ROGER Y. ANDERSON Department of Geology, the University of New Mexico, Albuquerque, New Mexico 87106 WALTER E

Home , Bell Canyon Formation, Castile Formation

ROGER Y. ANDERSON Department of Geology, The University of New Mexico, Albuquerque, New Mexico 87106 WALTER E. DEAN, JR. Department of Geology, Syracuse University, Syracuse, New Yor\ 13210 DOUGLAS W. KIRKLAND Mobil Research and Development Corporation, Dallas, Texas 75221 HENRY I. SNIDER Department of Physical Sciences, Eastern Connecticut State College, Willimantic, Connecticut 06226

Permian Castile Varved Evaporite Sequence,

West Texas and New Mexico

ABSTRACT is a change from thinner undisturbed anhydrite laminae to thicker anhydrite laminae that Laminations in the Upper Permian evaporite generally show a secondary or penecontem- sequence in the Delaware Basin appear in the poraneous nodular character, with about 1,000 preevaporite phase of the uppermost Bell to 3,000 units between major oscillations or Canyon Formation as alternations of siltstone nodular beds. These nodular zones are correla- and organic layers. The laminations then change tive throughout the area of study and underly character and composition upward to organi- halite when it is present. The halite layers cally laminated claystone, organically laminated alternate with anhydrite laminae, are generally calcite, the calcite-laminated anhydrite typical recrystallized, and have an average thickness of the Castile Formation, and finally to the of about 3 cm. The halite beds were once west anhydrite-laminated halite of the Castile and of their present occurrence in the basin but Salado. were dissolved, leaving beds of anhydrite Laminae are correlative for distances up to breccia. The onset and cessation of halite depo- 113 km (70.2 mi) and probably throughout sition in the basin was nearly synchronous. most of the basin. Each lamina is synchronous, The Anhydrite I and II Members thicken and each couplet of two laminated components gradually across the basin from west to east, is interpreted as representing an annual layer of whereas the Halite I, II, and III Members are sedimentation—a varve. thickest in the eastern and northeastern part of The thickness of each couplet in the 260,000- the basin and thicken from southeast to north- varve sequence (a total thickness of 447.2 m, west. This distribution and the synchroneity 1467 ft) has been measured individually and indicate a departure from the classical model of recorded and provides the basis for subdividing evaporite zonation. and correlating major stratigraphic units within the basin. The uppermost 9.2 m (30.3 ft) of the INTRODUCTION Bell Canyon Formation contains about 50,850 The Castile Formation (Upper Permian) in varve couplets; the Basal Limestone Member the Delaware Basin of Texas and New Mexico of the Castile about 600; the lowermost anhy- is often cited as perhaps the best example of a drite member of the Castile (Anhydrite I) con- large deep-water evaporite deposit for which tains 38,397; Halite I, 1,063; Anhydrite II, there are no modern analogs. In addition, the 14,414; Halite II, 1,758; Anhydrite III, 46,592; Castile is well known for its remarkably dis- Halite III, 17,879; and Anhydrite IV, 54,187. tinct laminations of calcite and anhydrite, The part of the Salado collected (126.6 m) con- which are assumed by many to reflect annual tains 35,422 varve couplets. The Bell Canyon- sedimentation. Castile sequence in the cores studied is appar- The regular interlamination of salts of dif- ently continuous, with no recognizable uncon- ferent solubilities (calcite and anhydrite; anhy- formities. drite and halite) implies that depositional con- The dominant petrologic oscillation in the trols must have fluctuated in response to some Castile and Salado, other than the laminations, periodic process or event. Udden (1924) sug-

Geological Society of America Bulletin, v. 83, p. 59-86, 18 figs., January 1972 59

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 O 100 200 300 KILOMETERS

Figure 1. Index maps showing location of the cores and P. R. R. Co. UNM-Pokorney no. 1, Blk. 61, T. studied and their general relationship to features of the and P. R. R. Co. UNM-Phillips no. 1, Sec. 3, Blk. 110, Delaware Basin. Well locations as follows: UNM- PSL. David Flood no. 1 Grisham and McAlpine, Sec. Cowden no. 2, Sec. 34, Twp. 1, Blk. 62, T. and P. R. R. 42, Blk. 54, PSL. Union Oil Co.-University "37" no. Co. UNM-Cowden no. 4, Sec. 33, Twp. 2, Blk. 62, T. 4, Sec. 37, Blk. 20, Univ. Lands Survey.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 PERMIAN CASTILE EVAPORITE SEQUENCE, WEST TEXAS AND NEW MEXICO 61

gested that each calcite-anhydrite couplet thors (Anderson and Kirkland, 1966; Kirkland represented an annual increment of sediment— and Anderson, 1970) revealed that the lamina- a varve. Most investigators who have discussed tions could be correlated with great precision the Castile agree with Udden's annual inter- over the entire basin (distances up to 113 km or pretation but have been unable to agree on a 70.2 mi). The laminations continue in an unin- periodic mechanism. Adams (1944) suggested terrupted sequence from the preevaporite that new sea water was introduced by seasonal phase below the Castile upward into the breaching and sealing of a barrier. Briggs Salado Formation in a series of some 260,000 (1957) suggested that freshening due to annual laminae couplets, and provide a reference scale maximum spring tides could produce the for determining the precise volume and distri- Castile laminations. Neither of these explana- bution relation of the various components in tions seems adequate to account for the great the system. The continuous time series of lateral continuity and synchroneity of deposi- laminations also provides a basis for examining tional conditions over an area the extent of the the behavior of such a basin over much of its Delaware Basin, as demonstrated by correlation life history. ol laminae. This report deals with the broader aspects of Organically rich layers are associated with the evaporite system and considers chiefly the calcite laminae in the Castile. Richter-Bern- petrology and stratigraphic relations of the burg (1964) explained a similar association in major units in the basin. These units have been calcite-anhydrite couplets in the Permian correlated within the basin on the basis of Zechstein Formation of Germany by assuming individual laminae and indexed to a master that organic matter represented mass killing of time series. The laminations themselves are an planktonic organisms. It seems likely, however, additional focal point in the study. Also, some that laminae concentrations of organic matter interpretations are made concerning basin are the result of a periodic (annual?) increase paleogeography, solution, and other problems. in plankton productivity (blooms). The inves- The study is based partly upon sonic, elec- tigations of Carpelan (1957) and Phleger (1969) tric, and sample logs, and field observations, but have shown that evaporite basins can have mainly on a number of cores collected from levels of primary productivity greater than Culberson County, Texas, in the west-central adjacent "normal marine" environments. If part of the basin and one core from Winkler such organisms were phytoplankton, then there County, Texas, in the east-central part (see Fig. is a mechanism for calcite deposition in the 1 for locations). One of these cores (University seasonal blooming process and the attendant of New Mexico-Phillips no. 1) includes part of removal of CC>2 from the water. the Salado Formation, all of the Castile, and Organically rich layers are also associated part of the underlying Bell Canyon Formation. with the anhydrite of anhydrite-halite couplets Each section of this 5 cm (2 in.) core was where seasonal evaporation can be invoked as marked as it was removed from the core barrel the mechanism for layered halite deposition. In in order to maintain proper sequence and fact, organic or organically rich layers are com- superposition. The core was slabbed, polished, mon to all the laminae types in the Bell Can- and marked off at 5.08-cm (2-in.) intervals. yon-Castile sequence, and form a basis for the Photographs of the core were enlarged three assumption that throughout the sequence the times, and printed on strips of photographic lamination process is in tune with, if not influ- paper. Each couplet (for example, calcite- enced by, seasonal and probably annual plank- anhydrite) was interpreted, marked, and ton productivity. measured on the photographs, and the core While it has never been conclusively demon- measurements were recorded on computer strated that laminae couplets such as those of cards. The result is a time series of approxi- the Castile are varves, no other hypothesis for mately 260,000 varve couplets beginning in the couplet timing in laminated evaporites has been Bell Canyon Formation, about 10.67 m (35 ft) given serious consideration, and this investiga- below the base of the Castile and continuing to tion is framed upon the assumption that each a basal limestone breccia, probably of the couplet (organic-siltstone, organic-calcite, cal- Rustler Formation, that rests on top of the cite-anhydrite, anhydrite-halite) represents an laminations in the lower part of the Salado annual cycle of sedimentation. Formation, a thickness of about 447.2 m Earlier investigations by several of the au- (1467 ft).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 62 ANDERSON AND OTHERS

