Вы находитесь на странице: 1из 25

Making sense of carbonate pore systems

Arve Lny

AUTHOR Arve Lny  Norsk Hydro Oil & Energy, Sandsliveien 90, N-5049 Sandsli, Bergen, Norway; arve.lonoy@hydro.com Arve Lny received a Cand. Real. degree in carbonate sedimentology from the University of Bergen in 1981. He has since then worked as a research geologist on carbonate sedimentology and, to a lesser extent, siliciclastic petrography in the Norsk Hydro Research Center. His main work on carbonates has been on Paleozoic carbonates of the Arctic (Norway, Canada, and Russia) and the Mesozoic of western Africa and the Middle East.

ABSTRACT The most widely used pore-type classification systems for carbonate reservoirs are limited by the fact that the relation between porosity and permeability is poorly defined. Existing classification schemes for porosity-permeability data do not, in many cases, optimally integrate sedimentology, diagenesis, and flow-related properties. In many carbonate reservoirs, it is therefore difficult to generate predictive models for reservoir-quality distribution, resulting in significant uncertainty in hydrocarbon reserve calculations. Based on empirical data, mostly from Europe and the Middle East, a new pore-type classification system has been developed. The new system not only uses elements from existing pore-type classification systems, but also introduces many new elements. The new pore-type system includes 20 pore-type classes that show a predictable relation between porosity and permeability. It combines sedimentologic and diagenetic features with flow-related properties, and reservoir-critical parameters can thus be predicted using sedimentologic and diagenetic models. A practical example based on data from a Devonian hydrocarbon field shows that pore-type variations may account for severalhundred-percent differences in calculated hydrocarbon reserves.

ACKNOWLEDGEMENTS I am grateful to my colleagues David Hunt, Ole Martinsen, and Brian A. Farrelly for their review of an early version of the manuscript and to Norsk Hydro for permission to publish the article. I also thank David Eby, Jerry Lucia, and Wayne M. Ahr for their helpful reviews and comments. I am particularly indebted to William A. Morgan, who gave constructive criticism and suggestions that sig nificantly improved the article.

INTRODUCTION The porosity classification system most widely used by carbonate petrographers and petroleum geologists is the one developed by Choquette and Pray (1970). This system is closely linked to sedimentological fabric, and pore types can therefore, to some extent, be predicted based on depositional setting and/or diagenetic evolution. This classification system is particularly useful for porosity evolution studies that are important for exploration. Among petrophysicists and reservoir engineers, the classification of Archie (1952) and, to a greater extent, Lucia (1983, 1995, 1999) is preferred because the pore classification is more directly linked to pore geometries and flow properties.

Copyright #2006. The American Association of Petroleum Geologists. All rights reserved. Manuscript received July 5, 2005; provisional acceptance September 12, 2005; revised manuscript received January 4, 2006; final acceptance March 13, 2006. DOI:10.1306/03130605104

AAPG Bulletin, v. 90, no. 9 (September 2006), pp. 1381 1405

1381

A major challenge in the evaluation of carbonate reservoirs is to understand the relationship between pore type and porosity and permeability. Such information, commonly derived from routine thin-section description and core-plug analysis, together with wireline-log data, is a basic constraint of both stock-tankoriginal-oil-in-place (STOOIP) calculations and dynamic simulations of hydrocarbon reservoirs. Lucia (1983, 1995, 1999) made a major contribution to understanding porosity-permeability relationships in carbonates. He showed how incorporating textural and particle size information helps to better understand porositypermeability relationships in carbonates. The importance of this work is highlighted by its widespread use in the hydrocarbon industry. The aim of this article is to present the results of a study that set out to analyze porosity and permeability relationships in carbonate reservoirs from a wide range of sedimentary basins, geological ages, and burial depths. As part of this study, core-plug thin sections were described using the semiquantitative classification of Lucia (1999) and the descriptive terminology of Choquette and Pray (1970). These data show a wide range of coefficient of determination values for permeability and porosity (R 2 in the range of 0.45 0.88), and it is apparent that some pore types are, in terms of porositypermeability relationships, poorly classified in comparison to others ( Table 1). This realization prompted the search for an alternative carbonate pore-system

classification that would provide a higher degree of correlation between porosity and permeability. The new pore-type classification presented here incorporates elements of both rock texture and pore size, reflecting depositional and diagenetic fabrics. As such, it builds from the important work of Choquette and Pray (1970) and Lucia (1983, 1995, 1999). However, the new classification provides a significant improvement over existing schemes in terms of a better correlation of matrix-related porosity and permeability, hence, for porosity cutoff values and permeability prediction. The latter is one of the most important elements of carbonate reservoir description. This paper first discusses the data types and methods employed followed by an illustrated description of the different pore classes. Then, a consideration of the observed porosity-permeability relationships for each pore class and a comparison to other pore classifications follow. A case study from a Devonian reservoir using the new classification scheme is used to illustrate its potential impact on reserve estimations.

DEVELOPMENT OF THE PORE CLASSIFICATION SCHEME The impetus for developing a new pore-classification scheme resulted from classification of the carbonate database in Norsk Hydro using the widely applied Lucia (1983, 1995, 1999) pore classification. However, a wide range and scatter of data in each of the pore class fields were apparent, with different pore classes showing very different coefficients of determination (R 2) for porosity and permeability (Table 1). This observation prompted the analysis of the data using the Choquette and Pray (1970) pore classification system. Although there is a wide range of R 2 using the Choquette and Pray classification (Table 1), this analysis led to the realization that there is a significant difference in permeability trends between intercrystalline and interparticle porosity, two pore types that are grouped together in the Lucia classification. One striking feature in the data set was the patchy distribution of porosity in a high proportion of the studied thin sections (i.e., >30% of the interparticle and intercrystalline pore-type samples). This observation prompted a first attempt at subdividing the porosity scheme of Choquette and Pray (1970) into uniform and patchy subclasses on the basis of both thin-section and plug description. This approach revealed natural pore-type subclasses that allow for a refined relationship between porosity and permeability and increased R 2 values.

Table 1. Porosity-Permeability Coefficients of Determination (R 2) for the Lucia (1983, 1995, 1999) and Choquette and Pray (1970) Classification Systems* Pore Type Lucia (1983, 1995, 1999) Classification System Interparticle, class 3 Interparticle, class 2 Interparticle, class 1 Vuggy, separate Vuggy, touching Choquette and Pray (1970) Classification System Interparticle Intercrystalline Moldic Intraparticle Vuggy
*Using samples from the present study.

