FABRIC/PETROPHYSICAL CLASSIFICATION The foundation of the Lucia (1983) and the Archie classifications is the concept that pore-size distribution controls permeability and saturation and that pore-size distribution is related to rock fabric. To relate carbonate rock fabrics to pore-size distribution, one must determine if the pore space belongs to one of the three major pore types: interparticle, separate vug, or touching vug (Figure 2). Each class has a different type of pore distribution and interconnection. One must also determine the vo lum e of po re sp ace in th es e va riou s cl ass es because pore volume relates to reservoir volume and, in the case of interparticle and separate-vug porosity, to pore-size distribution.
Petrophysics of Interparticle Pore Space In the absence of vuggy porosity, pore-size distribution in carbonate rocks can be described by particle size, sorting, and interparticle porosity (Figure 3). Lucia (1983) showed that particle size can be
Lucia
1279
B
A 0
0.5 mm
1.0
0
0.2 mm
0.4
0
0.5 mm
1.0
D
C 0
1.0 mm
2.0
Figure 6—Photomicrographs showing examples of nonvuggy limestone rock fabrics. (A) Grainstone, φ = 25%, k = 1500 md. (B) Grain-dominated packstone, φ = 16%, k = 5.2 md. Note intergranular cement and pore space. (C) Muddominated packstone, φ = 18%, k = 4 md. Note microporosity. (D) Wackestone.
related to mercury capillary displacement pressure in nonvuggy carbonates having more than 0.1 md permeability. Because displacement pressure is a function of the larger, well-connected pores, the particle size describes the size of the larger pores (Figure 4). Whereas the displacement pressure characterizes the larger pores sizes and is largely independent of porosity, the shape of the capillary pressure curve characterizes the smaller pore sizes and is dependent on interparticle porosity (Lucia, 1983). The relationship between displacement pressure and particle size (Figure 4) is hyperbolic and suggests important particle-size boundaries at 100 and at 20 µm. Lucia (1983) demonstrated that three
porosity-permeability fields can be defined using particle-size boundaries of 100 and 20 µm, a relationship that appears to be limited to particle sizes less than 500 µm (Figure 5). These three fields will be referred to as the greater than 100-µm permeability field, 100–20-µm permeability field, and less than 20-µm permeability field. Recent work has shown that permeability fields can be better described in geologic terms if sorting and particle size are considered. The approach to size and sorting used herein is similar to the grainsupport/mud-support principle upon which Dunham’s (1962) classification is built. Dunham’s classification, however, focused on depositional
1280
Classification of Carbonate Pore Space
(B)
(A) 1000
1000
) 100 d m ( y t i l i 10 b a e m r e P 1
) d 100 m ( y t i l i b 10 a e m r e 1 P
0.1
0.1 10
20
30
40
10
Interparticle porosity (%)
(C)
20
30
40
30
40
Interparticle porosity (%)
(D)
1000
1000
) d 100 m ( y t i l i 10 b a e m r e P 1
) d 100 m ( y t i l i b 10 a e m r e 1 P
0.1
0.1 10
20
30
40
Interparticle porosity (%)
10
20
Interparticle porosity (%)
Figure 7—Porosity-air permeability crossplots for nonvuggy limestone rock fabrics compared with the three permeability fields illustrated in Figure 5 (A) 400-µm ooid grainstone, Ste. Genevieve, Illinois (Choquette and Steiner, 1985). Low-permeability, high-porosity data are deleted because they are from oomoldic and wackestone rock fabrics (P. W. Choquette, 1993, personal communication). (B) Grain-dominated packstone data, Wolfcamp, west Texas (Lucia and Conti, 1987), a poorly sorted mixture of 80–300-µm grains and micrite. (C) Wackestones with microporosity between 5-µm crystals, Shuaiba, United Arab Emirates (Moshier et al., 1988). Data are associated with stylolites not shown. (D) Coccolith chalk, Cretaceous (Scholle, 1977). The presence of intragranular pore space in the coccoliths causes the data to plot below the less than 20-µm field.
texture, whereas petrophysical classifications are focused on contemporary rock fabrics that include depositional and diagenetic textures. Thus, minor modifications must be made in Dunham’s classification before it can be applied to a petrophysical classification. Instead of dividing fabrics into grain supported and mud supported as in Dunham’s classification, fabrics are divided into grain dominated and mud dominated (Figure 3). The important attributes of grain-dominated fabrics are the presence of open or occluded intergrain porosity and a grain-supported texture. The important attribute of mud-dominated fabrics is that the areas between the grains
are filled with mud, even if the grains appear to form a supporting framework. Grainstone is clearly a grain-dominated fabric, but Dunham’s (1962) packstone class bridges a boundary between large intergrain pores in grainstone and small interparticle pores in wackestones and mudstones. Some packstones have intergrain pore space, and some have intergrain spaces filled with mud; therefore, the packstone textural class must be divided into two rock-fabric classes: graindominated packstones that have intergrain pore space or cement, and mud-dominated packstones that have intergrain spaces filled with mud (Figure 3). The pore-size distribution is controlled by the
Lucia
1000
) 100 d m ( y t i l i 10 b a e m r e P 1
0.1 10
20
30
40
Interparticle porosity (%) Ooid grainstone Grain-dominated packstones Mud-dominated fabrics Figure 8—Composite porosity-air permeability crossplot for nonvuggy limestone fabrics compared with the three permeability fields illustrated in Figure 5. Chalks are not included because of the presence of intragrain pore space.
grain size in grain-dominated packstones and by the mud size in mud-dominated packstones.