REGIONAL SETTING calcite-anhydrite couplets of the Castile, and into the Castile halites. The arrangement and So much previous work has been done on the character of the laminations change in succes- regional aspects of evaporites in the Delaware sive lithologies and it is this change by the Basin that no attempt will be made here to addition or subtraction of individual laminae present a complete picture of the setting of types that results in the gross changes that are the basin. defined as stratigraphic units. This same type Regional aspects of the Delaware Basin and of lamina by lamina change, producing gross its evaporite sequences (Table 1) are discussed lithologic variations, also occurs in the Jurassic in the reports of J. E. Adams (1944, 1965, Todilto Formation of New Mexico (Anderson 1967), Adams and Frenzel (1950), Hills (1942), and Kirkland, 1960) and probably is character- P. B. King (1937, 1942, 1948), R. H. King istic of many laminated evaporite sequences. (1947), Lang (1935, 1937), Lloyd (1929), A system of identifying the position of a Newell and others (1953), and other investi- particular lamination or feature within the gators. preevaporite and evaporite time series has been The basin has generally been visualized as adopted that is based upon the position of a surrounded by a carbonate platform (reef) lamina above the base of a particular strati- with a marine opening to the south or south- graphic unit or member of the sequence. For west. Prior to evaporite deposition, fine clastic example, the designation Anhydrite I, To + sediments were deposited within the basin 1,187-1,190, 166.6 cm, indicates that the par- under what may have been deep water, ticular feature occurs 166.6 cm, or 1,187 to "starved basin" conditions (Adams and others, 1,190 laminae couplet units, above the base 1951). Sandstone and siltstone beds grade up- (To) of the Anhydrite I Member. ward into laminated claystone which is interrupted by limestone (Lamar Member of the Preevaporite Phase Bell Canyon Formation) apparently derived Just below the laminated zone, the Bell Can- from the margin of the basin (Tyrrell, 1969). yon Formation in the University of New Carbonate deposition at the basin margins and Mexico-Cowden no. 4 (Fig. 1) is composed of laminated clay and silt deposition within the well-sorted, angular quartz grains and minor basin apparently continued at the same time feldspar grains with a sparse clay matrix and that clastic-evaporite deposition occurred in carbonate cement (Fig. 2A). The first lamina- the "back-reef" areas (Artesia Group). tions appear as fragments of dark brown organic Finally, a sequence consisting mainly of beds material that are aligned in layers about 1 mm of laminated calcite and anhydrite intercalated apart (Fig, 2B). This condition prevails as the with beds of anhydrite-laminated halite was quartz grains diminish in size and frequency deposited within the basin as the Castile and upward in the sequence, and as the amount of Salado Formations. Eventually, the basin be- clay increases the organic laminae become came filled and Salado evaporite deposition better defined and more persistent. Eventu- spread northward and eastward over an area of TABLE 1. STRATIGRAPHIC TABLE OF PERMIAN ROCKS OF THE greater extent than the structural outline of DELAWARE BASIN, WEST TEXAS AND SOUTHEAST NEW MEXICO the basin. During Salado time, potassium salts were deposited within southeastern New Mex- Series Formations Members ico and a small part of Texas. Ochoa Dewey Lake Redbeds

Rustler Anhydrite IV 1 Halite III Salado Anhydrite III PETROLOGY Halite II Anhydrite II Halite I The laminations of the preevaporite and Anhydrite I evaporite phases of Bell Canyon-Castile For- Guadal upe Bell Canyon ' Basal Limestone mations provide a unique means for describing Cherry Canyon and interpreting petrologic variations. Lamina- Brushy Canyon tions of one sort or another occur in a continuous uninterrupted sequence from the Leonard Bone Spring Limestone organically laminated siltstone of the Bell Canyon, through the basal limestone and the Wolfcamp undifferentiated

ally, silt grains diminish to a point where the time progresses, the discrete brown organic unit is a laminated clay stone (Fig. 2C). Silt- laminae that are characteristic of the pre- stone and claystone beds then alternate with a evaporite phase (Fig. 2D) become unrecog- time frequency of several thousand laminae nizable and are replaced by diffuse brown couplets. organic material intimately mixed with the Two fossil components are found in the calcite laminae (Fig. 2E). laminated siltstone and claystone units: ex- The onset of sulfate deposition at the base of tremely minute fusulinids and shell(?) frag- the Anhydrite I Member is marked by two ments replaced by silica (Fig. 3C) and yellowish different types of anhydrite. One type occurs brown algal(?) remains (Fig. 3B). The algal(?) as layers of relatively large anhydrite grains remains are about 60 p in diameter. Their wall which have a yellowish brown coloration in thickness is about 10 n and the central part of transmitted light, probably resulting from the compressed bodies are often filled with organic stain. These anhydrite crystals, which silica. No aperture was observed. The yellowish are dark brown in reflected light, are associated brown grains are aligned parallel to the dark with small (25jii) calcite rhombs and are several brown laminae. They appear in the siltstone se- times larger than unstained anhydrite grains in quence shortly after lamination begins, but the accompanying thicker anhydrite laminae occur sparingly throughout most of the se- (Fig. 2E, F; Fig. 4A). The yellowish brown quence. A few cm below the base of the Castile, anhydrite crystals are frequently oriented with these organic remains are very abundant and their long dimension parallel to bedding, and form thin layers with a brown color in reflected are generally either in contact with or mixed light. with layers of calcite rhombs. These laminae The small fusulinids occur throughout the form some of the sharpest and most distinct claystone and siltstone units of the upper Bell laminae visible in outcrop and in core. Canyon Formation, but are more abundant in This mode of deposition continues for sev- the claystone. Apparently two forms are repre- eral thousand couplets at the base of the Anhy- sented; one is spherical, and the other which is drite I Member before giving way to alternat- less common, is fusiform. Structure is poorly ing laminae of anhydrite and organically preserved owing to partial or complete silicifica- stained calcite (Fig. 2G, H), which is typical of tion, but the ovoid form resembles the neo- the Castile. schwagerinid species Yabeina texana reported from the Lamar by Skinner and Wilde (1955), Calcite-Anhydrite Laminae but is less than one-tenth the size and appears The general petrologic description of Udden to have a few more volutions. The elongated (1924), Adams (1944), and expanded descrip- form may be a boultonid, but is even more tion by Anderson and Kirkland (1966) applies poorly preserved than the Ya&mzfl-like form. to most of the calcite-anhydrite laminations of the Castile Formation, and some additions to Evaporite Transition and refinements of these descriptions are pre- The base of the Castile is marked by the first sented below. appearance of calcite layers between dark Calcite. Calcite layers in the Basal Lime- brown organic laminae similar to those in the stone and lower part of the Anhydrite I Mem- siltstone and claystone units of the Bell Canyon ber (Fig. 2D, E, F, G) are composed of crystals (Fig. 2D). The contact between the Basal about 25 M m diameter and are remarkably con- Limestone Member of the Castile and the stant in size (Fig. 3D). laminated claystone of the Bell Canyon Forma- Examination of thin sections ground to a tion is very abrupt and occurs over a few thickness of 4 to 5 n shows the crystals to be couplets. Yabeina-\ike forms continue for a rhombs of calcite (Fig. 3E, F). Some crystals few centimeters into the Basal Limestone Mem- are euhedral; more commonly, the margins are ber of the Castile, but are confined to the crenulated and the corners rounded, giving the organically rich laminae between the calcite crystals a rounded appearance at low magnifica- layers. The yellowish brown algal(?) remains tion. Calcite "rosettes" were sometimes ob- are not found in the Castile Formation. served with six or more calcite rhombs around a Some basal calcite laminae contain a mosaic nucleus of calcite or unidentified material. of equidimensional calcite crystals with many The size of the small equidimensional calcite of the crystals having sutured boundaries. As rhombs does not change upward in the laminae