R2 0.68 0.62 0.79 0.86 0.45

0.70 0.50 0.88 0.86 0.50

1382

Making Sense of Carbonate Pore Systems

Table 2. New Porosity Classification System* Pore Type Interparticle Pore Size Micropores (10 50 mm) Mesopores (50 100 mm) Macropores ( >100 mm) Intercrystalline Micropores (10 20 mm) Mesopores (20 60 mm) Macropores ( >60 mm) Intraparticle Moldic Vuggy Mudstone microporosity Pore Distribution Uniform Patchy Uniform Patchy Uniform Patchy Uniform Patchy Uniform Patchy Uniform Patchy Pore Fabric Interparticle, uniform micropores Interparticle, patchy micropores Interparticle, uniform mesopores Interparticle, patchy mesopores Interparticle, uniform macropores Interparticle, patchy macropores Intercrystalline, uniform micropores Intercrystalline, patchy micropores Intercrystalline, uniform mesopores Intercrystalline, patchy mesopores Intercrystalline, uniform macropores Intercrystalline, patchy macropores Intraparticle Moldic micropores Moldic macropores Vuggy Tertiary chalk Cretaceous chalk Chalky micropores, uniform Chalky micropores, patchy R2 0.88 0.79 0.86 0.85 0.88 0.87 0.92 0.79 0.94 0.92 0.80 0.86 0.86 0.90 0.50 0.80 0.81 0.96

Micropores ( <10 20 mm) Macropores ( >20 30 mm) Micropores ( <10 mm) Uniform Patchy

*Partly based on Choquette and Pray (1970) and Lucia (1983, 1995, 1999). Porosity-permeability coefficients of determination (R 2) are based on samples from the present study.

The final step in the development of the new classification scheme was to incorporate the observation of Lucia (1983) that pore size is a primary factor in understanding porosity-permeability relationships. Whereas Lucia recognized the control of pore size on porositypermeability relationships, his classification scheme uses grain size (of particles or crystals) instead of pore size as the primary means of pore class division (i.e., his interparticle classes 1, 2, and 3). In this study, direct description of pore size was used (instead of particle or crystal size). This is in part because of the range of sorting observed in many samples that made classification of particle size problematic and partly because later cements commonly act to occlude pore space, making the relation of pore size to grain size indirect. The result of incorporating pore size into the system was a further increase in the R 2. During the evolution of the pore-type classification described herein, subdivision of the classes was made until a high coefficient of correlation was realized. This was a primary goal behind the project, instead of producing a genetic scheme that attempts to relate classes to depositional and diagenetic processes (e.g., Ahr and Hammel, 1999; Ahr et al., 2005). Given the above discussion, it follows that the pore system proposed here is

texturally derived mostly from Choquette and Pray (1970) and incorporates the pore-size differentiation that is underpinned by Lucias (1983) work. Important modifications include the addition of uniform and patchy porosity distribution and the incorporation of mudstone microporosity made to achieve high R 2 values. Mathematical analysis of the different pore classes reveals that some are described by linear relationships, whereas others show polynomial and logarithmic bestfit trends with a relatively high coefficient of correlation (Tables 1, 2). This is in contrast to the logarithmic boundaries to pore class fields described by Lucia (1983).

SAMPLES AND METHODS Approximately 3000 plug samples from a wide range of sedimentary basins and geological ages are included in the database of this study, including (in order of numeric representation in the database)
 

Maastrichtian and Danian well data, North Sea Upper Devonian and Lower Permian well data, Timan-Pechora Basin, Russia Lny 1383

      

Upper Carboniferous and Lower Permian well data, Barents Sea Cretaceous outcrops and well data, western Iran Tertiary well data, southeastern Asia Eocene well data, Libya Paleocene and Miocene outcrops, Majella, Italy Miocene outcrops, Almeria region, southeastern Spain Cretaceous well data, offshore Namibia

could be clearly defined by naked-eye observation of plugs, rock slabs, or thin section. Therefore, porosity should be classified as having a patchy distribution only when the patchy distribution can be seen without magnification. The absence of fractures was confirmed by thinsection observations and, in most cases, by visual examination of core plugs.

Most plugs have a diameter of 1 in. (2.5 cm), but some 1.5-in. (12.7-cm)-diameter plugs (especially from vuggy pore systems) are also included. Only plugs that fulfill all of the following selection criteria are included in the database: (1) thin sections are available from the plugs; (2) each plug has a dominance of only one pore type; (3) plugs are unfractured; and (4) porosity and permeability measurements are available for the plug. The first selection criterion ensures the accurate definition of pore types, thereby making it possible to directly relate pore type to measured porosity and permeability. Selection criteria 2 and 3 strongly reduce (but do not eliminate) the risk of reservoir parameters being the result of pore-type mixing. Thin-section porosity was impregnated with bluedyed epoxy for visualization of pore types. Helium porosity and air permeability (Klinkenberg corrected) were measured on horizontal and a few vertical plugs. The definition of pore-size categories (micro-, meso-, and macropores) was initially based on visual comparison and grouping of thin sections and later quantified from selected samples using both two- and three-dimensional (2-D and 3-D) techniques. The diameter of the volumetrically most common pore sizes was measured from thin sections (2-D) using a petrographic microscope. Three-dimensional pore-size distributions were estimated from core plugs using scanning electron microscopy image-processing techniques on slices cut at 10-mm increments. Interpolation was applied between the slices. Porosity distribution (uniform and patchy) was visually defined from both thin sections and core plugs. A thin section represents only a thin slice through a plug, and some thin sections therefore show uniform pore distribution in larger scale patchy pore networks. Porosity and permeability were measured on the entire plug, and it is the pore distribution in the plug instead of the thin section that is important in relating pore types to reservoir properties. With few exceptions, the patchy porosity distribution was on a scale where it 1384 Making Sense of Carbonate Pore Systems

A REVISED PORE-TYPE CLASSIFICATION SYSTEM As discussed, the new pore-type classification system developed here is based on the Choquette and Pray (1970) classification system, but incorporates elements from the Lucia (1983, 1995, 1999) system. However, the new system also introduces new elements that are critical for predicting reservoir properties. The main differences between the new carbonate pore system and those of Choquette and Pray (1970) and Lucia (1983, 1995, 1999) are as follows:
 

 

Porosity distribution is a major new element in the classification. Lucias subdivision of interparticle porosity has been partly incorporated into the new classification system, but is now based on pore size instead of grain size and sorting. Lucias three interparticle pore-type classes and Choquette and Prays interparticle and intercrystalline porosity types have been subdivided into 12 new classes (6 interparticle and 6 intercrystalline). Micromoldic and macromoldic pores are differentiated. A new pore-type category, consisting of four pore types, is introduced: mudstone microporosity.

Porosity-permeability crossplots for data sets used in this study yield significantly higher coefficients of determination (R 2) with the new classification system than with the classification systems of Choquette and Pray (1970) and Lucia (1983, 1995, 1999) (Tables 1, 2). The new classification system combines sedimentologic and diagenetic features with flow-related properties, and reservoir-critical parameters can thus be predicted using sedimentologic and diagenetic models. The new classification system is based on three main elements: pore type, pore size, and pore distribution. Age is an important factor for some of the mudstone micropore classes.

