Criteria for the recognition of salt-pan evaporites

zyxwvutsrqp zyxwvutsrqp Sedinientology ( 1 985) 32, 627--644 zyxw zyxwvutsrq zyxwvu Criteria for the recognition of salt-pan evaporites T I M K . LOWENSTEIN* Conoco, Inc., Ponca City, OK 74603 U . S . A . L A W R E N C E A. H A R D I E Department ojEurth and Planetary Sciences, The Johns Hopkins University, Baltimore, M D 21218, U . S . A . ABSTRACT Layered evaporites can accumulate in: ( I ) ephemeral saline pans, (2) shallow perennial lagoons or lakes, and (3) deep perennial basins. Criteria for recognizing evaporites deposited in these settings have yet to be explicitly formulated. The characteristics of the ephemeral saline pan setting have been determined by examining eight Holocene halite-dominated pans (salt pans) and their deposits (marine and non-marine) from the U.S., Mexico, Egypt and Bolivia. These salt pans are typified by alternating periods offlooding, resulting in a temporary brackish lake, evaporative concentration, when the lake becomes saline, and desiccafion,which produces a dry pan fed only by groundwater. The resulting deposits consist of alternating layers (millimetres to decimetres) of halite and mud. The layers of halite are characterized by: ( 1 ) vertical and horizontal cavities, rounded crystal edges and horizontal truncation surfaces, due to dissolution during flooding; (2) vertical ‘chevrons’ and ‘cornets’ grown syntaxially on the bottom during the saline lake stage; (3) halite cements (overgrowths and euhedral cavity linings) and disruption of layering into metre-scale polygons, produced during the desiccation stage. The muddy interbeds are characterized by displacive growth of halite during the desiccation stage. Immediately below the surface of the pan the halite layers are ‘matured’ by repeated episodes of dissolution and diagenetic crystal growth. This results in porous crusts with patches of ‘chevron’ and ‘cornet’ crystals truncated by dissolution, clear diagenetic halite cement, and internal sediment. These layers of ‘mature’ halite closely resemble the patchy cloudy and clear textures of ancient halite deposits. Holocene salt-pans are known to cover thousands of square kilometres and cap halite deposits hundreds of metres thick, so they are realistic models for ancient evaporites in scale, e.g. Permian Salado Formation of New Mexico-Texas, which preserves many primary salt-pan features. zyxwvutsr INTRODUCTION Layered evaporites can accumulate in three basic depositional settings : (1) deep perennial (density stratified) basins, such as the Dead Sea; (2) shallow perennial lakes or lagoons, such as Great Salt Lake (Utah) and the Gulf of Karabogaz; and ( 3 ) ephemeral saline pans, such as Death Valley (California). Specific criteria for recognizing ancient evaporites deposited in these types of settings have yet to be formally documented. Such criteria must come from the study of modern environments using the method of comparative sedimentology. With this in mind one of these *Present address : Department of Geological Sciences, State University of New York at Binghamton, Binghamton, New York 13901, U.S.A. basic evaporite settings, the ephemeral saline pan, has been examined in detail. Ephemeral saline pan is used here as a general term for any evaporitic setting that is normally dry and underlain by layered salts. Hardie, Smoot & Eugster (1978) called such a setting a salt pan but since salt is equated with halite by many workers it is probably best to use saline pan as a general term and salt pan or halite pan for halitedominated systems. Other saline pans might also be identified by their dominant mineral, for example, trona pan, gypsum pan, etc. Ancient halite deposits may reach impressive thicknesses and cover wide areas, and for this reason the present study was aimed at halite-dominatedsaline pans, referred to hereinafter as salt pans or halite 627 628 zyxwvutsrqp T. K . Lowenstein and L. A . Hardie zyx zy zyxwvutsr pans. Eight active, modern salt pans were studied in both marginal-marine settings (Salina Omotepec, Baja, California, Mexico; Tell Defana, eastern Nile delta, Egypt) and continental settings (Saline Valley and Death Valley, California; Carlsbad Salt Lake, New Mexico; Lake Uyuni, Bolivia; Lake Bonneville, Utah; and Great Salt Lake, Utah, artificial evaporating pans). Extensive field observations were made over several seasons only for Salina Omotepec, Saline Valley, Death Valley and Carlsbad Salt Lake. From observations of these modern halite pans and their deposits a general set of criteria for recognizing ancient halite pan deposits has been compiled. The field observations of Hunt et a / . (1966), Shearman (1970), Stoertz & Ericksen (1974), Hardie et a / . (1978) and Handford (1982) were used to