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 CASTILE FORMATION BELL CANYON FORMATION BASAL LIME- ANHYDRITE I MEMBER STONE MBR. i O r j > co r 1 CO O I p — o o-o q o 5 Q- 33 i ~ i _| 0- 33 r- a" CD -^ °

but the frequency often diminishes, giving the calcite crystal size occur even in high-carbonate laminae an upper boundary that is less sharp parts of the section except near the base of the than the lower boundary. The calcite crystals Castile where the crystals are smaller and have commonly appear to be suspended in the anhy- a different form. drite groundmass and removed from adjacent The habit of the calcite crystals ranges from grains by several diameters distance (Fig. ovoid rhombs (Fig. 3A), to larger and closely 3E, F). There is some mixing of rounded and packed but still distinct crystals, to laminae of euhedral calcite rhombs in the same lamina, calcite crystals that have become highly inter- but the degree of rounding or angularity within grown and sutured. The calcite crystals do not a particular lamina is usually the same over the usually show a preferred orientation except in distances between the Cowden no. 2 and very thin carbonate laminae (one or two crys- Phillips no. 1 cores (30 km). tals thick) where crystals are commonly imbri- Throughout much of the Castile, calcite cated subparallel to stratification. Organic laminae are composed of larger ovoid or fusi- matter that stains calcite often becomes con- form crystals usually about 75 n in diameter centrated, forming a more or less distinct or- and intimately mixed with and stained by ganic lamina in the upper, lower, or middle part brown organic matter (Fig. 2H; Fig. 3A). of a particular calcite lamina. These are the calcite crystals described by Ud- Anhydrite. Most anhydrite in the Castile den (1924). Typically, a calcite lamina has a Formation consists of an interlocking aggregate sharp basal contact and a less distinct upper of subhedral to euhedral crystals in a dense, contact that represents the mixing of calcite interlocking, fine-grained matrix (Fig. 4B). The crystals with the lower part of the overlying crystals have a distinct rectangular outline. anhydrite lamina. These rectangular crystals are larger than the Measurement of maximum lengths of 800 of crystals in the "matrix" and commonly form a these calcite crystals indicate that no important closely packed aggregate, the so-called "pile- differences in calcite crystal size occur between of-bricks" texture, and thought by many in- cores separated by 15 to 30 km, nor is there a vestigators to be the normal habit of primary significant vertical gradation of grain size anhydrite (Carozzi, 1960, p. 422). The smaller within particular carbonate layers. The mea- crystals of the matrix have the same habit as sured laminae are from a part of the section the larger crystals. The long and short dimen- which is relatively low in calcite (about 12 per- sions of 200 larger crystals were measured and cent), but apparently no major differences in were found to have a mean short dimension of 23 n, a mean long dimension of 30 /n, and a mean long-short ratio of 1.36. There appears to be no important difference in crystal size between Figure 2. Transitional lithology at the top of the the Phillips and the two Cowden sections and Bell Canyon Formation and at the base of the Castile there is no important vertical gradation in Formation (nomenclature is explained in text). (A) crystal size as described by Ogniben (1955 and Non-laminated siltstone, Siltstone I unit, Bell Canyon 1957) in gypsum and anhydrite laminae from Formation. (B) Laminae of siltstone and organic matter, the Sulfur Series of Italy. Siltstone III unit, T0 + 10,470, 233.5 cm, Bell Canyon Formation. (C) Laminae of claystone and organic mat- Although a vertical gradation in size of ter, Claystone II unit, To + 10,581, 115.6 cm, Bell anhydrite crystals does not occur within sulfate Canyon Formation. (D) Laminae of calcite and organic laminae, anhydrite crystals associated with the matter, Basal Limestone Member, T0 + 75, 2.0 cm, organic-anhydrite at the base of the Castile are Castile Formation. (E) Laminae of calcite, anhydrite, frequently larger than anhydrite crystals in the and organic-rich anhydrite, Anhydrite I Member, TO + purer sulfate laminae as noted in the discussion 1,187-1,190, 166.6 cm, Castile Formation. (F) Laminae of the transition zone. of organic-rich anhydrite and anhydrite, Anhydrite I In addition to matrix crystals and rectangular Member, T0 + 1,191-1,195, 166.6 cm, Castile-Forma- crystals, a third type occurs as laths with a tion. (G) Laminae of calcite and anhydrite, Anhydrite width of about 0.1 mm and a length up to I Member, T0 + 13,987-13,990, 1,201.0 cm, Castile Formation. (H) Laminae of calcite (ovoid crystals) and several millimeters long. The laths have indis- anhydrite, Anhydrite I Member, To + 10,791-10,793, tinct, irregular boundaries and often contain 859.9 cm, Castile Formation; the calcite laminae consist inclusions of rectangular anhydrite "blocks." of rounded rhombs of calcite. This couplet form is Laths can be found in all orientations, but the typical of most of the Castile. long dimensions of most are approximately

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 AN'DKKSON \NI) OTHKRS

------

------Figure 3. Fossils in the uppermost Bell Canyon 50.3 cm. (E)) Small ( ^25/i) calcite rhombs in calcite Formation and crystal textures in the Castile Formation. laminae near base of Anhydrite I {see Fig. 2G); note (A) Rounded-rhombohedrons of calcite; note that the organization of calcite crystals into lamina in center of frequency of calcite crystals diminishes upward into the photo. (E,F) Enlarged views of calcite rhombs in overlying anhydrite lamina, Anhydrite I Member, TO partially polarized light; note that some crystals retain a + 10,792, 859.9 cm. (B) Algal(?) remains, Claystone rhombic form and are "floating" in anhydrite ground III unit, To f 5,600, 74.3 cm. (C) Minute fusulinid, mass. compare Yabcinu sp., Claystone III unit, TO -\- 3,840,

normal lo slratihcutum. They commonly, but Halite Layers not invariably, occur within or adjacent to thin carbonate laminae, which are distorted or Halite layers from a tew millimeters to more pierced by the laths. than 10 cm thick are abruptly added to the At the lop ol the evaponte sequence in the calcite-anhydrite couplet pattern at the onset Salado Formation, the calcite-anhydnle cou- oi halite deposition. I he anhydrite laminae plets typical of most ol the Castile give way to continue at about the same thickness with the thicker anhydrite laminae separated by calcite introduction ol halite, but the calcite layers be- layers which have blebs of organic matter come less obvious and less well defined and are coating the calcite grains (Fig. 4Cj. only intermittently present as distinct laminae.

------Figure 4. Calcite and anhydrite crystal textures, 35,014-35,015. Note that organic matter forms a coating Castile and Salado Formations. (A) Anhydrite and on the calcite grains. (D) Calcite rhombs (dark) between organic-rich anhydrite laminae (see Fig. 2F); note con- nodules of anhydrite; note alignment of anhydrite trasting size of anhydrite crystals, with larger crystals in crystals adjacent to calcite band, Anhydrite I Member, organic-rich zones (polarized light); Anhydrite I To + 216-218, 35.6 cm. (E) Similar to (D), but in Member T0 + 21,630-21,631, 2,195.6 cm. (B) Typical polarized light, Anhydrite I Member, T0 + 216-218, blocky anhydrite (polarized light); Anhydrite I Mem- 35.6 cm. (F) Reticulate pattern formed by reorganiza- ber, To + 13,976, 1,200.0 cm. (C) Laminae of organic tion of anhydrite laminae into nodules, Anhydrite I stained calcite and anhydrite, Salado Formation; Member, T0 + 216-218, 35.6 cm. 12,737.8 cm above base of Salado Formation, To +

Some ol the halite layers from the upper part of laminae. They can be observed in insoluble the Castile retain the original crystal structures, residues and sometimes on polished surfaces and including internal laminae that are concentra- in thin sections, but are observed best on x- tions of organic material or anhydrite, and radiographs of slabs approximately 3 mm thick bubbles and vacuoles. Most halite layers, how- cut normally to stratification (see discussion in ever, have become recrystallized (Fig. 5D). Anderson and Kirkland, 1966). Very small quartz and zircon grains with Other Components maximum intercepts of approximately 50 yu Small crystals of pyrite are sparsely present have been observed in insoluble residues of in the Castile, generally at the base of calcite Castile material. Their quantity has not been