Pore Type Six main pore types are identified: interparticle, intercrystalline, vuggy, intraparticle, moldic, and mudstone microporosity (Table 2). The first five pore types are almost identical to those defined by Choquette and Pray (1970), whereas the last one is new. Ten additional pore types defined by Choquette and Pray are valid pore types to consider. However, because of insufficient data for fenestral, shelter, boring, burrow, and shrinkage porosity and/or analytical considerations for fracture, channel, cavern, growth framework, and breccia porosity, these pore types were not included in the new system. Insufficient data probably suggest that pore types are of minor significance in reservoirs examined to date. Analytical considerations are mostly related to high pore/plug size ratio, i.e., plug sizes are too small to adequately represent the pore system.

Pore Size Lucia (1983, 1995, 1999) realized that pore-size distribution controls permeability and is related to rock fabric. Therefore, he used average particle size and sorting to differentiate between different interparticle pore-type classes. The term particle was used as a general term for grains (multicrystalline particles) and crystals (single-crystal particles) (Lucia, 1983). The studied data set shows a relatively good relation between intercrystalline pore size and crystal size. However, the relation between intergrain pore size and particle size is sometimes poor, which, as pointed out by Lucia (1995, 1999), is partly because of poor sorting. Pore sizes in packstones may be controlled by allochem size and interparticle mud, resulting in a wide range of pore sizes. At the same average particle size, pore size is sometimes controlled by interparticle mud (pores between mud particles) and, in other cases, by the larger grains (pores between allochems). Average particle size is also difficult to define because it may be defined by the volume or by the number of grains. Lucia (1995, 1999) discussed sorting in relation to mud infilling of interparticle pore space by differentiating between grainstone, mud-lean packstones (graindominated packstones), and mud-rich packstones (muddominated packstones). However, this classification system does not consider the important effect allochem sorting has on pore size. Moderately to poorly sorted

grainstones and mud-lean packstones have a wide range of grain sizes even within the same grain-size class, resulting in closer packing of grains and smaller pore sizes. Another important effect on pore size that has not been covered by the Lucia (1983, 1995, 1999) classification system is cement. Both particle size and sorting are unaffected by interparticle cement, but the cement will reduce pore size and pore throats. Samples with similar particle size and sorting thus may have a significant spread in pore size and permeability-porosity relationship, depending on the cement volume and morphology. The foundation of Lucias (1983, 1995, 1999) classification system, that pore-size distribution controls permeability and is related to rock fabric, is therefore missing an important element. Based on these considerations, pore size instead of particle size has been applied in the new pore-type classification system. Pore-size differentiation provided a better coefficient of determination (R 2) in the porosity-permeability crossplots than particle-size differentiation. Pore-size differentiation was performed by visual examination of thin sections and has been applied to interparticle, intercrystalline, and moldic pore types (Table 2). Interparticle and intercrystalline pores have been visually subdivided into micro-, meso-, and macropores, whereas moldic pores have been subdivided into micro- and macropores. Pore sizes of the different classes vary with pore type (Table 2) and are not based on other published definitions. Interparticle pore diameters and size distributions were measured on the reference samples shown in Figure 1a c. Micropores have a dominant pore diameter in the 1060-mm range (70% of the pores). Mesopores have a dominant pore diameter of 40 100 mm, although approximately 30% of the pores are in the 100 300-mm range. Macropores are generally larger than 100 mm in diameter (approximately 75% of the pores). Interparticle pore-size groups may thus be defined by 10 50 (micropores), 50 100 (mesopores), and greater than 100 mm (macropores). Porosity with dominant pore diameters less than 10 mm is classified as mudstone microporosity. Intercrystalline micropores are commonly 10 20 mm in diameter, whereas mesopore diameters mostly are in the 20 60-mm range. Intercrystalline macropores have diameters larger than 60 mm. Moldic micropore diameters are typically less than 1020 mm, although they occasionally can be larger. Moldic macropores are larger than 2030 mm. Lny 1385

Figure 1. Thin-section micrographs showing samples with a predominance of interparticle porosity, some of which are solution enlarged. Minor intraparticle and moldic pores are present, but interparticle pores predominate: (a) microporosity (1050 mm pore diameter) with uniform porosity distribution, f = 17.6%, k = 0.84 md; (b) mesoporosity (50100 mm pore diameter) with uniform porosity distribution, f = 19.3%, k = 9.47 md; (c) macroporosity (>100 mm pore diameter) with uniform porosity distribution, f = 15.3%, k = 132 md; and (d) macroporosity (>100 mm pore diameter) with uniform porosity distribution and porelining calcite cement, f = 9.7%, k = 0.465 md. All photomicrographs were taken under planepolarized light.

Pore Distribution Porosity distribution is a new element in pore-type classification and has a significant effect on porositypermeability relationships (also noted by Lucia et al., 2004a, b). The distribution of interparticle pores, intercrystalline pores, and mudstone micropores has been visually classified as either uniform or patchy. At similar porosities, a patchy porosity distribution is observed to yield significantly higher permeability than uniform porosity distribution. The reason for this is that the porosity is concentrated over a smaller vol1386 Making Sense of Carbonate Pore Systems

ume and the pore system is better connected than for an equivalent, uniformly distributed pore volume. Furthermore, a patchy porosity distribution is often related to secondary dissolution with slight corrosion of pore throats, and this process also tends to favor connected pores. The effect of patchy porosity on porosity-permeability relationships may be illustrated by a porous, sucrosic dolomite with anhydrite nodules. Core plugs from this dolomite may show uniform porosity distribution or varying degrees of patchiness, depending on the location of the plugs. Plugs with uniform porosity

Figure 1. Continued.

distribution (no anhydrite) will have higher porosity than plugs with patchy porosity distribution (with anhydrite). The effect on permeability, however, is minor because pore-throat diameters are the same (independent of anhydrite). The only effect on permeability may be a slight increase in tortuosity. The permeability is thus nearly uniform in both examples, whereas the porosity is reduced in the anhydritic sample. Obviously, there may be various degrees of porosity patchiness, and ideally, there would be a relation between patchiness and porosity-permeability. Patchiness could probably be measured in some way from

thin sections. However, it is the patchiness in the plug and not the thin section that controls the porositypermeability relationship. In plugs with only two mineral phases, where grains and patchily distributed cement have significantly different grain densities, the grain density could be used as a measure of patchiness. An attempt to use grain density as a measure of patchiness in this study was not successful because very few samples had the appropriate mineral combinations. Controls on porosity distribution vary for different pore types and are therefore discussed under the respective pore-type description sections. Lny 1387