Permeability/Rock-Fabric Relationships Limestone Rock Fabrics
Examples of nonvuggy limestone petrophysical rock fabrics are illustrated in Figure 6. In grainstone fabrics, the pore-size distribution is controlled by grain size; in mud-dominated fabrics, the size of the micrite particles controls the pore-size distribution. In grain-dominated packstones, however, the poresize distribution is controlled by grain size and by the size of micrite particles between grains. Figure 7A is a crossplot between permeability and intergrain porosity for grainstones. The data are from Choquette and Steiner’s (1985) work on the Ste. Genevieve oolite (Mississippian). The average grain size of the oolite is about 400 µm. The points on this graph are concentrated within the greater than 100-µm permeability field. Figure 7C is a crossplot between air permeability and interparticle porosity from a Middle Eastern mud-dominated fabric containing microporosity. The data from this graph are abstracted from Moshier et al. (1988). The average crystal size of the mud matrix is about 5 µm (Moshier et al., 1988). The data are concentrated in the less than 20-µm permeability field. Because grain-dominated packstone is a new fabric class, data are difficult to find. Lucia and Conti
1281
(1987) reported on a nonvuggy, grain-dominated Wolfcampian packstone found i n a core taken i n Oldam County, Texas. The grain-dominated packstone is described as a poorly sorted mixture of 150–300-µm grains in a matrix composed of 80-µm pellets and 10-µm calcite crystals. A porosity-permeability crossplot of these data (Figure 7B) shows that it plots on the boundary between the 100–20µm and the less than 20-µm permeability fields. Additional data points have been gleaned from the literature, and they plot in the 100–20-µm permeability field. Figure 7D is a crossplot between permeability and porosity for North Sea coccolith chalk (Scholle, 1977). The average size of coccoliths is about 1 µm. The data points plot below the less than 20-µm permeability field. The presence of intrafossil pore space within the coccolith grains probably accounts for the lower-than-expected permeability in the high-porosity ranges. Figure 8 illustrates all the data for limestones compared with permeability fields. Although the data have considerable scatter, grainstone, graindominated packstone, and mud-dominated fabrics are reasonably well constrained to the three permeability fields. Whereas grain size and sorting define the permeability fields, the interparticle porosity defines the permeability within the field. Systematic changes in intergrain porosity by cementation, compaction, and dissolution will produce systematic changes in the pore-size distribution. Therefore, the permeability field is defined by interparticle porosity, grain size, and sorting. Dolomite Rock Fabrics
Examples of nonvuggy dolomite petrophysical rock fabrics are illustrated in Figure 9. Dolomitization can change the rock fabric significantly. In limestones, the grain and mud fabrics usually can be distinguished with little difficulty. If the rock has been dolomitized, ho wever, the overprint of dolomite crystals commonly obscures the grain and mud fabric. Grain and mud fabrics in fine crystalline dolostones are easily recognizable; however, as the crystal size increases, the precursor fabrics become more difficult to determine. Dolomite crystals (defined as particles in this classification) commonly range in size from several microns to more than 200 µm. Micrite particles are usually less than 20 µm in size. Dolomitization of a mud-dominated carbonate fabric therefore can result in an increase in particle size from less than 20 to more than 100 µm (Figure 9). The crossplot of interparticle porosity and permeability (Figure 10A) illustrates the principle that in mud-dominated fabrics, permeability increases as dolomite crystal size increases. Fine crystalline (average 15 µm)
1282
Classification of Carbonate Pore Space
A
D 0
0.2
0
0.5 mm
1.0
0
0.5 mm
1.0
mm
B
E 0
0.5 mm
1.0
Figure 9—Examples of nonvuggy dolomite fabrics. (A) Dolograinstone, 15-µm dolomite crystal size, φ = 16.4%, k = 343 md, Dune field (Bebout et al., 1987). (B) Dolograinstone, 30-µm dolomite crystal size, φ = 7.1%, k = 7.3 md, Seminole San Andres unit, west Texas. (C) Dolograinstone, crystal size 400 µm, φ = 10.2%, k = 63 md, Harmatton field, Alberta, Canada. (D) Grain-dominated dolopackstone, 10-µm dolomite crystal size, φ = 9%, k = 1 md, Farmer field, west Texas. (E) Grain-dominated dolopackstone, 30-µm dolomite crystal size, φ = 9.5%, k = 1.9 md, Seminole San Andres unit, west Texas. (F) Crystalline dolowackestone, 10-µm dolomite crystal size, φ = 11%, k = 0.12 md, Devonian, North Dakota. (G) Crystalline dolowackestone, 80-µm dolomite crystal size, φ =16%, k = 30 md, Devonian, North Dakota. (H) Crystalline dolowackestone, 150-µm dolomite crystal size, φ = 20%, k = 4000 md, Andrews South Devonian field, west Texas.
C 0
1.0 mm
2.0
Lucia
1283
Figure 9—Continued.
F 0
0.5 mm
0
0.5 mm
G
H 0
0.5 mm
mud-dominated dolostones from Farmer and Taylor-Link (Lucia et al., 1992b) fields in the Permian basin and from Lawrence (Bridgeport) field in the Illinois basin (Choquette and Steiner, 1985) plot within the less than 20-µm permeability field. Medium crystalline (average 50 µm) muddominated dolostones from the Dune field, Permian basin (Bebout et al., 1987), plot within the 100–20-µm permeability field. Large cr ystalline (average 150 µm), mud-dominated dolostones from An dr ew s Sou th De vo nia n fi el d, Permi an ba si n (Lucia, 1962), plot in the greater than 100-µm permeability field. 1.0 Grainstones are commonly composed of grains much larger than the dolomite crystal size (Figure 9), so that dolomitization does not have a significant effect on the pore-size distribution. This principle is illustrated in Figure 10B, a crossplot of interparticle porosity and permeability measurements from dolomitized grainstones. The grain size of the dolograinstones is 200 µm. The fine crystalline dolograinstone from Taylor-Link field, Permian basin, the medium crystalline dolograinstone from Dune field, Permian basin, and the large crystalline dolograinstone from an outcrop on the Algerita escarpment, New Mexico, all plot within the gr eater than 100-µm permeability field. The large cr ystalline mud-dominated fabrics also plot in this permeability field, indicating that they are petrophysically similar to grainstones (Figure 10A). Interparticle porosity and permeability measurements from fine to medium crystalline, grain-domi1.0 nated dolopackstones are crossplotted in Figure 10C. The samples are from the Seminole San Andres unit and Dune (Grayburg) field (Bebout et al., 1987), Permian basin. The data plot in the 100–20-µm permeability field. The medium crystalline mud-dominated dolostones also plot in this field (Figure 10A). Figure 11 illustrates all dolomite data compared with permeability fields. Dolograinstones and large crystal dolostones constitute the greater than 100µm permeability field. Grains are very difficult to recognize in dolostones with greater than 100-µm crystal size. However, because all large crystalline dolomites and all grainstones are petrophysically similar, whether the crystal size or the grain size is described makes little dif ference petrophysically. Fine and medium crystalline grain-dominated dolopackstones and medium crystalline mud-dominated dolostones constitute the 100–20-µm perme1.0 ability field. Fine crystalline mud-dominated dolostones constitute the less than 20-µm field.