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 ANDF.RSON AND OTHERS

determined accurately, but it is probably less mally thin or thick layers of carbonate or than 0.1 percent. The quartz grains are anhydrite that alternate in an irregular pattern rounded and do not show crystal outlines. The (see Anderson and Kirkland, 1966, Fig. 2). zircon grains are prismatic and often show The systematic changes in the proportions pyramidal terminations. of calcite, anhydrite, or organic matter may Small (50 to 500 ju) black magnetic particles result in /.ones or beds that appear to be almost have been extracted from the Castile. The entirely carbonate or almost entirely sulfate in particles are generally irregular in shape, and outcrop or core but that nevertheless contain are similar to magnetic particles described from some small proportion of the other materials. salt samples (Mutch, 1964 and 1966). The laminae associations o( these clastic and mag- Nodular Anhydrite Beds netic (factions have not been determined. Stratigraphic intervals of nodular anhydrite (Fig. 5 A) are associated with parts of the se- LITHOLOGIC VARIATION quence where sultate laminae are thick. Many but not all ol the prominent peaks in the graph Laminae Variation ot the time series (Fig. 6), which represent a Within the anhydrite members, the typical high rate of sulfate deposition, are associated pattern ol lamination is the alternation of with the development of nodules. The nodular calcite and anhydrite laminae previously de- /.ones are also characterized by a loss of defini- scribed. Changes in thickness or proportions of tion of carbonate laminae, and by a change in laminae generally occur in a regular or system- the appearance of the organic fraction from atic manner and produce oscillations in thick- brown to dark gray or black in reflected light; ness such as ihose illustrated in Figure 6. Parts changes which may also take place without the ol the sequence, however, may contain abnor- development of a nodular zone. The number of

------beds. (A) Nodular anhydrite in Anhydrite I, Castile 37,000. (C) Blanket solution breccia correlative with Formation. (B) "Collapse" type breccia in Anhydrite Halite I, Castile Formation. (D) Halite-anhydrite IV. Note angular fragments in tight packing with little couplets. Halite II, T0 f- 130-134. matrix. This breccia occurs above a blanket solution

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 UNION-UNIVERSITY UNIVERSITY OF NEW MEXICO "37" no. 4 PHILLIPS no. I VARVE THICKNESS 432 Imm GAMMA TIME SONIC . (YEARS)

-To+ 250,000

-T0+200,000

-T0+150,000

-TO+100,000 LU

(rt < 4500- O

5000- V. -TO+50,000 BELL CANYON T FORMATION ^

LAMAR LIMESTONE; 64cm thick: turbidite

Blanket solution breccia (varve thickness estimated) -Tn Figure 6. Correlation between sonic log of Union time series of the UNM-Phillips no. 1. Couplet thick- Oil Co .-University "37" no. 4 and smoothed (500-unit ness estimated for halite units, moving average) calcite-anhydrite couplet-thickness

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 70 ANDERSON AND OTHERS

couplets involved in a nodular zone vary from beds of anhydrite breccia (Fig. 5C) of a thick- less than 50 to several hundred. ness approximately equivalent to the sum of The nodular pattern is the result of sulfate the thicknesses of anhydrite laminae present in mobility and recrystallization within previously the correlative halite units. Hence the 140-ft existing sulfate laminae. The recrystallization (42.7 m) section of salt plus anhydrite in the and development of nodules exhibits many lower halite member of the Castile was re- stages, but the nodular anhydrite rarely com- duced to a blanket breccia about 3.4 m (11 ft) pletely loses its laminated appearance. Develop- thick after solution of salt. ment of nodules into a "chicken wire" stage Stratigraphic correlation of laminae show seldom occurs. The anhydrite crystals in the that the onset of halite deposition occurs at nodules lose their blocky character and re- almost the same Stratigraphic position in the crystalize into "felty" anhydrite (Fig. 4D, E). laminated sequence in the eastern and western In the lower part of the formation, where the parts of the basin. The development of nodular calcite crystals are often small equidimensional anhydrite beneath halite has altered original rhombs, the calcite crystals in the laminae be- thickness relations and made correlation diffi- tween nodules are sometimes drawn out into cult to demonstrate statistically. However, dis- thin strands only one crystal layer thick (Fig. tinctive groupings of distorted laminae, brown 4E, F). calcite-anhydrite couplets and gray calcite- The nodular growth clearly took place after anhydrite couplets with equivalent numbers of the deposition of undisturbed laminae and may laminae in the Union-University "37" and have been related, perhaps indirectly, to in- University of New Mexico-Phillips cores, and creased salinity of water. This is supported by a distinctive white lamina, have been used to the association of nodules with increased sulfate establish Stratigraphic correlation for the lami- deposition and by the fact that halite layers are nae immediately beneath Halite I (Fig. 7A). always immediately underlain by nodular Halite precipitation in the eastern sequence pre- zones, although there are many nodular zones ceded precipitation in the western part of the without overlying halite layers. Furthermore, basin, (as inferred from the Stratigraphic posi- those nodular zones underlying halite show the tion of equivalent fragmented layers of solution greatest nodular development. breccia) by only 15 years, or a halite thickness The nodular beds are perfectly correlative in of 85 cm (2.7 ft). all the cores examined and for distances up to Nodular anhydrite development beneath 113 km (70.2 mi) between cores located on Halite III did not completely distort the origi- opposite sides of the basin (see Fig. 7B). The nal thickness relations of calcite-anhydrite cou- onset of nodule formation is remarkably con- plets and Stratigraphic correlation can be stant throughout the basin, further suggesting established on the basis of synchronous thick- that some general control involving the entire ness changes in the two series (Fig. 7B). The basin, such as salinity, was a factor in their de- nodular zone at the base of Halite III begins at velopment. The nodular beds are present the same laminae couplet in both sequences but throughout the Castile-Salado sequence, and halite laminae occur in the Union-University although they are common in the upper part of "37" core 27 couplet units prior to the occur- the sequence, where interstratified beds of rence of solution breccia in the University of halite are more common, the first nodular bed New Mexico-Phillips no. 1 core, or a halite occurs about 1,000 couplets after the beginning thickness of 76 cm (2.5 ft). of anhydrite deposition at the base of the Beds of laminated anhydrite as thin as one Castile Formation. foot thick within halite beds more than 50 ft (15.2 m) thick are also present as unbrecciated Halite and Breccia Beds layers within correlative breccia units, involve Castile halite in the Union Oil Company- approximately the same number of laminae, University "37" no. 4 core consists of alter- and maintain their identity and remain un- nating layers of recrystallized halite several disturbed within overlying and underlying centimeters thick and layers of anhydrite up to breccia beds despite the removal of salt. The several millimeters thick (Fig. 5D). presence of thin anhydrite beds within halite Halite beds in the eastern part of the basin members can also be inferred from sonic logs, once extended beyond their present western and these beds can be correlated with breccia limit into the western part of the basin (Figs. 1 zones in the University of New Mexico- and 8). The dissolved salt intervals are now Phillips no. 1 core (Fig. 3).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 GRAY BROWN GRAY BROWN LAMINAE LAMINAE LAMINAE LAMINAE

1.0 cm-

0.5cm-

2 15

.5cm-

I.Ocm-

0.5cm-

NUMBER Figure 7. Correlation diagrams showing timing of Note that halite laminae occur in the Union University Note also that nodular development begins at the same halite deposition at the base of Halite I (A) and at the "37" sequence 15 yrs prior to their equivalent position time in both cores (see Fig. 6 for depth and time series base of Halite III (B) in the Union-University "37" no. as solution breccia in the UNM-Phillips no. 1 sequence locations). 4 (hachured) and UNM-Phillips no. 1 varve sequences, for Halite I and 27 yrs prior to breccia in Halite III.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 PERMIAN CASTILE EVAPORITE SEQUENCE, WEST TEXAS AND NEW MEXICO 73