PORE-TYPE DEFINITIONS AND ORIGINS Interparticle Porosity Choquette and Pray (1970) defined interparticle porosity as porosity occurring between grains (intergrain). Lucia (1983) extended the term interparticle to also include pore spaces between crystals (intercrystal). This redefinition thus included both the interparticle and intercrystalline porosity types of Choquette and Pray (1970). However, the results of this study show that the petrophysical properties of intergrain and intercrystal pores are different. Choquette and Prays definition of interparticle porosity has therefore been applied in the new pore-type classification system presented in this article. Solution-enlarged interparticle pores were included in the definition when the pores interparticle origin could be clearly defined. Interparticle pores are commonly primary in origin, but may also be formed by secondary dissolution of cements or matrix. To achieve a good porosity-permeability relationship, six natural interparticle pore-type classes are apparent, based on pore size and pore distribution (Table 2; Figures 1, 2). Three pore-type classes were defined based on pore size (micro-, meso-, and macropores; 1050, 50 100, and >100 mm, respectively), each being subdivided on the basis of a uniform or patchy pore distribution. In the studied data set, a patchy, interparticle pore distribution is commonly related to a diagenetic overprint. Diagenesis may include the selective dissolution of interparticle matrix and cement, and patchy cementation of both primary and secondary interparticle pores. Less commonly, a patchy pore distribution is primary in origin and is related to very large grains in poorly or bimodally sorted sediment (typically oncoid-rich samples in the studied data set). In the studied data set, selective dissolution of cements and matrix is strongly controlled by fluidmigration pathways (e.g., fractures and permeable grainstones), commonly yielding a patchy pore distribution in the early stages of dissolution. Patchy cement distribution is commonly controlled by selective precipitation of syntaxial calcite cement overgrowths on singlecrystal grains (typically echinoderm fragments), but also occurs where noncarbonate minerals precipitate within carbonate host sediments. Interparticle pores are normally associated with medium- to high-energy depositional settings in the studied data set, such as rimmed, platform-margin shoals, distally steepened ramp-margin shoals, inner-ramp fringing shoals, middle-ramp barrier shoals, local platform1388 Making Sense of Carbonate Pore Systems

interior shoals, gravity-driven flow deposits, beaches, wash-over fans, and others. Micropores and mesopores occur within grain-supported textures composed of extremely small bioclastic fragments, where larger interparticle pores have been partially occluded by cement, or within poorly sorted grainstones. Micropores also occur within recrystallized mud of mud-lean packstones. Macropores are most common in moderately to wellsorted, high-energy grainstones. Intercrystalline Porosity Intercrystalline porosity is the porosity between crystals that may be of either primary or secondary origin (Choquette and Pray, 1970). All intercrystalline pores included in the studied data set are secondary in origin and occur between crystals that have grown more or less in place by calcite recrystallization or dolomitization. Thus, only the pores of secondary origin are considered in the following discussion. In grain-rich dolostones, porosity may be classified as either intercrystalline or interparticle, depending on the dominant control on pore structure. When depositional grains (allochems) are the dominant control, the porosity is here classified as interparticle. Where pore structures are controlled by dolomite crystals, they are instead classified as intercrystalline. Based on these criteria, dolograinstones in the studied data set commonly (but not always) have interparticle porosity, whereas dolopackstones commonly have intercrystalline porosity. Intercrystalline pores are described by six poretype classes based on pore size and pore distribution, each with a distinct porosity-permeability relationship (Table 2; Figure 3). Three pore-type classes are defined based on pore size (micro-, meso-, and macropores; 10 20, 2060, and >60 mm, respectively). These are, in turn, subdivided into uniform or patchy pore distribution to optimize porosity-permeability best-fit trends. Crystals are commonly of uniform size in a sample, resulting in a typically good correlation between crystal size and pore size. In some cases, notably within microporous, argillaceous dolomites, the pore size is smaller than expected from the dolomite crystal size alone because clays partially fill the porosity. Patchy intercrystalline pore distribution is related to patchy cementation and, less commonly, to the spatial arrangement of crystals (i.e., texture). The mineralogy of patchy cements and host rock is commonly different. When mineralogies are identical, the cements tend to have more uniform distribution because more nucleation points are available during cement precipitation.

Figure 2. Example of patchy interparticle mesoporosity (50 100 mm pore diameter) in Devonian rocks from Russia. Left: core slab showing patchy pore distribution related to differential dissolution. Tight, calcite-cemented areas are gray; porous, oil-stained areas are brown. Right: thin-section photomicrographs showing patchy interparticle mesopores related to differential dissolution of cements. Lower photomicrograph is a close-up showing patchy porosity at a larger scale. Note the abundance of rhombohedral pore outlines, which are indicative of dissolution. Calcite cements are white. Both photomicrographs were taken under plane-polarized light.

The most common patchy cement mineralogies in the studied data set are chert and gypsum-anhydrite and, less commonly, calcite. The host rock is commonly dolomite. Patchy intercrystalline cementation is commonly controlled by depositional setting, but may also be controlled by fluid-migration pathways and reprecipitation of cements during later diagenesis. Restricted to hypersaline settings in an arid climate favor evaporite precipitation and patchy pore distribution in dolomites. Patchy intercrystalline pore distribution related to texture is, in the studied data set, commonly associated with early diagenetic, finely crystalline dolomites. Many

of the dolomite crystals are arranged in millimeter- to centimeter-size, relatively tight clusters, where each cluster is composed of several dolomitized grains and intergrain mud. Outside these clusters, dolomite grains are commonly still tight, but there is high intercrystalline porosity in dolomitized mud between the grains. Mudstone Microporosity Mudstone micropores have extremely small pore sizes, commonly a few micrometers in diameter. Individual pores cannot be seen with a standard petrographic Lny 1389

Figure 3. Thin-section photomicrographs showing examples of intercrystalline porosity: (a) microporosity (10 20 mm pore diameter) with uniform porosity distribution, f = 18.1%, k = 0.476 md; (b) mesoporosity (20 60 mm pore diameter) with uniform porosity distribution, f = 19.3%, k = 4.93 md; (c) macroporosity (> 60 mm pore diameter) with uniform porosity distribution, f = 11.0%, k = 16.7 md; and (d) mesoporosity (2060 mm pore diameter) with patchy porosity distribution related to gypsum cementation (white); f = 12.1%, k = 16.0 md. All micrographs were taken under planepolarized light.

microscope. However, using blue-dyed epoxy, the presence of these pores can commonly be discerned by a weak bluish stain to the thin section, especially when the porosity is high (Figure 4). Mudstone microporosity may be defined as interparticle or intercrystalline porosity. However, because of the extremely small pore sizes and variable pore structure (interparticle or intercrystalline), these pores were classified as a separate pore-type class. Mudstone microporosity includes both true chalks and chalky microporosity. Chalk micropores are pri1390 Making Sense of Carbonate Pore Systems

mary in origin and occur between grains of planktonic calcareous algae (coccospheres) or their component crystal plates (coccoliths). Chalky micropores are not related to chalk, but the pore structure is similar. These pores occur between recrystallized mud particles and may be formed either during early meteoric leaching or deeper burial diagenesis (Pittman, 1971; Budd, 1989; Moshier, 1989). The pores typically form in low-energy, muddy, platform-interior facies. Four natural classes of mudstone microporosity exist in the database: (1) Tertiary chalk, (2) Cretaceous

Figure 3. Continued.

chalk, (3) chalky micropores with uniform distribution, and (4) chalky micropores with patchy distribution (Figure 4). The distinction between Tertiary and Cretaceous chalks is important because a general decrease in the size of calcareous nannoplankton across the CretaceousTertiary boundary (Macleod et al., 1997) corresponds to a decrease in reservoir quality in Tertiary chalks (Hardman, 1983). Chalks with patchy porosity distribution have been identified, especially related to patchy chert cementation, but no porosity and permeability data are available in the studied data set.