1284
Classification of Carbonate Pore Space
(B)
(A) 1000
) 100 d m ( y t i l i 10 b a e m r e P 1
1000
Fine crystal size Medium crystal size Large crystal size
) d 100 m ( y t i l i b 10 a e m r e 1 P
Mud-dominated dolostones
0.1
0.1 10
20
30
40
Interparticle porosity (%)
(C) 1000
) d 100 m ( y t i l i 10 b a e m r e P 1
10
20
30
40
Interparticle porosity (%) Fine crystalline dolograinstone Medium crystalline dolograinstone Large crystalline dolograinstone
Fine to medium crystalline grain-dominated dolopackstone
0.1 10
20
30
40
Interparticle porosity (%) Figure 10—Porosity-air permeability crossplots of nonvuggy dolomite fabrics compared with the three permeability fields illustrated in Figure 5. (A) Mud-dominated dolostones with dolomite crystal sizes ranging from 10 to 150 µm. (B) Dolograinstones (average grain size is 200 µm) with dolomite crystal sizes ranging from 15 to 150 µm. (C) Grain-dominated dolopackstones with fine to medium dolomite crystal sizes.
The dolostone permeability fields are defined by dolomite crystal size, as well as grain size and sorting of the precursor limestone. Within the field, permeability is defined by interparticle porosity. Systematic changes in intergrain and intercrystal porosity by predolomite calcite cementation, dolomite cementation, and compaction will systematically change the pore-size distribution. Therefore, interparticle porosity defines the permeability, and dolomite cr ystal size, grain size, and sorting define the permeability field.
than 100-µm permeability field are (1) limestone and dolomitized grainstones, and (2) large crystalline grain-dominated dolopackstones and mud-dominated dolostones. These fabrics are called rock-fabric/petrophysical class 1. The effect of grain size in this field can be seen by comparing Figures 7A and 10B. Ooid grainstones, which have a grain size of 400 µm, are more permeable for a given porosity than dolograinstones, which have a grain size of 200 µm. Fabrics that make up the 100–20-µm field are (1) grain-dominated packstones, (2) fine to medium crystalline grain-dominated dolopackstones, Limestone and Dolomite Comparison and (3) medium crystalline mud-dominated doloData from limestone and dolomite rock fabrics are stones. These fabrics are called rock-fabric/petrocombined into one porosity-permeability crossplot physical class 2. A comparison of Figures 7B and in Figure 12. The fabrics that make up the greater 10C shows that the dolomitized grain-dominated
Lucia
1285
Unusual Types of Interparticle Porosity
Diagenesis can produce unique types of interparticle porosity. Collapse of separate-vug fabrics because of overburden pressure can produce fragments that are properly considered “diagenetic particles.” Large dolomite crystals with their centers dissolved can collapse to form pockets of dolomite rim crystals. Leached grainstones can collapse to form intergrain fabrics composed of fragments of dissolved grains. These unusual pore types typically are not areally extensive; however, the collapse of extensive cavern systems can produce areally extensive bodies of collapse breccia. Interbrecciablock pores produced by cavern collapse are included in the touching-vug category because of their association with fractures and other vugs formed by karsting processes (Kerans, 1989).
Capillary Pressure/Rock-Fabric Relationships Several methods have been presented for relating Figure 11—Composite porosity-air permeability crossplot for nonvuggy dolostone fabrics compared with the porosity, permeability, water saturation, and reserthree permeability fields illustrated in Figure 6. voi r h eight (Leverett, 1941; Au frich t a nd Koepf,
packstones tend to be more permeable for a givenporosity than limestone grain-dominated packstones. Further data may show this to be an important difference attributable to dolomite crystal morphology. The less than 20-µm permeability field is characterized by mud-dominated limestone and fine crystalline mud-dominated dolostones. These fabrics are called rock-fabric/petrophysical class 3. Reduced–major-axis permeability transforms are presented for each class (Figure 12). The transform for class 2 is slightly skewed to the field boundaries, and the following transform is more compatible with the field boundaries. Class 1: k = (45.35 × 108 ) × φ ip8.537
r = 0.71
k = (1.595 × 105 ) × φ ip5.184
r = 0.80
Class 2:
or this recommended transform for class 2: k = (2.040 × 106 ) × φ ip6.38
Class 3: k = (2.884 × 103 ) × φ ip4.275
r = 0.81
where k = millidarcys and φip = fractional interparticle porosity.
1957; Heseldin, 1974; Alger et al., 1989). These methods attempt to average the capillary pressure curves into one relationship among saturation, porosity, permeability, and reservoir height without regard to rock fabric. The data presented demonstrate that the three nonvuggy rock-fabric fields have unique porosity-permeability relationships, suggesting that capillary properties of nonvuggy carbonates should also be separated into rock-fabric categories. To characterize the capillary properties of the three rock-fabric fields, capillary pressure curves with different interparticle porosities from the three rock-fabric fields are compared in Figure 13. Figure 13A shows two cur ves representative of class 1. Both curves are from samples of fine to medium crystalline dolograinstones. The 9.2% porosity curve represents the average of two sets of capillary pressure data from dolograinstones in the Taylor-Link field (Lucia et al., 1992b), and the 17.6% porosity curve represents the average of two data sets, one from the Taylor-Link field and one from the Dune field (Bebout et al., 1987). Three curves representative of class 3 are presented in Figure 13C. The curves are from samples of fine crystalline dolowackestones and from the Farmer field, Permian basin. Three curves representative of class 2 are presented in Figure 13B. Class 2 represents a very diverse class of rock fabrics, and it is difficult to combine all the fabrics into a few simple curves. The curves presented here are from medium crystalline dolowackestones of the Seminole San Andres unit and may not be representative of all grain-dominated packstones. Pore-throat-size distribution for each capillary pressure curve was calculated using the following
1286
Classification of Carbonate Pore Space
1000 Class 1 Class 2 Class 3
Class 1
m
µ
0 0 1
Class 2
m
µ
2 0
100
Class 3
) d m ( y t i l i 10 b a e m r e P
m
µ
0 0 5
1
0.1 0.05
0.10
0.15
0.20
0.30
0.40
Interparticle porosity (fractional) Figure 12—Composite porosity-air permeability crossplot for nonvuggy l imestones and dolostones showing statistical reduced–major-axis transforms for each class (dashed lines). See text for equations.