The breccia generally consists of rectangular- probably extend over most of the Delaware shaped, subangular fragments of single laminae Basin and have been subdivided into six work- or groups of laminae embedded in a matrix of ing units. The type section for the subdivision anhydrite (Fig. 5C). The fragments, generally is a partial core from the Union-University less than one cm in length, occur in various "37" no. 4 supplemented by a sonic log, from orientations, but most occur with stratification, which correlation can be made throughout if visible, and long dimension near the hori- most of the basin. An additional supplement to zontal. Many of the fragments appear to have the type section is the University of New been only slightly displaced. Mexico-Phillips no. 1 core, which includes the In some of the blanket breccia beds it is entire Castile Formation and can be considered difficult to correlate the upper contact because a "master" or "type" time series for the basin. of solution collapse that resulted in a collapse- The relation between these two sequences and type breccia (Fig. 5B) consisting of larger, position of members is shown in Figure 6. The more angular fragments than the blanket solu- number and average thickness of varves in each tion breccia, and with little matrix. Good ex- unit are given in Table 2. amples of collapse-type breccia have been observed at the top of the Halite II Member and Upper Bell Canyon Formation in the upper part of the Anhydrite IV Member The upper part of the Bell Canyon Forma- above blanket solution breccia. tion can be subdivided into a number of units, Siltstone I through Claystone III (Fig. 6), SUBDIVISION AND DISTRIBUTION which are correlative over a large part of the OF THE CASTILE AND UPPERMOST Delaware Basin. BELL CANYON FORMATIONS The siltstone and claystone units of the up- The Castile Formation has been subdivided permost Bell Canyon varved sequence are here into eight members which permit exami- easily recognizable in the cores from the nation of the present areal distribution patterns western part of the basin, but they are not as of halite and anhydrite. Siltstone and claystone well defined in the core from the eastern part. units in the uppermost Bell Canyon Formation There is excellent correlation of laminae in

TABLE 2. SUBDIVISIONS OF THE UPPER BELL CANYOH-CASTILE SEQUENCE, DELAWARE BASIN, TEXAS AND NEW MEXICO

UNM-Phlllips #1 Number of Average thickness of varve couplets Thickness calcite-anhydrite varve couplets

Salado Formation 35,422 12,660 cm 0.36 cu (partial section, undifferentiated)

Castile Formation Anhydrite IV 54.187 9,842 cm 0.18 cm Halite III *17,879 2,748 cm 0.16 cm (including anhydrite beds) Anhydrite III 46,592 9,554 cm 0.21 m Halite II t 1,758 801 cm 0.45 m (including anhydrite beds} Anhydrite II 14,414 2,738 cm 0.19 m Halite I t 1,063 330 cm 0.31 m Anhydrite I 38,397 5,092 cm 0.13 m Basal Limestone 600 28 cm 0.04 m

Estimated totals (Castile Formation) 174,890 31,133 cm

Average thic ness of Units clastic-organic v rge couplets Bell Canyon Formation Claystone III 5,800 78 cm 0.01 m Siltstone III 24,814 551 cm 0.02 m Claystone II 15,650 166 cm 0.01 m Siltstone 11 1,086 44 cm 0.04 m Claystone I ttca. 2,000 24 cm ca. 0.01 m Siltstone I ttca. 1,500 61 cm ca. 0.04 m

Estimated totals (Bell Canyon Formation) 50,850 924 cm

Combined totals 261,162 44,717 cm

* Number of layers in halite fraction determined by extrapolation. t Number of layers determined in Union-University "37" #4 core; thickness of calcite-anhydrite fractions only. tt Number of layers in UNM-Cowden #4 core.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 74 ANDERSON AND OTHERS

these units between the Phillips no. 1 and isopach map constructed chiefly from sonic logs Cowden no. 4 cores, a distance of 24 km (Fig. (Fig. 10). The thickness is a fairly constant 170 9), although the Cowden no. 4 section is about ft (51.8 m) in the western part of the basin and one-third thicker and contains more calcium increases in the eastern part to about 350 ft carbonate than the Phillips no. 1 section. (106.7 m). Anhydrite I becomes more calcare- Laminae in the siltstone-claystone units in ous in the southwestern Delaware Basin the Union-University "37" no. 4 core are (Adams, 1944) and thickens radially to the about the same thickness but with less siltstone north and east from this area. and more carbonate than in the Phillips no. 1 The continuity of laminations within the and Cowden no. 4 cores. Only a two-ft (0.61 Anhydrite I Member and other anhydrite m) sequence of laminae could be correlated members of the Castile is illustrated by the cor- with certainty. Lamina proportions in the relations in Figure 11. The correlations are for Union-University "37" no. 4 core differ con- cores from widely separated parts of the basin siderably from those in the other two cores, and show the relatively small degree of change yet it is remarkable that laminae which have a in the amount of sulfate precipitated on a large clastic component would retain their lamina by lamina basis, even for opposite sides identity over a distance of 113 km (70.2 mi). of the basin as is the case for Figure 11C. Figure The Lamar Limestone Member of the Bell 11A and B show photomicrographs of thin Canyon Formation interrupts the Claystone II sections of correlative intervals with a north- unit of the varve sequence in the Phillips no. 1 south separation of about 65 mi (105 km; see core between T0 + 7,306 and 7,581; 33.0-147.0 Fig. 1). cm. The graded turbidite limestone beds occur This continuity of lateral distribution differs between the same laminae in both the Phillips for the organic, carbonate, and sulfate fractions no. 1 and Cowden no. 4 cores but are more that comprise the laminations. The three com- numerous and thicker in the Phillips section. A ponents can be separated from each other by few similar limestone beds several centimeters sampling and analyzing the material on a unit- thick are also found below Siltstone III in the time basis (see Anderson, 1967; Kirkland and Union-University "37" no. 4 core on the other Anderson, 1969). Correlation coefficients for side of the basin, but it could not be deter- the percent of each component in 10-couplet mined if they were precisely at the same strati- and 50-couplet samples from different parts graphic position. A short laminated section of of Anhydrite I and for the actual amount 275 clastic-organic couplets is interbedded with the Lamar Limestone. Castile Formation Basal Limestone Member. Many evaporite sequences begin with a basal carbonate, and the Castile is not an exception. The Basal Lime- stone Member of the Castile, however, is very thin and occupies only about 1/400 of Castile time and about 1/1500 of Castile stratigraphic thickness. This member, which contains no anhydrite, extends throughout much of the Delaware Basin and was recognized as a distinct unit by King (1942). In the Phillips no. 1 core, the member has a thickness of about 28 cm and in the University "37" no. 4 about 50 cm. The unit consists of about 600 calcite-organic couplets. It is considered a member because of its distinct character and persistence. An isopach COWDEN 4 PHILLIPS map was not constructed because the unit can- Figure 9. Laminated siltstone from correlative not definitely be delimited on wire-line logs. sections of the upper part of Bell Canyon Formation. Anhydrite I Member. The thickness distri- Quartz silt laminae alternate with laminae of organic bution of the lowermost anhydrite unit, which matter (dark); Siltstone III unit, To + 10,200-10,250, contains about 38,000 couplets, is shown by an 229.5-230.5 cm.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 EDDY

cjfr-.^W t V\ ^ 5-^5 0 10 20 30 KILOMETERS 0 5 10 15 20 MILES

Well control Contour interval: 50 feet (15.2 m ) 175 foot contour added

Figure 10. Thickness distribution of Anhydrite I Member, Castile Formation.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 76 ANDERSON AND OTHERS

(thickness) of each component in the same moving correlation coefficient for couplet samples show that lateral continuity is greatest thickness between the three cores (Fig. 12). for the sulfate (Tables 3 and 4). Carbonate The greater continuity in a northwest-south- distribution is more variable than sulfate but east direction agrees with the thickness trend correlation coefficients are still high and in that part of the basin as illustrated by the significant, whereas the organic fraction (as 175 ft (53.3 m) isopach in Figure 10. determined by weight loss) has a more variable Statistical correlation studies have not yet distribution. been done for the Union Oil Company-Univer- A comparison of the correlation coefficients sity "37" no. 4 cores from the eastern part of for the different cores (Table 3) also reveals the basin, which includes only the uppermost that there is greater continuity between the and lowermost part of Anhydrite I. Strat- Phillips no. 1 core and the two Cowden cores, igraphic correlations of laminae, however, which are 24 km (14.9 mi) and 32 km (19.8 mi) reveal that couplets in certain parts of the to the northwest, than between the two varve sequence maintain almost exactly the Cowden cores which are separated from each same thickness proportions and general ap- other by 14 km (8.7 mi) in a north-south pearance (for example, contact relations, color) direction. This difference in continuity with over the 113 km (70.2 mi) distance. Other direction in the basin is best illustrated in the parts of the sequence have couplets with ------