Moldic Porosity Moldic pores are secondary pores formed by the selective, complete, or partial dissolution and recrystallization of grains or crystals. This definition is slightly modified from Choquette and Pray (1970) by including pores formed by partial dissolution and recrystallization. A distinctive difference in solubility between grains and/or crystals and the surrounding matrix is commonly needed and is commonly related to mineralogical differences (Moore, 2001). Lny 1391

Figure 4. Thin-section photomicrographs showing examples of mudstone microporosity (<10 mm pore diameter): (a) chalk, f = 25.4%, k = 0.734 md; (b) chalky microporosity, f = 4.9%, k = 0.01 md; and (c) patchily distributed chalky microporosity, no porosity, or permeability measurements are available. Individual pores are too small to be seen, but porosity can be discerned because of impregnation with blue-dyed epoxy. The micrographs were taken under planepolarized light.

1392

Making Sense of Carbonate Pore Systems

Figure 5. Thin-section photomicrographs showing examples of moldic porosity: (a) microporosity, f = 21.7%, k = 3.45 md; and (b) macroporosity, f = 24.74%, k = 35.2 md. Some of the moldic micropores in (a) are marked by m. Both micrographs were taken under plane-polarized light.

Moldic pores are divided into two natural classes on the basis of intramold pore sizes: moldic micropores (Figure 5a) and moldic macropores (Figure 5b). Moldic micropores occur as small intramold pores (typically less than 1020 mm in diameter) developed between intramold mud particles or small crystals (Figure 5a). The pores are commonly associated with partially dissolved or recrystallized peloids, ooids, and foraminifera with microgranular structure. Moldic micropores resemble mudstone micropores in structure, but the porosity distribution is different, resulting in

different porosity-permeability relationships. Mudstone micropore distribution is controlled by micritic matrix, whereas moldic micropores are controlled by allochem distribution. Moldic macropores may have relatively small intermold pore-throat diameters, although individual molds can be relatively large (>2030 mm) (Figure 5b). The pores contain no or subordinate fine crystals and mud and are commonly formed through dissolution. Both diagenesis and depositional setting control the macromoldic porosity distribution. Mineralogically unstable Lny 1393

Figure 6. Thin-section photomicrograph showing an example of intraparticle porosity in fusulinid foraminifera. f = 16.0%, k = 1.25 md. The micrograph was taken under planepolarized light.

grains, e.g., aragonite and evaporite minerals, are commonly more susceptible to dissolution than the surrounding calcitic or dolomitic matrix (Matthews, 1974; Moore, 1989; Moore, 2001). Aragonite grains are common in euphotic skeletal assemblages, and shallow water depth, warm climate, and low nutrient supply are important factors for the generation of this assemblage (see James, 1997, and references therein). Partially dolomitized sequences are also susceptible to secondary porosity development, including moldic macropore formation, because of contrasts in mineralogical stability (calcite versus dolomite) (Moore, 2001).

pores of diagenetic origin, commonly not enclosed by skeletal walls (moldic pores)

The distribution of intraparticle pores is fauna dependent. Some of the more common faunal constituents that exhibit this pore type in the studied data set are foraminifera, gastropods, bryozoans, corals, and calcispheres.

Vuggy Porosity The definition of vuggy porosity in this article follows the definition of Choquette and Pray (1970). Vuggy pores are secondary solution pores that are not fabric selective (i.e., the pores cut across grains and/or cement boundaries) (Figure 7). The pores are of irregular size and shape and may or may not be interconnected. Many vugs are solution-enlarged molds where the outlines of the precursor grains are poorly defined. Lucia (1983, 1995) defined vuggy porosity as pore space significantly larger than or within particles. He distinguished between separate vugs that typically are fabric selective in origin and touching vugs that typically are nonfabric selective in origin. Separate vugs are connected only through interparticle or intercrystalline porosity and include intraparticle, shelter, and moldic porosity. Touching vugs include interconnected larger cavities, channels, fenestrae, and fractures. In this study, the vuggy pore definition of Choquette and Pray (1970) has been applied, and the vuggy

Intraparticle Porosity Intraparticle pores are pore spaces occurring within grains (Figure 6), either of primary origin or formed through the decay of organic material in carbonate skeletons. A skeletal wall will therefore enclose, at least partly, most intraparticle porosity. The definition of intraparticle porosity is slightly modified from Choquette and Pray (1970) by not including porosity related to dissolution (this porosity is part of the moldic pore system in the classification system presented here). The reason for this modification is to differentiate between


pores enclosed by skeletal walls (at least partly), mostly of primary origin (intraparticle pores) Making Sense of Carbonate Pore Systems

1394

Vuggy porosity (as defined by Choquette and Pray, 1970) is formed by the dissolution of cement, matrix, and grains. This typically occurs under the influence of near-surface meteoric waters (Loucks and Handford, 1992; Saller et al., 1994), but may also be related to deep-burial fluids (Moore and Druckman, 1981; Choquette and James, 1987; Moore and Heydari, 1993). Meteoric diagenesis is commonly associated with sea level lowstands and subaerial exposure surfaces in humid climates (Loucks and Handford, 1992; Saller et al., 1994). Deep-burial dissolution may be related to hydrocarbon maturation and shale dewatering (Moore, 1989). Fluid-migration pathways, such as fractures and faults, are the main controls on vuggy porosity distribution during deep burial.

POROSITY-PERMEABILITY RELATIONSHIPS Porosity-permeability crossplots are used for predicting permeability from porosity or vice versa. Borehole porosity can be estimated from wire-line logs, but apart from the nuclear magnetic resonance log, there is no log that measures permeability. Therefore, it is common practice to establish porosity-permeability relationships through core analysis and then estimate permeability where cores do not exist using porosity logs and corederived porosity-permeability relationships. This method works well when there is a simple relationship between porosity and permeability, as in many sandstone reservoirs. In carbonate reservoirs, however, the porosity-permeability relationship is very complex because of the great variability of pore types. Porosity cutoffs are applied in reservoir evaluation to exclude porosity that is thought to have insufficient permeability to contribute to flow. The cutoff is commonly defined from porosity-permeability crossplots by defining a critical permeability. Application of porosity cutoffs is an easy and quick method for calculating reserves, and cutoffs give a simple representation of the pore-type control on reservoir properties. However, the validity of using a single porosity cutoff value is questionable when multiple pore types are present because different pore types may be associated with different permeabilities for the same porosity value. This is indicated by the data presented here. Porosity-permeability crossplots were made for each of the 20 pore-type classes defined in this article (Figures 8, 9). If the pore types are known or can be predicted from sedimentologic and/or diagenetic models, Lny 1395

Figure 7. Core slabs showing vuggy porosity. Plug-derived porosities and permeabilities are strongly dependent on the location of core plugs because of the large vug sizes. Representative measurements will require whole-core analysis.

pores are therefore nonfabric selective in origin. The vuggy pores mostly fall into Lucias (1983, 1995) touchingvug pore system, although in some low-porosity samples, the vugs appear to be nontouching and fall into Lucias separate-vug pore system. For many samples, it is difficult to tell if the vugs are touching in 3-D, and one is tempted to classify the sample based on the petrophysical properties. This exemplifies the difficulties in applying the Lucia vuggy pore system classification.