supplement the documentation of criteria for halite pan deposition. Additional evidence was derived from published experimental work on halite grown under conditions similar to those operating in salt pans (Wardlaw & Schwerdtner, 1966; Arthurton, 1973; Southgate, 1982). While the results presented in this paper are confined to halite pans, work is in progress on gypsum pans (Castens-Seidell & Hardie, 1983, 1984; CastensSeidell, 1984) because they are believed to be the major setting for gypsum-anhydrite caps to sabkha cycles found so abundantly in the geologic record. T H E S A L I N E P A N CYCLE Saline pans are characteristic features of arid basins. They are flat, shallow depressions floored with layered salts and are normally dry except when storm flooding turns the pan and its surrounding mudflats into a temporary lake (theephemeral saline lake subenvironment of Hardie et a / . , 1978, pp. 21- 23). Saline pans occur in both continental and marginal marine (sabkha) settings. Modern saline pans vary in size from tiny salt flats less than 1 km2 in area to giants thousands of square kilometres in area, such as the salt pan at Lake Uyuni, Bolivia (Rettig, Jones & Risacher, 1980), which covers about 10,000 km2. zyxwv THE SALINE PAN CYCLE STAGE 1. FLOODING (8) zyxwvuts zyxwvutsrqpon Fig. 1. (A) Oblique aeriai view of Saline Valley, California, looking north. The white area is the salt pan and its surrounding salt-encrusted saline mudtlats. Note the linear fault scarps and the apron of braided alluvial fans that reaches down to the pan. The Inyo Mountains that make up the western buttress of Saline Valley stand more than 3000 m above the floor of the salt pan. (B) Summary of the basic elements of the saline pan cycle. 630 zyxwvutsrqp zyxwvutsr zyxwvutsrq zyxwvuts T. K. Lowenstein and L. A . Hardie Continental saline pans occupy the lowest areas of closed arid basins (Fig. 1A). The flat, salt-encrusted pan is surrounded by a brine-soaked mudflat permeated with evaporite minerals that grew within the sediment, and this saline mudflat in turn grades outward into a dry mudflat and sandflat (Hardie et ul., 1978). A small but noticeable topographic break normally separates the sandflat from the sloping, gravelly surface of surrounding alluvial fans. Marginal marine saline pans occupy very shallow depressions on arid supratidal flats (marine sabkhas) and pass into saline mudflats and tidal flats on their seaward side. On their landward side marginal marine pans commonly grade into the same mudflat-sandflat-alluvial fan complex that exists in continental settings (Castens-Seidell, 1984; Castens-Seidell & Hardie, 1984; Gavish, 1974). The sediments found in saline pans consist of alternating layers of crystalline salts (of variable thickness but typically of a centimetre scale) and detrital siliciclastic-rich mud (typically a millimetre to centimetre scale). The mineralogy of the crystalline salt layers depends on the composition of the inflow waters. In most instances a single mineral dominates and not uncommonly the layers are monominerallic. Principal salts found in modern arid basin saline pans are halite (NaCI), gypsum (CaS0,.2H20), mirabilite (Na2S0,. 10HzO), thenardite (Na,SO,), epsomite (MgS04.7H,0) and trona (NaHCO, . NaZCO3.2H20),but saline pans in which halite is the dominant mineral are by far the most abundant. Modern potash-rich saline pans have not been described in detail but they have been reported from the Chaidam Basin of China (Yuan, Chengyu & Keqin, 1983). Study of these sylvite (KCl) and carnallite (KCI . MgClz.6 H 2 0 ) bearing basins could be of great importance in deciphering the origin of ancient potash evaporites. The key to understanding saline pan deposition lies in the recognition that such pans repeatedly cycle through a well defined sequence of stages that leaves a characteristic and legible petrographic record. This saline pan cycle is as follows: desiccation stage (dry saline pan) -+flooding stage (brackish lake) + evaporative concentration stage (saline lake) -+ return to desiccation stage (dry saline pan). The normal state is that of desiccation when the surface of the saline pan is hard and dry. The dry saline pan stage is interrupted by flooding during storms and, in some arid basins, during spring thaws in the surrounding snow-capped mountains. Flooding may occur as frequently as several times a year but most commonly major flood events are years and even decades apart. In continental basins the flood waters are dilute meteoric waters (rainstorm runoff and snow meltwater). Marginal marine saline pan3 are flooded both by meteoric waters (from rainstorms, etc.) and by seawater washed on to the pan by storm surges and unusually high spring tides. These floodwaters are ponded on the saline