formula. The results (Figure 13) show decreasing pore-throat size with decreasing porosity within a petrophysical class and a general decrease in porethroat size from class 1 to class 3. Microporosity [pore-throat size less than 1 µm (Pittman, 1971)] is concentrated in class 3 and decreases in importance from class 3 to class 1. Ri = (2Τ × cosθ × C )/ Pc
where Ri = pore-throat radius ( µm); T = air-mercury interfacial tension = 480 dynes/cm; θ = air-mercury contact angle = 140°; C = unit conversion constant = 0.145; and Pc = mercury injection pressure (psia). Each group of curves is characterized by similar displacement pressures and a systematic change in curve shape and saturation characteristics with changes in interparticle porosity. The relationship among porosity, saturation, and rock-fabric class can be demonstrated by selecting a reservoir height of 150 m (which equates to a mercury capillary
pressure of about 650 psia) and plotting saturation against porosity for each rock-fabric class. The results (Figure 14) show that in nonvuggy carbonate reservoirs, a plot of porosity vs. water saturation can separate rock fabrics into three classes. Equations relating water saturation to porosity and reservoir height are developed in two steps (Lucia et al., 1992b). First, mercury capillary pressure is converted to reservoir height using generic values (Table 2). Second, wetting-phase saturations from the capillary pressure curves are plotted against porosity for several reservoir heights. Third, lines of equal reservoir height are drawn assuming equal slopes and a relationship between intercepts and reservoir height is developed. This relationship is substituted for the intercept term in the porosity vs. saturation equation, resulting in a relationship among water saturation, porosity, and reservoir height. The resulting equatio ns follow, and threedimensional representations for classes 1 and 3 are presented in Figure 15. These equations are
(A) ) 2000 a i s p ( e r 1600 u s s e r 1200 p n o i t 800 c e j n i y r 400 u c r e 0 M 100
(A') 100
) % ( n o i t a r u t a s y r u c r e M
9.2% porosity
17.60% porosity
80
60
40
20
17.6% porosity Permeability = 166 md
80 60 9.2% porosity Permeability = 72 md
40 20 0
0
2.0
1.5
Mercury saturation (%)
0
-0.5
-1.0
-1.5
(B') 100 8.7% porosity
) % ( n o i t a r u t a s y r u c r e M
10.3% porosity 12.5% porosity 13.7% porosity
13.7% porosity Permeability = 5.1md 80 12.5% porosity Permeability = 2.3 md
60
10.3% porosity Permeability = 1.3 md
40 20
8.7% porosity Permeability = 0.05 md
0 80
60
40
20
2.0
0
1.5
1.0
0.5
0
-0.5
-1.0
-1.5
-1.0
-1.5
Log pore-throat radius (µm)
Mercury saturation (%)
(C')
(C) ) 2000 a i s p ( e r 1600 u s s e r 1200 p n o i t c 800 e j n i y r 400 u c r e 0 M 100
0.5
Log pore-throat radius (µm)
(B) ) 2000 a i s p ( e r 1600 u s s e r 1200 p n o i t 800 c e j n i y r 400 u c r e 0 M 100
1.0
100
) % ( n o i t a r u t a s y r u c r e M
9.0% porosity 11.8% porosity 15.4% porosity
15.4% porosity Permeability = 2 md
80
11.8% porosity Permeability = 0.4 md
60 40
9.0% porosity Permeability = 0.04 md 20 0
80
60
40
Mercury saturation (%)
20
0
2.0
1.5
1.0
0.5
0
-0.5
Log pore-throat radius (µm)
Figure 13—(A–C) Capillary pressure curves and (A ′–C′) pore-size distribution for petrophysical classes. (A) Class 1. Data are from dolograinstones. (B) Class 2. Data are from medium crystalline dolowackestones. (C) Class 3. Data are from fine crystalline dolowackestones.
1288
Classification of Carbonate Pore Space
100
Table 2. Values Used to Convert Mercury/Air Capillary Pressure to Reservoir Height
) % ( n o i t a r u 10 t a s r e t a W
Laboratory (mercury/air/solid)
Class 3 Class 1
Class 2
T (dynes/cm)
θ (°) Water density
1 5
10
20
30
40
Figure 14—Crossplot of porosity and water saturation for the three rock-fabric/petrophysical classes at a reser voir heig ht of 150 m. Water saturation (1 mi nus mercury saturation) and porosity values are taken from capillary pressure curves illustrated in Figure 13.
specific to the capillary pressure curves used in this paper and will not necessarily apply to specific reservoirs. However, the equations will provide reasonable values for original water saturations when only porosity and rock-fabric data are available. Class 1: Sw = 0.02219 × H –0.316 × φ –1.745
Reservoir (oil/water/solid)
480 140 1.04
28 44 0.88
Class 2 comprises grain-dominated packstones, fine and medium crystalline grain-dominated dolopackstones, and medium crystalline mud-dominated dolostones: k = (2.040 × 106 ) × φ ip6.38 Sw = 0.1404 × H –0.407 × φ –1.440
Class 3 contains mud-dominated limestones and fine crystalline mud-dominated dolostones: k = (2.884 × 103 ) × φ ip4.275 Sw = 0.6110 × H –0.505 × φ –1.210
Petrophysics of Separate-Vug Pore Space
The addition of vuggy pore space to interparticle pore space alters the petrophysical characterClass 2: istics by altering the manner in which the pore –0.407 –1.440 space is connected, all pore space being connect×φ Sw = 0.1404 × H ed in some fashion. Separate-vug pore space is Class 3: defined as pore space that is (1) either within particles or is significantly larger than the particle Sw = 0.6110 × H –0.505 × φ –1.210 size (generally greater than 2× ), and (2) is inter where H = height above capillary pressure equal to connected only through the interparticle porosity (Figure 17). Separate vugs (Figure 18) are typically zero and φ = fractional porosity. fabric selective in origin. Intrafossil pore space, such as the living chambers of a gastropod shell; Rock-Fabric/Petrophysical Classes moldic pore space, such as disso lved grains (oomolds) or dolomite crystals (dolomolds); and Because the three rock-fabric groups define per- intragranular microporosity are examples of intrameability and saturation fields, the three groups, particle, fabric-selective separate vugs. Molds of together with interparticle porosity and reservoir evaporite crystals and fossils found in mud-domiheight, can be used to relate petrophysical proper- nated fabrics are examples of fabric-selective sepaties to geologic observations. These rock-fabric rate vugs that are significantly larger than the pargroups, herein termed rock-fabric/petrophysical ticle size (Figure 18). In mud-dominated fabrics, classes (Figure 16), are described, with their gener- shelter pore space is typica lly much larger than ic transform equations as follows: the particle size and is classified as separate-vug Class 1 is composed of grainstones, dolograin- porosity. stones, and large crystalline dolostones: In grain-dominated fabrics, extensive selective leaching of grains may cause grain boundaries to k = (45.35 × 108 ) × φ ip8.537 dissolve, producing composite molds. These composite molds may have the petrophysical characterSw = 0.02219 × H –0.316 × φ –1.745 istics of separate vugs. If, however, dissolution of
Lucia
1289
Figure 15—Three-dimensional displays of (A) class 1 and (B) class 3 equations relating water saturation to reservoir height and porosity.