Figure 8. Porosity-permeability crossplots for interparticle and intercrystalline pores: (a) interparticle micropores (10 50 mm pore diameter); (b) interparticle mesopores (50 100 mm pore diameter); (c) interparticle macropores (>100 mm pore diameter); (d) intercrystalline micropores (10 20 mm pore diameter); (e) intercrystalline mesopores (20 60 mm pore diameter); and (f) intercrystalline macropores ( >60 mm pore diameter). Note the significant differences in porositypermeability relationships related to porosity distribution and poresize variations. Porosity cutoffs characterize these differences. Cutoffs are derived from inverted axes (permeability on x-axis). Circled data points in (c) are from samples with pore-lining cements, which probably explains the lower permeabilityto-porosity ratio. R 2 = coefficient of determination; n = number of samples.

1396

Making Sense of Carbonate Pore Systems

Figure 8. Continued.

Lny

1397

Figure 9. Porosity-permeability crossplots for: (a) chalk, (b) chalky micropores, (c) moldic pores, (d) intraparticle pores, and (e) vuggy pores. Few samples were available with chalky microporosity. Note the difference in porosity-permeability for Cretaceous versus Tertiary chalks (a). Porosity cutoffs are derived from inverted axes (permeability on x-axis). R 2 = coefficient of determination; n = number of samples.

1398

Making Sense of Carbonate Pore Systems

Figure 9. Continued.

permeability can be predicted from wire-line-log porosities using the crossplots. Furthermore, porosity cutoffs can be more accurately defined. Porosity-permeability is traditionally plotted with porosity along the x-axis, which is suitable for estimating permeability by a regression of permeability on porosity. However, to estimate the porosity from a given permeability (e.g., to estimate a porosity cutoff) by a regression of porosity on permeability, permeability has to be plotted along the x-axis if the usual formulae are to be employed (Jensen et al., 2000). All plots presented here are plotted with porosity along the x-axis. However, the given that porosity cutoffs are derived

from plots with inverted axes (porosity on the y-axis), permeability is plotted as a logarithmic value (log k). The scatter of data points around the porositypermeability trendlines can be the result of several factors: Although plugs dominated by only one pore type have been selected, most samples do have other pore types that may affect porosity and permeability.  Pore types observed in thin sections may not be representative of the entire plug, although this has been minimized through the visual evaluation of the plugs.


Lny

1399

Three pore-size groups have been defined, whereas in reality, there is a continuous range of pore sizes.  Various degrees of porosity patchiness exist.  Analytical error, especially at lower permeabilities (repeat measurements of some plugs show minor differences in porosity and permeability).


Interparticle Pore-Type Classes The best relation between porosity (f) and log k for all interparticle pore types is a logarithmic trendline (Figure 8a c). R 2 (coefficient of determination) ranges between 0.79 and 0.88. The effect of both pore size and porosity distribution can be clearly seen in the figures, with a significant increase in k/f when pore size increases and/or porosity distribution becomes patchier. If 1 md is used as the critical flow parameter, the porosity cutoff for interparticle microporosity (10 50-mm pore diameter) is reduced from 15.3% at uniform porosity distribution to 8.1%, when the porosity is patchily distributed. Similarly, for meso- (50 100-mm pore diameter) and macroporosity (>100-mm pore diameter), the porosity cutoff is reduced from 11.1 to 5.8 and 8.2 to 5.5%, respectively, for the corresponding 1-md permeability. Two of the studied samples (data points encircled in Figure 8c) have isopachous, pore-lining cements (Figure 1d) and a lower permeability than expected for the pore type and porosity values. These cements are typical of early-marine diagenesis in high-energy beach rock and hardgrounds (e.g., Moore, 1977; James and Choquette, 1990) and are thought to have occluded pore throats and reduced permeability while having only a minor effect on porosity (e.g., Moore, 2001; Budd, 2002). The effect of interparticle pore-lining cements on porosity-permeability relationships is also documented from siliciclastic sediments (Aase et al., 1996; Bloch et al., 2002) and affects fine-grained sediments more than coarse-grained sediments (Bloch et al., 2002). Interparticle pores partially cemented by porelining cements are thus expected to have a lower k/f than seen in Figure 8a c.

uniformly distributed mesoporosity, which shows a clear linear trend. Why this pore type shows a different trendline relation is unclear. R 2 ranges between 0.79 and 0.94. No data are available for patchily distributed macroporosity. As for interparticle pores, pore size and porosity distribution have a clear control on k/f (Figure 8d f). A significant increase in k/f exists when pore size increases and/or porosity distribution becomes patchier. If 1 md is used as the critical flow parameter, the porosity cutoff for intercrystalline microporosity (10 20-mm pore diameter) is reduced from 22.4% at uniform porosity distribution to 18.6%, when the porosity is patchily distributed. Similarly, for mesoporosity (20 60-mm pore diameter), the porosity cutoff is reduced from 16.9 to 9.1%. Uniformly distributed macroporosity ( > 60-mm pore diameter) has a porosity cutoff of 5.4%.

Mudstone Micropore Classes Chalk shows a second-order polynomial best fit between porosity and log k, whereas chalky micropores with uniform distribution show a linear relationship (Figure 9a, b). R 2 ranges between 0.80 and 0.96 for mudstone micropore classes. If 1 md is used as the critical flow parameter, the porosity cutoff is 25.8% for Cretaceous chalk, 31.3% for Tertiary chalk, and 27.0% for chalky microporosity with uniform distribution. The only sample with patchy distribution of chalky micropores shows a significantly higher k/f than for samples with a uniform porosity distribution (Figure 9b).

Moldic Pore-Type Classes Moldic macropores show a third-order polynomial best fit between porosity and log k (Figure 9c). Micropores show both linear and third-order polynomial best-fit trendlines, but for simplicity, the linear trendline was chosen. R 2 is 0.90 for macropores and 0.86 for micropores. If 1 md is used as the critical flow parameter, the porosity cutoff is 13.1% for macropores and 16.2% for micropores. Differences in k/f for the two pore-type classes result from a slight overall reduction in pore-throat diameters, as confirmed by mercury capillary-pressure measurements, and is probably related to the presence of intramold matrix in moldic micropores.

Intercrystalline Pore-Type Classes Intercrystalline pores are similar to interparticle pores in that they generally show a logarithmic best fit between porosity and log k (Figure 8df). An exception is the 1400 Making Sense of Carbonate Pore Systems

Intraparticle Porosity Intraparticle pores show a linear best fit between porosity and log k, with an R 2 = 0.86 (Figure 9d). If 1 md is used as the critical flow parameter, the porosity cutoff is 14.1%.