pan to form a temporary shallow brackish lake. The depths of these lakes are typically no more than a few tens of centimetres, but in some cases may reach the metre scale. For example, after the huge flood of 1949 the Lake Eyre (Australia) salt pan turned into a temporary lake over 8000 km’ in area, but the maximum depth wasonly4 m(Johns&Ludbrook, 1963).The inflowing dilute floodwaters partially dissolve the old surface saline crust, leaving an important record of the temporarily undersaturated conditions. The relatively insoluble salts such as gypsum are eroded, corroded and redeposited as rounded crystal clasts, such as has been observed at Salina Omotepec (Castens-Seidell & Hardie, 1983; Castens-Seidell, 1984). Fine-grained siliciclastic sediment brought in by the floods is deposited over the etched and eroded saline crust surface as a blanketing layer of mud. With time, the shallow ephemeral lake becomes concentrated by evaporation and the saline lake stage is entered. In NaC1-rich systems, continued evaporation concentrates the lake brines until they reach saturation with halite and a period of salt deposition follows. Crystallization starts at the brine surface as small plates and hopper crystals (Dellwig, 1955)which sink to the bottom. These serve as nuclei for further growth and widespread syntaxial overgrowth takes place on the lake floor, ultimately resulting in the development of chevrons (Wardlaw & Schwerdtner, 1966) and cornets (Arthurton, 1973). As evaporation and halite crystallization continue, the saline lake shrinks, ultimately dryingout altogether and the desiccation stage is reached. This is the normal state of the salt pan during which a dense, residual brine, saturated with halite and sustained by perennial groundwater inflow, permeates the salt and sediment layers just below the surface of the pan. This brine continues to evolve in response to evaporation (probably the evaporative pumping mechanism of Hsii & Siegenthaler, 1969),and this leads to the growth of void-filling and displacive halite within the vadose and upper phreatic zones. With time, pervasive growth of halite in the subsurface causes the surface of the pan to fracture and buckle by volume expansion, resulting in the development of a network of metre- Recognizing salt-pan eouporites 63 1 scale polygons rimmed by pressure ridges. A spongy efflorescence of halite then develops in the polygon cracks and continues to grow in response to evaporation of the subsurface brine drawn up the cracks. This phase of halite crystallization continues until the'next flooding event, at which point the cycle of stages is renewed. This saline pan cycle is summarized graphically in Fig. l(B). zyxwvutsrq zyxwvuts CRITERIA FOR RECOGNITION OF HALITE P A N D E P O S I T S Halite pan deposits are most effectively recognized by the specific petrographic features that result from the processes operating at each of the stages of the saline pan cycle (Fig. 1B).For this reason the criteria will be discussed on a stage by stage basis. (1) Dissolution and deposition features produced during the flooding stage The unequivocal signature of salt pan deposition is found in the dissolution textures andjabricsof the halite layers. It is these important features that indicate contact of halite layers with undersaturated waters and clearly distinguish the saline pan halite from that deposited in perennial subaqueous environments. At the start of a new cycle storm flooding of the dry salt pan by either meteoric waters or seawater causes dissolution of the exposed halite layer that floors the pan because these waters are considerably undersaturated with respect to halite. This produces a smooth dissolution surface across the top of the halite layer. Such dissolution surfaces are most dramatically displayed where the tops of vertically oriented chevron and cornet crystals are sharply truncated (Fig. 2A; see also Arthurton, 1973, fig. 9). With time, the ponded waters begin to dissolve the internalportions of halite layers. Dissolution pathways generally follow a pre-existing halite fabric and elongated cavities or pipes develop at grain boundaries between vertically oriented halite crystals (Shearman, 1970)(Figs2A and 3). Thesedissolution pipescontinue to grow a t the expense of surrounding crystals, and are most apparent where bands of fluid inclusions within chevrons and cornets are sharply truncated on their sides against dissolution voids (Fig. 2B). Halite is also preferentially dissolved in a pattern that follows original crystal growth banding, which, in its extreme, results in skeletal halite spines (Fig. 4). Such secondary zy zyxwv Fig. 2. Thin section photomicrographs of halite crust (depth <0.5 m) from Salina Omotepec, Baja California, Mexico (collected 