W a t e r s a t ur a t i on ( % )
W a t e r s a t u r a t io n ( % )
the grain boundaries is extensive, the pore space may be classified as interparticle porosity. Grain-dominated fabrics may contain grains with intragranular microporosity (Pittman, 1971; Keith and Pittman, 1983; Asquith, 1986). Intragranular microporosity is classified as a separate vug because it is within the particles of the rock and is interconnected only through the intergrain pore network. Mud-dominated fabrics also may contain grains with microporosity, but they present no unique petrophysical condition because of the similar pore sizes between the microporosity in the mud matrix and in the grains.
) y ( % o s i t P o r
) y ( % o s i t r o P
The addition of separate-vug porosity to interparticle porosity increases total porosity, but does not significantly increase permeability (Lucia, 1983). Figure 19A illustrates this principle in that permeability of a moldic grainstone is less than would be expected if all of the porosity were interparticle and, at constant porosity, permeability were to increase with decreasing separate-vug porosity (Lucia and Conti, 1987). This principle is also true for intragranular microporosity. Figure 19B is a crossplot of data from a San Andres dolograinstone from the Permian of west Texas. The crossplot shows that the permeability of the grainstone with
1290
Classification of Carbonate Pore Space
Figure 16—Petrophysical and rock-fabric classes based on similar capillary properties and interparticle-porosity/ permeability transforms.
intragranular microporosity is less than would be expected if all the porosity were interparticle. Separate vugs found within the oil column are usually considered to be saturated with oil because of their relatively large size; oil migration into separate vugs is controlled by interparticle pore size. Intragranular microporosity, however, is unique because of small pore size. The small pore size produces high capillary forces that trap water and lead to anomalously high water saturations within a productive interval (Pittman, 1971; Dixon and Marek, 1990). Figure 20 shows the capillary characteristics of a Cretaceous grainstone composed of grains having microporosity between 2-µm crystals (Keith and Pittman, 1983) and a San Andres dolograinstone with grains having microporosity between 10-µm dolomite crystals. This figure illustrates that within the oil column, water saturations in grainstones with intragranular microporosity are significantly higher than would be expected if no intragranular microporosity were present. The
capillary pressure curves typically have a bimodal character.
Petrophysics of Touching-Vug Pore Space Touching-vug pore systems are defined as pore space that is (1) significantly larger than the particle size, and (2) forms an interconnected pore system of significant extent (Figures 17, 18). Touching vugs are typica lly nonfab ri c selective in or ig in . Cavernous, breccia, fracture, and solution-enlarged fracture pore types commonly form an interconnected pore system on a reservoir scale and are typical touching-vug pore types (Figure 18). Fenestral pore space is commonly connected on a reservoir scale and is grouped with touching vugs because the pores are normally much larger than the grain size (Major et al., 1990). Fracture porosity is included as a touching-vug pore type because fracture porosity is an important
Lucia
1291
Figure 17—Geological and petrophysical classification of vuggy pore space based on vug interconnection. The vol ume of separate-vug pore space is important for characterizing the petrophysical properties.
contributor to permeability in many carbonate reservoirs and, therefore, must be included in any petrophysical classification of pore space. Although fracturing is often considered to be of tectonic origin and thus not a part of carbonate geology, diagenetic processes common to carbonate reservoirs, such as karsting (Kerans, 1989), can produce extensive fracture porosity. The focus of this classification is on petrophysical properties rather than genesis, and must
include fracture porosity as a pore type irrespective of its origin. Touching vugs are thought to be typically filled with oil in the reservoirs and can increase permeability well above what would be expected from the interparticle pore system. Lucia (1983) illustrated this fact by comparing a plot of fracture permeability vs. fracture porosity to the three porosity-permeability fields for interparticle porosity (Figure 21). This graph shows that permeability in
A
B 0
0.5 mm
1.0
C
0
1.0 mm
2.0
D 0
1.0 mm
2.0
E
0
F 0
0.2 mm
1.0 mm
2.0
20 microns
H
G 1 in.
1 in.
I
J 1 in.
3 in.
K 0
1.0 mm
2.0
Figure 18—Photomicrographs and photographs showing examples of vug pore types. Separate-vug types: (A) oomoldic porosity, φ = 26%, k = 3 md, Wolfcampian, west Texas; (B) intrafossil pore space in a gastropod shell, Cretaceous, Gulf Coast; (C) fossil molds in wackestone, φ = 5%, k = 0.05 md; (D) anhydrite molds in grain-dominated packstone, φ = 10%, k = <0.1 md, Mississippian, Montana; (E) fine crystalline dolograinstone with intergranular and intragranular microporosity pore types, φ = 10%, k = 3 md, Farmer field, west Texas; (F) scanning electron photomicrograph of dolograins in (E) showing intragranular microporosity between 10-µm crystals. Touching-vug types: (G) cavernous porosity in a Niagaran reef, northern Michigan; (H) collapse breccia, Ellenburger, west Texas; (I) solution-enlarged fractures, Ellenburger, west Texas; (J) cavernous porosity in Miami oolite, Florida; (K) fenestral porosity in pisolitic dolostone. Note that the fenestral pores are more than twice the size of the enclosing grains.
1294
Classification of Carbonate Pore Space
(A)
(B)
1000
1000
Grainstone field
) 100 d m ( y t i l i b 10 a e m r e 1 P
Svug porosity average 8%
Grainstone field
) 100 d m ( y t i l i b 10 a e m r e 1 P
Svug porosity average 20%
0.1
Dolograinstone with intragrain microporosity
0.1 5
10
20
30
40
Porosity (%)
5
10
20
30
40
Porosity (%)
Figure 19—Crossplot illustrating the effect of separate-vug porosity on air permeability. (A) Grainstones with separate-vug (Svug) porosity in the form of grain molds plot to the right of the grainstone field in proportion to the vol ume of separate-vug porosity. (B) Dolograin stones with separate vug s in the form of intragranular microporosity plot to the right of the grainstone field.
touching-vug pore systems is related principally to fracture width and is sensitive to extremely small changes in fracture porosity. There is no effective method for measuring fracture width using the rock-fabric approach.