Vuggy Porosity Vuggy pores show a linear best fit between porosity and log k (Figure 9e), with a low coefficient of determination (R 2 = 0.50). If 1 md is used as the critical flow parameter, the porosity cutoff is 6.2%. The poor relationship between porosity and permeability is most likely related to a low sample-tovug size ratio (many vugs are more than 1 cm [0.4 in.] in diameter). It may also be argued that the scatter in data points results from a variable degree of connectivity between the vugs, but such an explanation could not be confirmed by the present data set.

Choquette and Pray (1970) classification systems, R 2 was reduced to 0.620.79 (Lucia) and 0.500.70 (Choquette and Pray). Chalk is overrepresented in the data set of Lucia pore-type class 3, and equal proportions of mudstone micropores, intergrain pores, and intercrystal pores would thus probably give a lower coefficient of determination for class 3 than given in Figure 10c. Several reasons why the new classification system yields a significantly better correlation between porosity and permeability exist: 1. The effect of patchy porosity distribution on the porosity-permeability relationship (Figure 8) 2. The split between interparticle and intercrystal porosity (as in Choquette and Pray, 1970) and the introduction of mudstone microporosity (the effect of this can be seen in Figure 10ac) 3. The use of pore-size differentiation instead of particle size and sorting differentiation (samples in the studied data set show widely different pore sizes within each of Lucias, 1995, 1999, interparticle classes because of the variable extent of interparticle and intercrystalline cementation and allochem sorting) To evaluate the influence of points 2 and 3 on the porosity-permeability relationship, samples with patchy porosity distribution were excluded from the data set. Using the Lucia (1983, 1995, 1999) classification system, this gave R 2 = 0.77 for interparticle class 1, R 2 = 0.74 for interparticle class 2, and R 2 = 0.68 for interparticle class 3. This is a significant improvement for class 2, but all classes still have a considerably lower coefficient of determination than for those in the new classification system (Figures 8, 9a, b). Intraparticle, moldic (both micro- and macromoldic), and some of the low-porosity vuggy pores in the new classification system are classified as separate-vug pores in the Lucia (1983, 1995, 1999) classification system. The separate-vug pore system has an R 2 = 0.86 (Figure 10d), which is comparable to R 2 values for intraparticle pores (R 2 = 0.86), moldic micropores (R 2 = 0.86), and moldic macropores (R 2 = 0.90). However, as seen in Figure 9c and d, different pore types may have important differences in porosity cutoffs for the same permeability. In some hydrocarbon fields, this change in cutoff may have an important effect on net/gross ratios. Vuggy pores in the new classification system, except for some low-porosity samples, are classified as touching-vug pores in the Lucia (1983, 1995, 1999) classification system. The classification systems give Lny 1401

Mixed Pore Types As discussed previously, the data presented here include samples that are strongly dominated by only one pore type. The selected samples were thus chosen from a significantly larger data set, many of which have a mixed pore-type composition. Samples with mixed pore types plot between the end members, but tend to be biased toward the pore type of highest permeability.

Comparison to Other Classification Systems All samples in the present database were classified in accordance with the Choquette and Pray (1970), Lucia (1983, 1995, 1999), and the new classification system presented here to evaluate which system has the best predictive properties for porosity and permeability. The coefficient of determination (R 2) was used as a measure, with higher R 2 indicating better correlation between porosity and permeability. The results are presented in Tables 1 and 2 and Figure 10. Interparticle pores, intercrystalline pores, and mudstone micropore systems of the new classification system give R 2 = 0.790.96 in porosity-permeability crossplots (Figures 8, 9a, b). Classified according to the Lucia (1983, 1995, 1999) (Figure 10ac) and

Figure 10. (ac) Data from pore systems of this study plotted as interparticle classes 13 of Lucia (1983, 1995, 1999). In these plots, the complete spread of data corresponds to the coefficient of determination shown in the back text boxes (R 2 values of 0.79, 0.62, and 0.68, in plots ac, respectively). The blue and red data points in (a c) demonstrate the effect of differentiating the data into intercrystalline and intergrain pore types (following the approach used by Choquette and Pray, 1970). The effect on the coefficients of determination is shown in red and blue text boxes. The plots show that the distinction between intercrystalline and intergrain pore types varies as a function of grain size and porosity, the greatest differentiation being within classes 2 and 3. The figure in (c) shows the need for incorporating mudstone microporosity pore types into any classification scheme. The figure in (d) plots both the touching-vug and separate-vug pore types of Lucia. These pore types correspond to intraparticle, moldic (both micro- and macromoldic) and vuggy pores of this study. For all plots, the trendline yielding the highest R 2 value was chosen.

comparable coefficients of determination (R 2 = 0.50 in the new classification system; R 2 = 0.45 in the Lucia classification system).

IMPLICATION FOR RESERVE ESTIMATES Carbonate reservoirs commonly have a wide range of porosities, and net/gross (N/G) is commonly calculated by applying a porosity cutoff. Porosity cutoff ranges significantly for different pore types in the data set described here, from 5.4 to 31.3% at 1-md permeability. Pore-type definition is thus crucial for hydrocarbon reserve calculation. A Devonian hydrocarbon field in Russia will be used to illustrate the effect of pore-type definition on 1402 Making Sense of Carbonate Pore Systems

reserve calculation. For simplicity, the reserve estimates were calculated based on data from a single well (regional variation across the field was not considered). Furthermore, porosity and permeability data were derived from core plugs, assuming that the plugs were representative for the reservoir section. Reserves were calculated by applying a variable porosity cutoff, defined bed by bed, based on pore type. The reservoir section is dominated by interparticle micro- and mesoporosity. Core and thin-section studies show that the porosity has a patchy distribution and is related to the dissolution of interparticle cement. Failure to recognize the patchy distribution of the interparticle porosity may have a severe effect on reserve estimates because porosities commonly are above the cutoff values for patchily distributed porosity but

Figure 10. Continued.

below the cutoff for uniformly distributed porosity, using a 1-md permeability cutoff (Figure 11). Calculations show that estimated reserves are 370% higher for the patchy pore distribution compared to the uniform pore distribution (Figure 11). This example illustrates the importance of poretype evaluation in reserve estimation. In many cases, pore-type evaluation can be one of the most important factors in the economic evaluation of a prospect or field.

CONCLUSIONS A new pore-type classification system that better combines sedimentologic and diagenetic features and flowrelated properties has been developed. This system

significantly improves porosity to permeability relationships for individual pore types compared to previous classification systems and yields coefficients of determination (R 2) mostly in the range of 0.79 0.96. Porosity cutoffs derived from porosity-permeability crossplots show significant variability among the different pore types, ranging from 5.4 to 31.3% at 1-md permeability. This has a major impact on reserve calculations because it strongly affects net/gross. A simple model run on a Devonian hydrocarbon reservoir indicates that estimated reserves may vary by several hundred percent because porosity cutoffs will likely vary by pore type. Knowing the dominant pore types and their distribution in the reservoir are thus critical in field evaluation, both in the exploration and Lny 1403

Figure 11. Pore-type control on reserve calculations in a 35-m (114-ft)thick Devonian reservoir, Russia. Vertical lines on porosity-permeability crossplot represent porosity cutoffs at 1-md permeability and are based on global trendline curves (Figure 8a, b). Marked trendlines are based on the Devonian reservoir and differ slightly from the global trendlines. The table shows differences in calculated net-to-gross and average phi for patchy versus uniform pore distributions based on coreplug data. BP = interparticle porosity, N/G = net/gross pay.

production phases. Pore type could be one of the most important parameters in the economic evaluation of a prospect or field.