1972). Dark banded parts of chevrons contain a multitude of tiny fluid inclusions, trapped during crystal growth. (A) Dissolution surface (between arrows) truncates tops of vertically oriented chevrons. Subsequent syntaxial overgrowth rebuilt the chevrons above the dissolution surface. Vertically oriented voids (V) are partly filled with clear halite cement. Scale bar is 2 mm. (B) Void (V) enlarged by dissolution of the framework of chevron halite (arrow). Clear halite forms syntaxial overgrowths (0)on the top and sides of the same crystal. Note muddy internal sediment at the base of the solution cavity in the upper right of the photo. Scale bar is 2 mm. textures are readily distinguished from primary skeletal halite crystals formed during very rapid growth (e.g. Southgate, 1982) by the rounded solution truncations. Further evidence of dissolution can also be found in the form of horizontal cavities that appear in plan view as a network oftubular vugs and in sectional view as elongated cave-like openings (Figs 5 and 6). The orientation of these cavities is inherited from 632 zyxwvutsrqp zyxwv zyxwv T. K . Lowenstein and L. A . Hurdie Fig. 3. Thin section photomicrograph of halite crust (depth <0.5 m), Salina Omotepec, Baja California, Mexico, showing vertically oriented solution cavities (dark tubular areas) between vertically oriented chevrons and cornets of halite (clear areas with faint bands of inclusions). Uppermost surface is a smooth horizontal dissolution surface. Dark granular zone at the bottom is a layer of gypsum crystal sand, containing siliciclastic mud and insect larval cases, and cemented with clear poikilotopic halite (light coloured areas). Scale bar is 10 mm. zyxwvuts zy Fig. 5, Close up view of the surface of the halite crust covering the Saline Valley salt pan (photograph taken 198 1). The circular and worm-like markings are tubular dissolution cavities in the crust. The white ridge in the upper right of the picture is part of a polygonal fracture. Pencil (0.18 m) for scale. Fig. 6. Halite crust (< 0. I m depth, see Fig. 9) from the Saline Valley salt pan, showing large horizontal dissolution cavities with remnant vertical pillars. Note that the large cavity at the bottom centre of the picture (beneath the prominent pillar) is lined with euhedral halite cement and is floored with internal muddy sediment (black). Coin is 20 mm in diameter. zyxw Fig. 4. Thin section photomicrograph of surface halite crust, near Tell Dafana on the shores of Lake Manzala, eastern Nile delta, Egypt. The large chevron crystal with fluid inclusions bands (centre) shows preferential dissolution (clear rounded voids, labelled V) along original crystal growth banding, giving a ‘fishbone’ pattern, scale bar is 1 mm. horizontal shelter voids and layer boundaries in the original halite fabric. A more subtle dissolution feature observed in coarse, vertically oriented chevron crystals is the rounding of top crystal faces (Fig. 7). This may result from the settling of heavy night-time dew over the salt pan when the top portions of halite crystals are exposed to the atmosphere (D. Shearman, 1983, pers. comm.), or it may be the result of local flooding by small storms. Finer grained halite crystallized at the air-brine interface and which then settles to the bottom also exhibits distinct dissolution features. For example, layers with accumulations of laterally-linked crystal Recognizing salt-pan eiuporites rafts develop horizontal cavities by solution enlargement of pre-existing shelter voids beneath settled rafts (Fig. 8). Individual plates, pyramidal hoppers and crystalscomprising rafts are all rounded by dissolution. Where no growth fabric exists, randomly oriented networks of tubular voids develop. For such finely crystalline halite, repeated episodes of solution results in the eventual destruction of the original crystal textures. In this manner, evidence of bottom accumulations of fine-grained plates, cubes and coalesced rafts may be selectively removed from halite layers leaving only coarser, less rapidly dissolved chevrons and cornets as the dominant constituents. As a result of dissolution a vuggy crystalline halite layer is formed with porosity commonly up to 50%. In addition to dissolution, the flooding stage leaves a record of deposition. As noted above, halite pans are layered, typically on a centimetre scale, with alternations of detrital mud and crystalline halite. For example, in the top 0.25 m of the salt pan of Saline Valley, California (Hardie, 1968; Hardie et a[., 1978), there are up to six layers of dark brown, clay-rich mud sandwiched between crystalline halite layers (Fig. 9). Each mud layer presumably was deposited in the shallow ephemeral lake that formed when muddy floodwaters inundated the pan