------

Figure 11. Correlative Castile Sections (A,B) Cor- mi) north-northwest of the Flood-Grishairi no. 1. (C) relative intervals (thin sections, plain light); couplets Correlative couplets of organic-rich calcite (dark) and of organic-rich calcite (dark) and anhydrite. The anhydrite (core slabs); the Union-University "37" no. Pokorny no. 1 is 29.0 km (18.0 mi) north-northwest of 4 is 91.0 km (56.4 mi) east of the Phillips no. 1. the Phillips no. 1, and the Phillips no. 1 is 32 km (19.8

thickness proportions sufficiently different Halite I thickens gradually from south to to make lamina by lamina correlation extremely north in the eastern part of the basin and has a difficult, although the longer trends and more maximum thickness of more than 400 ft (122 variable couplets can be readily matched (Figs. m) in Lea County, New Mexico (Fig. 13). 7Band 11C). Most of this northward thickening is probably Halite I Member. The lowermost halite due to an increase in thickness of individual member of the Castile is the thickest and most laminae judging from the near synchroneity of extensive of the Castile halite units. In the halite deposition in the eastern and western University "37" no. 4 core, the Halite I parts of the basin (Fig. 7A). Even if halite be- Member contains 1063 + 20 anhydrite-halite gan precipitating 200 yrs earlier in the north couplets with an average thickness of 3.1 cm. than in the University of New Mexico- The anhydrite-halite couplets are thickest at Phillips no. 1 core, this would mean an annual the base of Halite I and decrease gradually up- deposition rate of as much as 10 cm (3.9 in.) of ward within the unit. halite in the northern part of the basin. The original thickness of Halite I in the western TABLE 3. CORRELATION COEFFICIENTS FOR 10-COUPLET SAMPLES part of the basin cannot be determined. The FROM ANHYDRITE I, ^n+ 33,921-34,871 FROM COWDEN 2, COWDEN 4, AND PHILLIPS CORES OF THE CASTILE FORMATION thickness of the solution breccia zone equivalent to the Halite I is about 330 cm in the Phillips Cowden 2 Cowden 2 Cowden 4 vs. vs. vs. no. 1, Cowden no. 2, and Cowden no. 4 cores Variable Cowden 4 Phillips Phillips and all that can be determined about the past

Couplet thickness 0.68 0.70 CK77 thickness of halite in this area is that enough halite was interstratified with anhydrite Percent CaCO- 0.54 OJ1 0.63 laminae to cause brecciation upon solution. Percent organic -0.04 0.14 0.04 Anhydrite II Member. The Anhydrite II Percent CaSO^ OJ6 OJ5 0.61_ Member in the University of New Mexico- Absolute carbonate CL29 0,39 0.3L Phillips no. 1 core contains about 14,000 calcite-

Absolute organic 0.10 -0.05 0.15 anhydrite couplets. The thickness of Anhydrite II (Fig. 14), like the thickness of Anhydrite I Absolute sulfate (hSi 0..73 0.77 (Fig. 10), increases from west to east, with N = 95; 99« confidence limits = ±0.27 lines of equal thickness nearly paralleling the Values which are significant at the 998 level are under- eastern and northern margins of the basin lined. Thickness values are summations of 10 Individual laminae; percent carbonate, organic, and sulfate were (Figs. 10 and 14). However, the rate of east- determined by loss on ignition; absolute carbonate, or- ward thickening of Anhydrite II is much less ganic, and sulfate were calculated by multiplying the percent value by couplet thickness. than the rate of thickening of Anhydrite I. Halite II Member. The Halite II Member

TABLE 4. CORRELATION COEFFICIENTS FOR 50-COUPLET SAMPLES is about 200 ft (61.0 m) thick in the northern FROM ANHYDRITE I, T0 + 24,770-26,620 FROM COWDEN 2, part of the basin and about 115 ft (35.0 m) COWDEN 4, AND PHILLIPS CORES OF THE CASTILE FORMATION thick in the Union Oil Company-University

Cowden 2 Cowden 2 Cowden 4 "37" no. 4 core. The halite is interrupted by vs. vs. vs. five beds of carbonate-laminated anhydrite, Variables Cowden 4 Phillips Phillips ranging from a few centimeters to over 1 m Couplet thickness 0.99 0.99 0.99 thick that can be observed readily on sonic logs.

Percent CaC03 0.99 0.99 0.99 The entire Halite H Member including the Percent organic -0.01 0.81, -0.09 couplets in the anhydrite beds encompasses 1758 + 10 couplets of which about 1139 + 10 Percent CaSO,, 0.99 0.99 0.99 are anhydrite-halite and the others calcite- Absolute CaC03 0.93 0.94 0.95 anhydrite. The average thickness of the halite- Absolute organic 0.54 0.86 0.55 anhydrite couplets is 2.3 cm with the thickness Absolute CaSO^ 0.99 0.99_ 0.99 decreasing gradually upward within each halite N = 28; 99? confidence limits = ±0.50 unit between the five laminated anhydrite beds. Values which are significant at the 99% level are under- lined. Thickness values are summations of 50 individual The same five beds of laminated anhydrite couplets; percent carbonate, organic, and sulfate were occur between breccia beds in the Phillips no. 1 determined by loss on ignition; absolute carbonate, organic, and sulfate were calculated by multiplying the core, but the relation is vague in the upper few percent value by couplet thickness. feet of breccia because of faulting and collapse.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 EDDY

Well control Well control Contour interval: 100 feet (30.5m) Contour interval: ZS feet{7.6m) 50 foot contour added 90 foot contour added Figure 13. Thickness distribution of Halite I Member, Castile Formation. Figure 14. Thickness distribution of Anhydrite II .Member, Castile Formation.

The distribution pattern of Halite II is in Figure 16. The University "37" no. 4 core similar to that of the Halite I Member, but collected only the lower three halite units in Halite II extends farther south and not as far Halite III. These three beds contain 297 + 10 west (Fig. 15). The western limit, however, halite-anhydrite couplets with an average does not correspond to the original depositional thickness of 4.6 cm per couplet. The sonic log limit because the corresponding solution of the Union Oil Co.-University "37" no. 4 breccia units are well developed in the western well (Fig. 6) indicates that the Halite III part of the basin and Halite II may have Member contains approximately 72 m of halite originally extended as far as, or perhaps farther and 40 m of interbedded anhydrite. Projecting than, Halite I. the rate of 4.6 cm per couplet obtained for the Anhydrite HI Member. The Anhydrite III cored halite units to all salt in the Halite III Member in the Phillips no. 1 core is a sequence gives a total time of halite deposition of ap- of calcite-lammated anhydrite above the proximately 1,600 years. The total time of Halite II Member, contains about 46,600 deposition of the Halite III Member is es- couplets, and is generally 280 (85.3 m) to 300 timated to be about 18,000 years. ft (91.4 m) thick. It thickens from west to east Anhydrite IV Member. Anhydrite IV con- but at a lesser rate than either Anhydrite I or tains about 54,000 calcite-anhydrite couplets. Anhydrite II. The lowest halite bed of the The number of couplets assigned to Anhydrite overlying Halite III Member is absent in the IV depends upon which breccia beds within the western part of the basin and an anhydrite bed Phillips core are selected as representing the within the Halite III Member lies directly onset of dominant halite deposition in the upon the anhydrite of Anhydrite III, therefore, Salado Formation. The thick breccia beds at isopach map was not constructed. about TO + 240,000 in the time series (Fig. 6) Halite III Member. This member is a mixed correlate with halite beds within the Salado. halite-anhydrite unit with more time involved The breccia bed in the varve sequence selected in anhydrite than halite deposition, but with as the top of Anhydrite IV (Salado boundary) halite occupying a greater thickness. The occurs at TO + 53,979; 9,842 cm above the top distribution of halite within this unit is shown of Halite III. In the Phillips no. 1 core, this

1.0 - 0.8: 0.6 - UJ 0.4 - COWDEN O

UJ 1.0 8 0.8 - Z 0.6 - COWDEN 2 vs PHILLIPS I

LJ oc 1.0 -, cc o 0.8 - o 0.6 - 0.4 - COWDEN 2 vs COWDEN 4

0 + 34,020 34,420 34J320 years TIME Figure 12. Moving correlation coefficients for Phillips 1 cores, To +33,921-34,871 zone of Anhydrite couplet thickness between Cowden 4, Cowden 2, and I member, Castile Formation (N = 51 yrs).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 Figure 15. Thickness distribution of Halite II Member, Castile Formation. Figure 16. Thickness distribution of halite units within Halite III Member.