IMPLICATIONS FOR CONSTRUCTING A GEOLOGIC MODEL A fundamental problem in constructing a reser voir model for input into numerical fluid-flow simulators is converting geologic observations into petrophysical rock properties, namely porosity, permeability, and saturation. The problem has been discussed by Lucia et al. (1992a), Senger et al. (1993), and Kerans et al. (1994). These workers suggested that the petrophysical properties within a rock-fabric unit are nearly randomly distributed and thus form the basic building blocks for reservoir models. Therefore, geologic models described in terms of rock-fabric units can be most effectively converted into petrophysical parameters and into reservoir models. On the basis of data presented herein, the most important rock-fabric characteristics to describe and map in nontouching-vug reservoirs are (1) grain size and sorting using the modified Dunham classification, (2) dolomite crystal size using 20 and 100 µm as size boundaries, (3) separate-vug porosity, (4) separate-vug type with special attention to intergrain microporosity, and (5) total porosity. The stacking patterns of rock-fabric units are systematic within a depositional and diagenetic
environment (Lucia, 1993). The simplest example is a shoaling-upward bar complex. Mud-dominated fabrics grade upward and laterally into grain-dominated fabrics with a corresponding increase in permeability, decrease in water saturation, and little change in porosity (Figure 22). The permeability and water saturation are a function of total porosity and rock fabric. The introduction of moldic porosity (separate vugs) by selective dissolution of grains in the grain-dominated fabric, caused perhaps by shoaling and the introduction of meteoric water at a subaerial exposure surface, can produce a diagenetic overprint that results in no significant change in porosity, little change in water saturation assuming the molds are filled with hydrocarbons, and drastically reduced permeability in the grainstone (Figure 22). The permeability is a function of interparticle porosity determined by subtracting separate-vug porosity from total porosity. Conversion of the shoaling-upward bar complex from limestone to 50-µm dolostone will not change the grain size and sorting characteristics of the grain-dominated fabrics, but will alter the particle size characteristics of the mud-dominated fabrics, changing the petrophysical class from class 3 to class 2. Dolomitization may reduce the total porosity of the grain-dominated fabrics, but the permeability and water saturation will still be a function of grain size, sorting, and intergrain porosity. If the precursor limestone is a moldic grain-dominated fabric, the resulting grain-dominated dolostone will be moldic (Figure 22). Dolomitization of the muddominated fabric to a 50-µm dolomite may reduce
Lucia
(A')
(A)
) 2000 a i s p ( e r 1500 u s s e r p n 1000 o i t c e j n 500 i y r u c r e M 0
1295
100 Unimodal pore system 11.2%, 44.9 md
Bimodal pore system 16.4%, 5.4 md
) % ( n o i t a r u t a S y r u c r e M
80 Unimodal pore system 11.2%, 44.9 md
60 40
Bimodal pore system 16.4 %, 5.4 md
20 0
100
80
60
40
20
1.5
0
1.0
(B)
100
0
-0.5
-1.0
-1.5
Log pore-throat radius (µm)
Mercury saturation (%)
) 2000 a i s p ( e r 1500 u s s e r p 1000 y r a l l i p a c 500 y r u c r e M 0
0.5
(B') 100
) % ( n o i t a r u t a S y r u c r e M
Bimodal pore system 18.2%, 14.0 md Unimodal pore system 17.6%, 166 md
Unimodal pore system 17.6%, 166 md 80 60 40 Bimodal pore system 18.2%, 14.0 md
20 0
80
60
40
20
0
Mercury saturation (%)
2.0
1.5
1.0
0.5
0
-0.5
-1.0
-1.5
Log pore-throat radius (µm)
Figure 20—(A, B) Capillary pressure curves and (A ′, B′) pore-size distribution illustrating the effect of intragranular microporosity on capillary properties. Notice the bimodal character of the samples with both intergranular and intragranular microporosity pore types. (A) Ooid grainstone with intergranular porosity compared with ooid grainstone with both intergranular and intragranular microporosity pore types, Rodessa limestone, Cretaceous, east Texas (after Kei th and Pittman, 1983). (B) Dolograinstone with intergranular porosity compared with dolograinstone with both intergranular porosity and intragranular microporosity pore types, San Andres dolomite, Permian (Farmer San Andres field), west Texas.
the porosity, but will increase the permeability and decrease the water saturation significantly. Another simple example is the high-energy tidalflat–capped cycle where intertidal and supratidal sediments overlie a shoreface grain-dominated packstone (Figure 23). Carbonate cement and compaction have reduced porosity in the tidal-flat sediments, resulting in a low-permeability, high–watersaturation flow unit overlying a permeable grain-dominated limestone and a low-permeability mud-dominated limestone. In a seaward direction, shoaling-upward bar complexes replace the tidalflat–capped cycles.
Evaporitic tidal-flat environments are a source of dolomitizing water, and downward f low of the hypersaline water can dolomitize underlying carbonate sediments (Figure 23). In Permian basin reservoirs, intertidal and supratidal dolostones are dense because of compaction and occlusion of pore space by anhydrite and dolomite, and graindominated dolostones underlying the tidal-flat sediments commonly are dense because of pore-filling anhydrite. High porosity is confined to subtidal mud-dominated dolostones, and permeability and water saturation are related to intercrystal porosity and dolomite crystal size. In this example (Figure
1296
10 5 10 4 10 3
Classification of Carbonate Pore Space
m
porosity rather than total porosity less separate-vug porosity, and permeability and saturation can be seriously underestimated if the changes in particle size resulting from dolomitization are not included in the geologic model.
m
c c kf= 84.4 x 10 5 W 3 /Z 1 . 5 0 = φf= W/Z x 100 = w w z = Fracture spacing c m m w = Fracture width c 1 . 0 0 5 m Kf= Fracture permeability . = 0 c w = φf= Fracture porosity 0 0 1 = Z
10 2
) d ( 10 y t i l i b 1 a e m r e 10 -1 P
SUMMARY
m c 0 1 = Z c m Inter m 0 1 5 c particle . 0 0 = . 0 w = 0 >100µm w m c µ 1 100 to 20 m = Z <20µm
m c 0 0 0 1 = Z
10 -2
w
φ
c m 1 0 0 0 . = w
10 -3 10 -4 10 -5
10 -4
10 -3
10 -2
10 -1
1
10
100
Porosity (%) Figure 21—Theoretical fracture air permeability/porosity relationship compared to the rock-fabric/petrophysical porosity, permeability fields (Lucia, 1983).