during a major storm flooding stage, as has been observed by the authors in Salina Omotepec. In the Saline Valley example the detrital interbeds are dominated by mud (silt and clay) zyxwv 633 zyxwvu zyxwvutsrq zyxwvuts zyxwvut zyxwvu Fig. 8. Thin section photomicrograph of surface halite crust from the artificial evaporation pans of Great Salt Lake, Utah. The lower half of the crust is composed of cumulates of long, flat halite rafts (between arrows). The horizontal cavities (V) were formed by dissolution and enlargement of original shelter voids beneath the rafts. The dark circles are air bubbles trapped during sample preparation. Scale bar is 5 mm. Fig. 9. Vertical cross-secticn through the uppermost 0.1 m of the Saline Valley salt pan (January 1984) showing the interbedding of halite (white) and siliciclastic mud (dark) so characteristic of salt pans. The mud layers contain halite euhedra that grew displacively. The halite layers are full of dissolution cavities. The thicker halite layer is actually a compositeof three separate halite crusts, each distinguishable by slight differences in the pattern of vertical growth and dissolution cavities. Coin is 18 mm in diameter. Fig. 7. Thin section photomicrograph of shallow subsurface halite crust ( < 0 . 3 m depth) from Salina Omotepec, Baja California, Mexico. Chevrons with fluid inclusion bands have rounded coigns (such as the one indicated by the arrow) repaired by syntaxial overgrowth. The open voids (dark grey) are floored by internal sediment (black). Some of the clear void-filling halite cements (A) contain patches of internal sediment (black within clear halite) incorporated during growth. Scale bar is 2 mm. rather than sand. This is typical of many salt pans in both marginal marine and continental settings and it emphasizes that during flooding the bedload and coarse suspension load are dropped in the surrounding subenvironments (sandflats, beach ridges, etc.) and only the finer suspension load is carried on to the pan itself. Coarse sediment is found in some saline pans but it is usually composed of wind-blown sand, organic debris, etc., and locally derived clasts made either of 634 zyxwvut zyxwvutsrqpo zyxwvut T. K . Lowenstein and L. A . Hardie mud aggregates or of the less soluble salts precipitated in the pan, such as gypsum. As the dilute floodwater sweeps over the dry pan, halite is dissolved but any gypsum present is reworked and redeposited as a lag of corroded and etched gypsum crystal sand. This is an important feature of the marginal marine saline pan of Salina Omotepec, Baja California (Shearman, 1970; Castens-Seidell, 1984), where layers of rounded clastic gypsum crystal sands containing insect larval cases and other detrital material are interlayered with halite crusts (Figs 3 and 18; Castens-Seidell & Hardie, 1983; Castens-Seidell, 1984). Whatever the source of the detrital sediment, some of the transported grains will be washed into dissolution cavities and other voids in the underlying halite crust to make pockets of internal sediment (Fig. 7). The bulk of the transported sediment, however, comes to rest as a detrital layer separated sharply from the underlying halite layer by a dissolution surface, as described above. Excellent ancient analogues of this are found in the Devonian Prairie halite of Canada (Wardlaw & Schwerdtner, 1966). Minor flooding is a much more common occurrence on saline pans than major flooding, and can create a temporarily undersaturated saline lake without deposition of a detrital mud layer. Sediment-free water may alsg be fed to the lake from springs along the edges of the pan (e.g. Saline Valley). These episodes of sediment-free-flooding are recorded within the halite crust as centimetre or millimetre scale layers separated by distinct horizontal dissolution surfaces (Fig. 2A). Repetition of such flooring can result in decimetre scale halite crusts lacking detrital mud partings. This situation is probably common in the central portionsof very large salt pans where only rare catastrophic floods succeed in transporting mud to the centre of the pan (e.g. Lake Uyuni, Bolivia). In contrast, the thickest and largest number of mud layers separating crystalline halite are observed on smaller salt pans such as in Saline Valley, or on the margins of larger pans in close proximity to a detrital source. phase of the salt pan cycle begins. Initially, halite crystallizes at the air-brine interface as millimetre to centimetre-sized pyramidal hoppers (Fig. 10; see also Dellwig, 1955), and rectangular- and square-shaped plates (Fig. 11A; see also Arthurton, 1973). The crystals are commonly flattened and horizontally elongated, a feature attributed by Phillips (1956) and Llewellyn (1968) to growth in a thin layer of supersaturated