breccia bed occurs in a sequence of couplets the basin had been made on the basis of sonic in which the calcite laminae are thinner than log correlations prior to the availability of the for typical Castile anhydrite and in which Union-University "37" halite core. The sulfate lamina thickness increases by about 50 Winkler County core, however, revealed that percent (see Fig. 6, nearT0 + 230,000). There- thin anhydrite beds of only a few decimeters after, the calcite-anhydrite couplet thickness thick within more massive halite units main- remains high in the Salado except for a few tained their position and character after halite brief intervals, which occur at about 240,000, solution. This fact, and the observation that and 250,000 couplets above the base of the single anhydrite laminae, once separated by time series. several centimeters of halite, were sometimes A thick breccia bed occurs in the Phillips little disturbed upon solution, showed that the core, well down into Anhydrite IV, at To + withdrawal of halite was a very gentle process. 37,214. This bed marks the presence of a dis- With the exception of one halite bed in solved halite. Based on the thickness of other Anhydrite IV, every halite bed observed in the breccia beds which can be correlated to halite Winkler County core from the eastern part of beds of known thickness and depositional rate, the basin has an equivalent breccia bed in the the dissolved halite bed probably contained University of New Mexico-Phillips no. 1 core. at least 400 couplets and was probably 40 to Inasmuch as this core locality is only about 32 50 ft (12.2 to 15.2 m) thick. A halite bed of this km (20 mi) from the western edge of the basin, thickness should be recorded on sonic logs, there is every reason to suppose that halite but appears only as a slight' 'kick" on some logs deposition once extended to, or nearly to, the in the northeastern part of the basin. However, western margin. The present western solution in the southeastern part of the basin, this inter- margin of halite units within the Castile shifts val contains more than 300 ft (91.4 m) of halite progressively eastward, with Halite II more (Fig. 17) and is interrupted by several anhy- areally restricted than Halite I. The halite in drite beds. The correlative breccia in the the Salado, however, extends farther westward Phillips no. 1 core contains no nonbrecciated than the present western solution limit of beds suggesting that some halite beds in Anhy- Castile halite (Fig. 8). This suggests that an drite IV, unlike halite beds in all lower units, episode of solution might have taken place may have had no equivalents in the western prior to Salado deposition. The isopach map of part of the basin. the halite beds within Anhydrite IV (Fig. 17) shows a very irregular distribution of halite in Salado Formation the east-central and northeastern part of the Halite is the dominant lithology in the basal basin that is not present in any of the lower Salado Formation in the eastern part of the Castile halites and could also represent solution basin, whereas anhydrite with blanket breccia prior to Salado deposition. beds is the dominant lithology in the western It seems more likely, however, that all of the part where the Phillips no. 1 core contains the solution took place after Salado time and that lower one-third of the Salado Formation. The the irregular distribution pattern in Anhydrite Salado distribution pattern (Fig. 18) contrasts IV developed later. A comparison of the Anhy- markedly with that of the halite and anhydrite drite IV isopach for halite and the published members in the underlying Castile. The Castile map of Tertiary basin fill of Maley and Huf- members are confined to the Delaware Basin fington (1953) shows a very close agreement proper. The Salado, however, overlaps the between the locus of Cenozoic basin fill in the Delaware Basin and is present on adjoining Delaware Basin and the areas of thin or missing areas to the north and east. The thickest deposi- halite in Anhydrite IV. Similarly, there is also tion in the Salado is north of the locus of the a correlation between the Cenozoic basins and thickest Anhydrite IV Member and overlies thin areas in the Salado. the thickest part of the Halite II Member, but covers a broader area. SYNCHRONEITY AND VARIATION OF STRATIGRAPHIC UNITS EFFECTS OF SOLUTION A comparison of the isopachs of the Anhy- The interpretation that breccia beds in the drite I and II and the Halite I and II Members University of New Mexico-Phillips no. 1 core reveals that halite gradually thickens toward represented halite beds in the eastern part of the north-northeast with a trend that differs

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 Figure 17. Thickness distribution of halite units within Anhydrite IV Figure 18. Thickness distribution of Salado Formation. Member, Castile Formation.

by about 90° from the anhydrite trend. Com- tion. The so-called ripple marks (Lang, 1937; parison of laminae at the base of Halite I and Porch, 1917) are not sedimentary structures, Halite III in the Union-University "37" core but are minor tectonic features that originated in Winkler County and the same laminae in after consolidation (Kirkland and Anderson, the Cowden cores in the west-central part of the 1970). The nodular beds in the Castile, while basin, a distance of 113 km (70.2 mi), shows a superficially resembling the nodular beds 15- and 27-yr difference in the onset of halite associated with tidal flat sedimentation, are deposition (Fig. 7). closely associated with normal varving and in Halite and anhydrite beds within Halite II fact are varved themselves and show no primary appear to show the same degree of synchroneity. breaks in the continuity of sedimentation. It is more difficult to observe the end of halite Estimates of the depth of water have deposition in major units owing to collapse of ranged from 150 to 700 m (King, 1934; Adams nonbrecciated laminated anhydrite im- and Frenzel, 1950; Adams, 1944; Kroenlein, mediately above solution breccia, but judging 1939) and are based chiefly on the present-day from correlations of individual anhydrite beds relief between the top of the Capitan Formation within Halite II, approximately the same (the "reef") and the base of the Castile. number of laminae are involved, suggesting Newell and others (1953, p. 189) and Adams that the end of halite deposition in different and Frenzel (1950) discuss this method. parts of the basin was also nearly synchronous. Ideally, the depth of water within an evap- The synchroneity of halite deposition and orite basin should have little effect on the the markedly different trends for halite and precipitation process (Schmalz, 1969) and the anhydrite, suggest that the classical model of accumulated sequence should reflect changes in evaporite salt zonation, as described, for environmental conditions of the water body. example, by Scruton (1953) must be modified In the Castile sequence, however, there is a for the Delaware Basin. Some lateral zonation progressive change in the proportion of mate- exists, but factors that triggered halite deposi- rials over an interval of several hundred thou- tion seem to have affected almost the entire sand years. basin simultaneously. Within the three halite members, for The isopach maps of the Anhydrite I and II example, intercalated beds of halite and anhy- Members indicate a thickening from west to drite become more and more common higher east in the form of a fan-shaped wedge. The in the formation. Halite I is a single bed of fanlike shape is best illustrated by the 175 ft halite. Halite II is interrupted by five thin contour in Anhydrite I and the 90 ft contour anhydrite beds and Halite HI by six major in Anhydrite II (Figs. 10 and 14). According anhydrite beds. In addition, the time series to basin reflux models of King (1947) and plot of couplet thickness (Fig. 6) shows a Scruton (1953) the thickness of a particular progressive increase in the amplitude of a evaporite facies should thicken radially from dominant oscillation in sulfate thickness that the marine connection. If this is the case for has a frequency between 1,000 and 3,000 years the Delaware Basin, then the distribution (compare, for example, the tendency toward patterns of anhydrite in the Castile suggest oscillation in Anhydrite I and IV). that marine water entered the basin from the These progressive changes within the basin west over or through the reef, rather than from could be attributed to prolonged trends in the south as suggested by Kroenlein (1939), climate or sea level or they could simply be the King (1942) and Adams (1944). The halite result of a progressive shallowing of the basin distribution patterns would favor the inter- and the increasing impact of climatic change or pretation of a southern source but inasmuch as freshening upon a smaller water volume within anhydrite represents about 97 percent of the basin. Castile time it may be more reasonable to look for an alternate explanation for the differing CONCLUSIONS halite distribution. The lamination (varving) process began prior to evaporite deposition and continued BASIN DEPTH uninterrupted throughout the deposition of a No sedimentary features observed or re- basal limestone member, four anhydrite ported from the Castile Formation can be members, and three halite members of the construed as evidence for shallow water deposi- Castile Formation. Individual laminations