REFERENCES CITED
Aase, N. E., P. A. Bjrkum, and P. H. Nadeu, 1996, The effect of grain-coating microquartz on preservation of reservoir porosity: AAPG Bulletin, v. 80, no. 10, p. 1654 1673. Ahr, W. M., and B. Hammel, 1999, Identification and mapping of flow units in carbonate reservoirs: An example from Happy Spraberry (Permian) field, Garza County, Texas, U.S.A.: Energy Exploration and Exploitation, v. 17, p. 311 334. Ahr, W. M., D. Allen, H. N. Bachman, E. A. Clerke, K. B. Gzara, J. K. Hassell, C. R. K. Murty, and R. Ramamoorthy, 2005, Confronting the carbonate conundrum: Schlumberger Oilfield Review, Spring 2005, v. 17, no. 1, p. 18 29. Archie, G. E., 1952, Classification of carbonate reservoir rocks and

petrophysical considerations: AAPG Bulletin, v. 36, no. 2, p. 278 298. Bloch, S., R. H. Lander, and L. Bonnell, 2002, Anomalously high porosity and permeability in deeply buried sandstone reservoirs: Origin and predictability: AAPG Bulletin, v. 86, no. 2, p. 301 328. Budd, D. A., 1989, Micro-rhombic calcite and microporosity in limestones A geochemical study of the Lower Cretaceous Thamama Group of the Arabian Gulf, in C. R. Handford, R. G. Loucks, and S. O. Moshier, eds., Nature and origin of microrhombic calcite and associated microporosity in carbonate strata: Sedimentary Geology, v. 63, p. 293 311. Budd, D. A., 2002, The relative roles of compaction and early cementation in the destruction of permeability in carbonate grainstones: A case study from the Paleogene of west-central Florida, U.S.A.: Journal of Sedimentary Research, v. 72, no. 1, p. 116 128. Choquette, P. W., and N. P. James, 1987, Diagenesis #12. Diagenesis in limestones 3. The deep burial environment: Geoscience Canada, v. 14, p. 3 35. Choquette, P. W., and L. C. Pray, 1970, Geologic nomenclature

1404

Making Sense of Carbonate Pore Systems

and classification of porosity in sedimentary carbonates: AAPG Bulletin, v. 54, no. 2, p. 207 250. Hardman, R. F. P., 1983, Chalk hydrocarbon reservoirs of the North Sea An introduction; Joint Association for Petroleum Exploration Courses Chalk Seminar, December 15, 1983, p. D1 D33. James, N. P., 1997, The cool-water carbonate depositional realm, in N. P. James and J. A. D. Clarke, eds., Cool-water carbonates: SEPM Special Publication 56, p. 1 20. James, N. P., and P. W. Choquette, 1990, Limestones The seafloor diagenetic environment, in I. A. McIlreath and D. W. Morrow, eds., Diagenesis: Ottawa, Ontario, Canada, Geological Association of Canada, p. 13 34. Jensen, J. L., L. W. Lake, P. W. M. Corbett, and D. J. Goggin, 2000, Statistics for petroleum engineers and geoscientists, 2d ed.: Amsterdam, The Netherlands, Elsevier Science B.V., 338 p. Loucks, R. G., and C. R. Handford, 1992, Origin and recognition of fractures, breccias, and sediment fills in paleocave-reservoir networks, in M. P. Candeleria and C. L. Reed, eds., Paleokarst, karst-related diagenesis and reservoir development: Examples from Ordovician Devonian age strata of west Texas and the mid-continent: Permian Basin Section, SEPM Publication 9233, p. 31 44. Lucia, F. J., 1983, Petrophysical parameters estimated from visual descriptions of carbonate rocks: A field classification of carbonate pore space: Journal of Petroleum Technology, v. 216, p. 221 224. Lucia, F. J., 1995, Rock-fabric/petrophysical classification of carbonate pore space for reservoir characterization: AAPG Bulletin, v. 79, no. 9, p. 1275 1300. Lucia, F. J., 1999, Carbonate reservoir characterization: Berlin, Springer-Verlag, 226 p. Lucia, F. J., J. W. Jennings, Jr., and S. C. Ruppel, 2004a, South Wasson Clear Fork reservoir modeling: The rock fabric method for constructing flow layers for fluid flow simulation (abs.): AAPG Hedberg Research Conference, March 15 18, 2004, El Paso, Texas.

Lucia, F. J., R. H. Jones, and J. W. Jennings, 2004b, Poikilotopic anhydrite enhances reservoir quality (abs.): AAPG Annual Meeting Program, v. 13, p. A88. Macleod, N., et al., 1997, The Cretaceous Tertiary biotic transition: Journal of the Geological Society (London), v. 154, p. 265 292. Matthews, R. K., 1974, A process approach to diagenesis of reefs and reef-associated limestones, in L. F. Laporte, ed., Reefs in time and space: SEPM Special Publication 18, p. 234 256. Moore, C. H., 1989, Carbonate diagenesis and porosity: New York, Elsevier, 338 p. Moore, C. H., 2001, Carbonate reservoirs: Porosity evolution and diagenesis in a sequence-stratigraphic framework: Developments in Sedimentology, v. 55, 444 p. Moore, C. H. Jr., 1977, Beach rock origin: Some geochemical, mineralogical, and petrographic considerations: Geoscience and Man, v. 18, p. 155 163. Moore, C. H., and Y. Druckman, 1981, Burial diagenesis and porosity evolution, Upper Jurassic Smackover, Arkansas and Louisiana: AAPG Bulletin, v. 65, p. 597 628. Moore, C. H., and E. Heydari, 1993, Burial diagenesis and hydrocarbon migration in platform limestones: A conceptual model based on Upper Jurassic of Gulf Coast of U.S.A., in A. D. Horbury and A. G. Robinson, eds., Diagenesis and basin development: AAPG Studies in Geology 36, p. 213 229. Moshier, S. O., 1989, Development of microporosity in a micritic limestone reservoir, Lower Cretaceous, Middle East, in C. R. Handford, R. G. Loucks, and S. O. Moshier, eds., Nature and origin of micro-rhombic calcite and associated microporosity in carbonate strata: Sedimentary Geology, v. 63, p. 217 240. Pittman, E. D., 1971, Microporosity in carbonate rocks: AAPG Bulletin, v. 55, no. 10, p. 1873 1881. Saller, A. H., D. A. Budd, and P. M. Harris, 1994, Unconformities and porosity development in carbonate strata: Ideas from a Hedberg conference: AAPG Bulletin, v. 78, p. 857 872.

Lny

1405

Вам также может понравиться