brine at the air-water interface. During periods of rapid growth, skeletal hoppers are formed by preferential development of cube edges over cube faces. Such rapid cube edge growth occurs in supersaturated surface brines where filling of cube faces cannot be completed before supersaturated conditions cease (Cooke, 1966; Llewellyn, 1968), and results in the trapping of a multitude of tiny fluid inclusions along growth bands that can be seen under the microscope as cloudy laminae (Figs 11-14). While floating, crystals may coalesce and become weakly attached to form rafts which then continue to grow (Fig. 11B; see also Shearman, 1970; Arthurton, 1973). Eventually, the surface grown halite sinks to the bottom of the shallow brine lake when the combined effects of increasing weight together with gusts of wind overcome the effects of surface tension that buoy up the crystals. The sinking crystals accumulate to form layers composed of horizontally aligned rafts, plates and randomly oriented hoppers (Fig. 12). These accumulations may thicken in depressions and thin over the tops of chevron crystals in the underlying halite crust. With time, slower crystallization continues on the bottom and accumulated cloudy skeletal crystals zyxwvutsrq (2) Crystal growth features produced during the saline lake stage The initial flooding of the pan results in the establishment of a shallow brackish lake. This is followed by a period when the shallow lake waters become concentrated both by evaporation and dissolution of halite. When halite saturation is reached, the crystal growth zyxwvu Fig. 10. View looking straight down on to the surface of a small brine pool on the salt pan at Salina Omotepec, Baja California, Mexico. Four-sided skeletal hoppers of halite float on the brine surface as inverted pyramids held up by surface tension. Knife is 90 mm long (phototaken by R. W . Mitchell). zy zyxwvuts Recognizing salt-pun euuporites 635 zyxwvu zyxwv zy zyxwvutsrqp Fig. 12. Thin section photomicrograph of surface halite crust from Salina Omotepec, Baja California, Mexico, showing alternating layers of fine-grained cumulates of halite (sunken rafts, plates, hopper cubes) and coarse-grained, vertically oriented chevrons and cornets that grew upwards from the saline lake bottom. Dark grey areas are voids. In the chevrons, bands of fluid inclusion can be traced stratigrdphically across crystal boundaries from one crystal to the next. Figure 11(A) is taken from the lower left corner and Fig. 11(B) from the upper left corner of this thin section. Scale bar is 10 mm. Fig. 11. Thin section photomicrographs of a cross-section of a pristine surface crust from Salina Omotepec, Baja California, Mexico, showing flattened rectangular- and squareshaped plates (1 1A) and laterally linked rafts (1 1 B). These originally grew at the surface and then sank to form a horizontally oriented, grain-supported framework (see Fig. 12). Note that the halite crystals are clouded with bands of fluid inclusions and that a large amount of void space (dark grey areas) is preserved. Scale bars are 2 mm. become filled with clear halite free of fluid inclusions (Figs 11 and 13; see also Shearman, 1978). These modified rafts, plates and hopper cubes lying on the brine bottom then become nuclei for widespread syntaxial growth into the evaporating brine (Fig. 13). As bottom growth progresses, competition for space forces preferential upward growth and a layer of vertically oriented and elongated crystals is produced (Fig. 14). These crystals are zoned by alternating cloudy bands rich in tiny fluid inclusions and clearer bands relatively free of inclusions. Such zoning is probably the result of varying crystal growth rates where cloudy bands form rapidly during periods of intense evaporation and clear bands crystallize slowly during periods of lower evaporation rates (Shearman, 1978; Holser, 1979; Roedder, 1984). The morphology Fig. 13. Thin section photomicrograph of surface halite crust from Lake Bonneville, Utah, showing a raft (centre) and a flattened plate (arrow on the right) which acted as nuclei for syntaxial growth of chevrons (A) and cornets (C). Note the downward syntaxial growth into the void beneath the raft (arrow on the left). Both the host crystals and their overgrowths show typical primary growth layering accented by dark bands of fluid inclusions. Note the streaky clear patches with hazy boundaries within the dark bands (as for example in the chevron on the extreme left). These clear patches are characteristic of primary growth layering and represent differential exclusion of fluid inclusions from an accreting layer during growth of the crystal (mechanism unknown). They are different from sharp edged patches of secondary halite cement. Scale bar is 2 mm. 