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021 84 ANDERSON AND OTHERS

persist laterally for 113 km (70.2 mi) and New Mexico: Am. Assoc. Petroleum Geologists probably extend throughout the basin. Bull., v. 28, p. 1592-1625. The calcite-anhydrite laminations that are 1965, Stratigraphic-tectonic development of typical of the Castile changed character during Delaware Basin: Am. Assoc. Petroleum times of high sulfate deposition. The same Geologists Bull., v. 49, p. 2140-2148. thick anhydrite layers developed into beds of 1967, Semi-cyclicity in the Castile Evaporite, nodular anhydrite after formation; nodular in Elam, J. G., and Chuber, S., eds., Cyclic sedimentation in the Permian Basin: West laminae and zones are also correlative within Texas Geol. Soc. Symposium, p. 197-203. the basin. The episodes of high sulfate deposi- Adams, J. E., and Frenzel, H. N., 1950, Capitan tion or nodule development are separated by barrier reef, Texas and New Mexico: Jour. 1,000 to more than 3,000 laminae couplets. Geology, v. 58, p. 289-312. Halite deposition in each member was of Adams, J. E., Frenzel, H. N., Rhodes, M. L., and short duration (1,000 to 2,000 yrs) and the Johnson, D. P., 1951, Starved Pennsylvanian timing of deposition was in response to the Midland Basin: Am. Assoc. Petroleum Geol- same changes that produced thick sulfate ogists Bull., v. 35, no. 12, p. 2600-2607. laminae. Halite beds originally extended Adams, S. S., 1967, Bromine in the Salado Forma- tion, Carlsbad Potash District, New Mexico throughout the basin and are represented now [Ph.D. thesis]: Cambridge, Harvard Univer- by blanket beds of solution breccia in the sity, 202 p. western part of the basin. Anderson, R. Y., 1967, Sedimentary laminations in Anhydrite members of the Castile thicken time-series study, in Mernam, D. F., ed., eastward and halite members thicken north- Computer applications in the earth sciences: ward, with a trend difference of about 90°; the Colloquium on time-series analysis: Kansas onset and end of halite and anhydrite deposi- Geol. Survey Computer Contr. 18, p. 68-72. tion is nearly synchronous over 113 km (70.2 Anderson, R. Y., and Kirkland, D. W., 1960, mi) and probably over the entire basin, sug- Origin, varves and cycles of Jurassic Todilto Formation, New Mexico: Am. Assoc. Petro- gesting that the classical model of evaporite leum Geologists Bull., v. 44, p. 37-52. zonation must be modified for the Castile 1966, Intrabasin varve correlation: Geol. Soc. sequence. Also, influx of water into the basin America Bull., v. 77, p. 241-256. was apparently from over or through the west- Briggs, L. I., 1957, Quantitative aspects of evaporite ern reef or platform. deposition: Michigan Acad. Sci., Arts and A progressive upward increase in episodes Letters Paper, v. 42, p. 115-123. of halite deposition and an increase in the Carpelan, L. H., 1957, Hydrobiology of the Aluiso fluctuation of sulfate deposition with time sug- Salt Ponds: Ecology, v. 38, p. 375-390. Carrozzi, A. V., 1960, Microscopic sedimentary gest a prolonged and sustained change in petrography: New York, John Wiley and environment or progressive shallowing of the Sons, Inc., 458 p. basin. Hills, J. M., 1942, Rhythm of Permian Seas, a paleogeographic study: Am. Assoc. Petroleum Geologists Bull., v. 26, no. 2, p. 217-255. ACKNOWLEDGMENTS King, P. B., 1934, Permian stratigraphy of trans- The coring, data collection, and lab operat- Pecos Texas: Geol. Soc. America Bull., v. 45, ing expenses have been supported by the Earth p. 697-798. 1937, Geology of the Marathon region, Texas: Sciences Section of the National Science U. S. Geol. Survey Prof. Paper no. 187, 148 p. Foundation. The work of Dean and Snider 1942, Permian of west Texas and southeastern was partly supported during tenure as National New Mexico: Am. Assoc. Petroleum Geologists Aeronautic and Space Administration trainees. Bull., v. 26, p. 535-763. The authors are indebted to William T. 1948, Geology of the southern Guadalupe Holser of the University of Oregon and to Mountains, Texas: U.S. Geol. Survey Prof. Chevron Research, Standard Oil Company of Paper 215, 183 p. California, for contributing slabs from the King, R. H., 1947, Sedimentation in Permian Union Oil Company-University "37" no. 4 Castile sea: Am. Assoc. Petroleum Geologists Bull., v. 31, p. 470-477. cores, Winkler County, Texas. Kirkland, D. W., and Anderson, R. Y., 1969, Com- position and origin of Rita Blanca varves, in Anderson, R. Y., and Kirkland, D. W., eds., REFERENCES CITED Paleoecology of an Early Pleistocene lake on Adams, J. E., 1944, Upper Permian Ochoa Series of the high plains of Texas: Geol. Soc. America Delaware Basin, west Texas and southeastern Mem. 113, 215 p.

1970, Microfolding in the Castile and Todilto 1957, Secondary gypsum of the sulfur series, evaporites, Texas and New Mexico: Geol. Sicily and the so-called integration: Jour. Sed. Soc. America Bull., v. 81, p. 3259-3282. Petrology, v. 27, p. 64-79. Kroenlein, G. A., 1939, Salt, potash, and anhydrite Phleger, F. B., 1969, A modern evaporite deposit in Castile Formation of southeastern New in Mexico: Am. Assoc. Petroleum Geologists Mexico: Am. Assoc. Petroleum Geologists Bull., v. 53, p. 824-829. Bull., v. 23, p. 1682-1693. Porch, E. L., Jr., 1917, The Rustler Springs sulfur Lang, W. T. B., 1935, Upper Permian formation of deposits: Texas Univ. Bull. no. 1722, 71 p. Delaware Basin of Texas and New Mexico: Richter-Bernburg, G., 1964, Solar cycle and other Am. Assoc. Petroleum Geologists Bull., v. 19, climatic periods in varvitic evaporites, in p. 262-270. Nairn, A.E.M., ed., Problems in palaeocli- 1937, The Permian formations of the Pecos matology: New York, Interscience Publishers, Valley of New Mexico and Texas: Am. Assoc. p. 510-519. Petroleum Geologists Bull., v. 21, p. 833-898. Schmalz, R. F., 1969, Deep-water evaporite deposi- Lloyd, E. R., 1929, Capitan Limestone and asso- tion: A generic model: Am. Assoc. Petroleum ciated formations in New Mexico: Am. Assoc. Geologists Bull., v. 53, no. 4, p. 798-823. Petroleum Geologists Bull., v. 13, p. 645-657. Scruton, P. C., 1953, Deposits of evaporites: Am. Maley, V. C., and Huffington, R. M., 1953, Assoc. Petroleum Geologists Bull., v. 37, no. Cenozoic fill and evaporite solution in the 11, p. 2498-2512. Delaware Basin, Texas and New Mexico: Geol. Skinner, J. W., and Wilde, G. L., 1955, New Soc. America Bull., v. 64, p. 539-546. fusulinids from the Permian of west Texas: Mutch, T. A., 1964, Extraterrestrial particles in Jour. Paleontology, v. 29, no. 6, p. 927-940. Paleozoic salts: New York Acad. Sci. Annals, Tyrrell, W. W., Jr., 1969, Criteria useful in inter- v. 119, p. 166-185. preting environments of unlike but time- 1966, Abundance of magnetic spherules in equivalent carbonate units (Tansill-Capitan- Silurian and Permian salt samples: Earth and Lamar), Capitan Reef Complex, west Texas Planetary Sci. Letters, v. 1, p. 325-329. and New Mexico, in Depositional environ- Newell, N. D., Rigby, J. K., Fischer, A. G., ments of carbonate rocks: Soc. Econ. Paleontol- Whiteman, A. J., Hickox, J. E., and Bradley, ogists and Mineralogists Spec. Pub. no. 14, p. J. S., 1953, The Permian reef complex of the 80-97. Guadalupe Mountains region, Texas and New Udden, J. A., 1924, Laminated anhydrite in Texas: Mexico: San Francisco, Freeman and Co., Geol. Soc. America Bull, v. 35, p. 347-354. 236 p. Ogniben, Leo, 1955, Inverse graded bedding in MANUSCRIPT RECEIVED BY THE SOCIETY OCTOBER primary gypsum of chemical deposition: Jour. 19, 1970 Sed. Petrology, v. 25, p. 273-281. REVISED MANUSCRIPT RECEIVED JULY 19, 1971

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/1/59/3428556/i0016-7606-83-1-59.pdf by guest on 25 September 2021