23), dolomite crystal size i ncreases from fine to medium with depth and the petrophysical properties change correspondingly. Grain-dominated bodies seaward of the tidal-flat deposits are dolomitized by hypersaline water from an overlying cycle and are not cemented by anhydrite. The permeability and saturation characteristics are related to grain size, sorting, and intergrain porosity (Figure 23). The underlying mud-dominated dolostone is composed of medium dolomite crystals and has higher permeability than a muddominated limestone with the same interparticle porosity. The underlying limestone has low porosity and permeability in this example. These examples illustrate the importance of mapping rock-fabric units within the geologic framework. If the shoaling-upward cycle is mapped as a bar facies or the tidal-flat–capped cycle is mapped as a peritidal facies, the permeability and saturation structure is compromised and the resulting model is oversimplified. In addition, if the diagenetic facies are not included as mapping parameters, the permeability and saturation values in the resulting reservoir model may be in serious error. For example, permeability can be significantly overestimated if the estimate is based on total
The goal of reservoir characterization is to describe the spatial distribution of petrophysical parameters, such as porosity, permeability, and saturation. The rock-fabric approach presented here is based on the premise that pore-size distribution controls the engineering parameters of permeability and saturation, and that pore-size distribution is related to rock fabric, which is a product of geologic processes. Thus, rock fabric integrates geologic interpretation with engineering numerical measurements. To determine the relationships between rock fabric and petrophysical parameters, one must define and classify pore space as it exists today in terms of petrophysical properties. This is best accomplished by dividing pore space into pore space located between grains or crystals, called interparticle porosity, and all other pore space, called vuggy porosity. Vuggy pore space is further subdivided into two groups, depending on how the vugs are interconnected: (1) vugs interconnected only through the interparticle pore network are termed separate vugs, and (2) vugs in direct contact are termed touching vugs. The petrophysical properties of interparticle porosity are related to particle size, sorting, and interparticle porosity. Grain size and sorting of grains and micrite are based on Dunham’s (1962) classification and are modified to make the classification compatible with petrophysical considerations. Instead of dividing fabrics into grain-supported and mud-supported categories, fabrics are divided into grain-dominated and mud-dominated categories. The important attributes of grain-dominated fabrics are the presence of open or occluded intergrain porosity and a grain-supported texture. The important attribute of mud-dominated fabrics is that the areas between the grains are filled with mud even if the grains appear to form a supporting framework. Grainstone is clearly a grain-dominated fabric, but Dunham’s (1962) packstone class bridges an important petrophysical boundary. Some packstones, as we see them now, have intergrain pore space, and some have the intergrain spaces filled with mud. Therefore, the packstone textural class must be divided into two rock-fabric classes: (1) grain-dominated packstones having intergrain pore space or cement, and (2) mud-dominated packstones having intergrain spaces filled with mud.
Lucia
1297
Figure 22—Three examples of how the stacking of rock-fabric units affects the distribution of porosity, permeability, and water saturation in a shoaling-upward sequence with selective dissolution and dolomitization overprints.
1298
Classification of Carbonate Pore Space
Figure 23—Examples of how the stacking of rock-fabric units affects the distribution of porosity, permeability, and water saturation in a tidal-flat–capped seque nce with hypersali ne reflux dolomitization and sulfate emplacement overprints.
Lucia
The important fabric elements to recognize for petrophysical classification of dolomites are precursor grain size and sorting, dolomite crystal size, and interparticle (intergrain and intercrystal) porosity. Important dolomite crystal size boundaries are 20 and 100 µm. Dolomite crystal size has little effect on the petrophysical properties of grain-dominated dolostones. The petrophysical properties of muddominated fabrics, however, are significantly improved when the dolomite crystal size is less than 20 µm. Permeability and saturation characteristics of interparticle porosity can be grouped into three rock-fabric/petrophysical classes. Class 1 is composed of limestone and dolomitized grainstones, and large crystalline grain-dominated dolopackstones and mud-dominated dolostones. Class 2 is composed of grain-dominated packstones, fine to medium crystalline grain-dominated dolopackstones, and medium crystalline mud-dominated dolostones. Class 3 is composed of mud-dominated limestone and fine crystalline mud-dominated dolostones. Generic porosity-permeability transforms and water saturation, poro sity, reservoir-height equations for each rock-fabric/petrophysical class are presented in the section “Rock-Fabric/Petrophysical Classes.” The addition of separate-vug porosity to interparticle porosity increases total porosity, but does not significantly increase permeability; therefore, one must determine interparticle porosity by subtracting separate-vug porosity from total porosity and use interparticle porosity to estimate permeability. Separate-vug porosity is thought to be typically filled with hydrocarbons in the reser voir. Intergranular microporosity, however, may contain significant amounts of capillar y-bound water, resulting in water-free production of hydrocarbons from intervals with higher than expected water saturation. Touching-vug pore systems cannot be related to porosity, but are related principally to fracture width. Bec ause the re is no effective met hod for making this observation in the reservoir, the rockfabric approach cannot be used to characterize touching-vug reservoirs. The key to constructing a geologic model that can be quantified in petrophysical terms is to select facies or units that have unique petrophysical qualities for mapping. Petrophysical properties in rockfabric facies or units are nearly randomly distributed and can be legitimately averaged, making these geologic units ideal for petrophysical quantification. In nontouching-vug reservoirs, the most important rock-fabric characteristics to describe and map are (1) grain size and sorting using the modified Dunham classification, (2) dolomite crystal size
1299
using 20 and 100 µm as size boundaries, (3) separate-vug type with special attention to intergrain microporosity, (4) total porosity, and (5) separate vug porosity. In touching-vug reservoirs, characterizing the pore system is difficult because the pore system is not related to a precursor depositional fabric, but is typically wholly diagenetic. Although the pore system may conform to bedding, as in evaporite collapse brecciation, it more commonly cuts across stratal boundaries. Recognizing the presence of a touching-vug pore system is paramount, however, because it may dominate the flow characteristics of the reservoir.