636 zyxwvutsrqp zyxwv T. K . Lowenstein and L. A . Hardie vessel as a cumulate. No significant syntaxial overgrowth of the settled cumulates took place during periods of rapid evaporation. This is probably because it was prevented by the continuous settling of fresh cryst'als on the bottom, which swamped any upward growth. Without the fan, lamp and dehumidifier, a lower rate of evaporation was produced and surface nucleation and settling of crystals was greatly suppressed. This allowed time for syntaxial overgrowth of cumulates to take place, and over a period of several weeks cornets and chevrons of about the same size as found on modern salt pans were formed (see also Wardlaw & Schwerdtner, 1966 and Arthurton, 1973). It would appear, therefore, that surface grown halite develops in response to rapid evaporation and can crystallize, sink and accumulate in a matter of hours, whereas bottom growingchevrons and cornets develop in response to slower evaporation spanning a period of several days to weeks. Comparable suites of crystals are found in modern salt crusts (such as from Salina Omotepec, Baja California, Fig. 12), documenting that a detailed record of conditions prevailing during desiccation at the shallow saline lake stage is preserved in salt pan crusts and can be used in the interpretation of ancient halite. The time involved for the saline lake stage, from initial flooding to complete desiccation, is quite variable and depends on the following factors: ( I ) the initial depth of the saline lake; (2) the rate of evaporation, which is a function of air temperature, relative humidity, brine salinity, and wind velocity; and (3) the amount of subsurface fluid supplied from groundwaters after flooding. The lifespan of modern ephemeral saline lakes ranges from weeks to at most a few years. Friedman, Smith & Hardcastle (1976), for example, studied Owens Dry Lake in south-eastern California, from its initial flooding stage until it was nearly dried out. From January to August 1969, waters from the Owens River emptied into an initially dry lake bed, creating a standing body of water up to 2.4m deep that covered an area of over 100km' (Friedmanetal., 1976). By August 1971, after a period of about two years, the lake had virtually dried out. At the other end of the spectrum, a far shorter saline lake interval was observed in the Carlsbad Salt Lake located about 15 km east of Carlsbad, New Mexico. On 9 June 1980, the pan was flooded to a depth of up to 0.1 m and the underlying salt crust had dissolved (Fig. 15A). By 4 July 1980, after a period of less than one month, the saline lake had disappeared, leaving a 20 mm thick carpet of halite crystals (Fig. 15B). This salt pan, however, has been modified by channel zyxwv zyxwvutsrqp Fig. 14. Thin section photomicrograph of surface halite crust from an artificial evaporation pan at Great Salt Lake, Utah, showmg typical competitive growth habit of halite that results in a vertical crystalline fabric. Growth layering is traced by bands of fluid inclusions (dark) that outline chevrons (upward pointing bands) and cornets (flat, upward widening bands). Scale bar is 5 mm. of the upward growing crystals is dependent on the attitude of the parent crystals. Where syntaxial overgrowth begins on a halite cube lying on edge (e.g. a randomly oriented hopper), the resulting overgrowth will be chevron-shaped with an upward pointing coign (Figs2,4,13and 14; seealso Wardlaw&Schwerdtner, 1966). But if the parent crystal is oriented with an upward facing cube face (e.g. horizontally oriented rafts and plates), syntaxial overgrowth results in a flattopped, upward widening cornet-shaped crystal (Figs 13 and 14; see also Arthurton, 1973). In some cases, shelter voids preserved beneath the rafts also provide space for competitive syntaxial growth of crystals, resulting in downward facing cornets (Fig. 13). Layers of freshly crystallized salt pan halite, not modified by other salt pan events, thus consist of a basal zone of surface-nucleated hoppers, rafts and plates overlain by syntaxially grown chevrons and cornets (Fig. 12). The relative proportions of surface nucleated and bottom grown crystals probably depend on the rate of evaporation of the saline lake. This is suggested by a series of laboratory experiments in which halite-saturated brine 0.15 m deep was evaporated at varying rates from large glass vessels. Evaporation was encouraged by the combination of a fan blowing continuously over the brine surface, a heating lamp placed directly above the brine, and a room dehumidifier. With fan, lamp and dehumidifier all working in combination, evaporation was rapid and plates, hoppers and rafts formed continuously at the surface and settled to the bottom of the reaction Recognizing salt-pan