REFERENCES CITED Alge r, R . P. , D. L. L uffel, and R. B . Tr uman , 19 89, N ew u nifi ed method of integrating core capillary pressure data with well logs: Society of Petroleum Formation Evaluation, v. 4, no. 2, p. 145–152. Arch ie, G. E. , 195 2, Cl assi fica tion of c arbo nate reservoir rocks and petrophysical considerations: AAPG Bulletin, v. 36, no. 2, p. 278–298. Asquith, G. B., 19 86, Micropo rosity in the O’Ha ra oolit e zone of the Mississippian Ste. Genevieve Limestone, Hopkins County, Kentucky, and its implications for formation evaluation: Carbonates and Evaporites, v. 1, no. 1, p. 7–12. Aufricht, W. R., and E. H. Koepf, 1957, The interpretation of capillary pressure data from carbonate reservoirs: Transactions of the American Institute of Mining, Metallurgical, and Petroleum Engineers, v. 210, p. 402–405. Bebout, D. G., F. J. Lucia, C. F. Hocott, G. E. Fogg, and G. W. Vander Stoep , 1987, Chara cteri zati on of the Gr ayburg reser vo ir , U nive rs it y L an ds Du ne fi el d, Cr an e C ou nt y, Te xa s: University of Texas at Austin, Bureau of Economic Geology Report of Investigations No. 168, 98 p. Choquette, P. W., and L. C. Pray, 1970, Geologic nomenclature and classification of porosity in sedimentary carbonates: AAPG Bulletin, v. 54, no. 2, p. 207–250. Choquette, P. W., and R. P. Steiner, 1985, Mississippian oolite and non-supratidal dolomite reservoirs in the Ste. Genevieve Formation, North Bridgeport field, Illinois basin, in P. O. Roehl and P. W. Choquette, eds., Carbonate petroleum reservoirs: New York, Springer-Verlag, p. 209–238. Dixon, F. R., and B. F. Marek, 1990, The effect of bimodal pore size distribution on electrical propert ies of some Middle Eastern limestones: Society of Petroleum Engineers Technical Conference, SPE 20601, p. 743–750. Dunham, R. J., 1962, Classification of carbonate rocks according to depositional texture, in W. E. H am, e d., Cla ssifi cation s of carbonate rocks—a symposium: AAPG Memoir 1, p. 108–121. Heseldin, G. M., 1974, A method of averaging capillary pressure curves: Society of Professional Well Log Analysts Annual Logging Symposium, paper E, p. 1–7. Keith, B. D., and E. D. Pittman, 1983, Bimodal porosity in oolitic reservoir—effect on productivity and log response, Rodessa limestone (Lower Cretaceous), East Texas basin: AAPG Bulletin, v. 67, no. 9, p. 1391–1399. Kerans, C., 1989, Karst-controlled reservoir heterogeneity in the Ellenburger Group carbonates of west Texas: AAPG Bulletin, v. 72, no. 10, p. 1160–1183. Kerans, C., F. J. Lucia, and R. K. Senger, 1994, Integrated characterization of carbonate ramp reservoirs using Permian San Andres Formation outcrop analogs: AAPG Bulletin, v. 78, no. 2, p. 181–216. Leverett, M. C., 1941, Capillary beha vior in porous solids:
1300
Classification of Carbonate Pore Space
Transactions of the American Institute of Mining, Metallurgical, and Petroleum Engineers, v. 142, p. 151–169. Lucia, F. J., 1962, Diagenesis of a crinoidal sediment: Journal of Sedimentary Petrology, v. 32, no. 4, p. 848–865. Lucia, F. J., 1983, Petrophysical parameters estimated from visual description of carbonate rocks: a field classification of carbonate pore space: Journal of Petroleum Technology, March, v. 35, p. 626–637. Lucia, F. J., 1993, Carbonate reservoir models: facies, diagenesis, and flow characterization, in D. Morton-Thompson and A. M. Woods, ed s., Dev elopment geology r efere nce ma nual: AA PG Methods in Exploration 10, p. 269–274. Lucia, F. J., and R. D. Conti, 1987, Rock fabric, permeability, and log relationships in an upward-shoaling, vuggy carbonate sequence: University of Texas at Austin, Bureau of Economic Geology Geological Circular 87-5, 22 p. Lucia, F. J., C. Kerans, and R. K. Senger, 1992a, Defining flow units in dolomitized carbonate-ramp reservoirs: Society of Petroleum Engineers, SPE 24702, p. 399–406. Lucia, F. J., C. Kerans, and G. W. Vander Stoep, 1992b, Characterization of a karsted, high-energy, ramp-margin carbonate reservoir: Taylor-Link West San Andres unit, Pecos County, Texas: University of Texas at Austin, Bureau of Economic Geology Report of Investigations No. 208, 46 p.
ABOUT THE AUTHOR F. Jerry Lucia F. Jerry Lucia retired from Shell Oil Company in 1985, where he had worked as a consulting geological engineer, after 31 years in research and operations. Currently, he is a senior research fellow with the University of Texas at Austin, Bureau of Economic Geology, developing new techniques and methods of characterizing carbonate reservoirs by applying outcrop studies to subsurface reservoirs.
Major, R. P., G. W. Vander Stoep, and M. H. Holtz, 1990, Delineation of unrecovered mobile oil in a mature dolomite reservoir: East Penwell San Andres unit, University Lands, west Texas: University of Texas at A ustin Bureau of Economic Geology Report of Investigations No. 194, 52 p. Moshier, S. O., C. R. Handford, R. W. Scott, and R. D. Boutell, 1988, Giant gas accumulation in “chalky”-textured micritic limestones, Lower Cretaceous Shuaiba Formation, eastern United Arab Emirates, in A. J. Lo ma nd o an d P. M. Harris, eds., Giant oil and gas fields: Society of Economic Paleontologists and Mineralogists Core Workshop 12, v. 1, p. 229–272. Murray, R. C., 1960, Origin of porosity in carbonate rocks: Journal of Sedimentary Petrology, v. 30, no. 1, p. 59–84. Pittman, E. D., 1971, Microporosity in carbonate rocks: AAPG Bulletin, v. 55, no. 10, p. 1873–1881. Scholle, P. A., 1977, Chalk diagenesis and its relation to petroleum exploration: oil from chalks, a modern miracle?: AAPG Bulletin, v. 61, no. 7, p. 982–1009. Senger, R. K., F. J. Lucia , C. Kerans, and M. A. Ferris, 1993, Dominant control of reservoir-flow behavior in carbonate reservoirs as determined from outcrop studies, in B. Linville, R. E. Burchfield, and T. C. Wesson, eds., Reservoir characterization III: Tulsa, Oklahoma, PennWell, p. 107–150.