evaporites zyxwv zy 637 line solution cavities and intergranular voids (Fig. 16 and Shearman, 1970,fig. 6). Theseclear halite cements then transform the rounded solution cavities into angular, partly cemented vugs. Chevrons and cornets that have been rounded by earlier phases of dissolution may also be repaired at this stage by syntaxial overgrowth (Figs 2 and 7), and the dissolved internal portions of crystals may also become filled with clear halite. In all cases, a sharp, often curved, dissolution boundary separates the clear halite cement from earlier formed cloudy halite. This type of cement should be differentiated from a similar clear halite texture that exists in the internal portions of some primary crystals where cloudy, fluid inclusionrich halite passes laterally along the length of an individual growth band into clear, fluid inclusionpoor halite (Fig. 13). These transitions along a single growth band are not marked by sharp dissolution truncations, and they result from differential incorporation of fluid inclusions during halite crystal growth (mechanism unknown). Displacive growth of halite may also take place at this stage in muddy and sandy layers located at or beneath the brine table. The resulting crystals are either isolated euhedra or interlocking aggregates of randomly oriented cubes. Host sediment is commonly partially incorporated along growth faces. This process is particularly important in the shallow subsurface mud of the Saline Valley salt pan (Fig. 17), and in the gypsum sand layers beneath the Salina Omotepec salt pan (Fig. 18; see also Shearman, 1970). The continued growth of halite cement immediately beneath the dry surface of the pan during this stage causes lateral expansive growth of the surface crust, and leads to disruption of the crust into metre scale zyxwvutsrqp zyxwvuts Fig. 15. Salt pan at Carlshad Salt Lake near Carlsbad, New Mexico. (A) View of the pan flooded to a depth of 0.1 m, 9 June 1980. (B) View of the pan after the lake had dried out leaving a 20 mm thick layer of halite, 4 July, 1980. Car tyre for scale. dredging for a small salt mining operation, and therefore some waters may have drained into the channels. (3) Syndepositional diagenetic growth features produced during the desiccation stage The terminal stage of the saline pan cycle is reached on the complete desiccation of the surface brine body. However, halite crystallization continues as the newly formed crust, together with underlying salt crusts, gypsum and mud layers, undergo a phase of diagenetic growth of halite from the residual brine that lies beneath the surface of the porous crust. This phase of growth produces clear halite with no preferred crystal orientation, growth direction or elongation with respect to bedding and is readily distinguishable from the cloudy halite formed during the saline lake stage (Shearman, 1970). Thisdiagenetic halite forms inward growing euhedrally terminated cubic crystals which Fig. 16. Shallow subsurface halite crust from Salina Omotepec, Baja California, Mexico, showing a horizontal solution cavity lined with clear halite euhedra that grew inward as a cement. Coin is 20 m m in diameter. 638 zyxwvutsrqp zyxwvutsr zyxwv T. K . Lowenstein and L. A . Hurdle zyxwvutsrqp zyxwvutsr Fig. 17. Thin section photomicrograph of halite in a muddy layer from just beneath the surface of the salt pan of Saline Valley, California. The halite occurs as randomly oriented interlocking cubes with incorporated mud (black) and patches of fluid inclusions (arrows). Aggregates of interlocking small rhombs of glauberite (CaSO, . Na2S0,) are present at the top of the picture. Scale bar is 5 mm. Fig. 19. View of desiccated surface of the halite pan, Saline Valley, California, showing polygonal cracks filled with a fluffy white halite efflorescence. Even at this early stage of disruption the edgesof polygonsare upwarped and overthrust on to neighbouring polygons to produce pressure ridges. It can also be seen that the propagating tips of newly formed cracks are overthrust, so that compression must be the driving force for the disruption of the crust (cf. the contractional model of Tucker, 1981). Presumably this compression results from expansive cementation (as in tepee structures in carbonates). The surface of the crust is pockmarked with small dissolution cavities (see Fig. 5 for a close up view). Pencil (0.18 m) for scale. ridges (see Eugster & Hardie, 1978, fig. 7b), and results in buckling of the polygons to form dished lenses with tepee-like lateral contacts. Preservation of these features, however, is uncommon because flooding preferentially removes the porous upturned edges of the polygons before subsequent layers of sediment can be deposited. The cracks themselves normally survive the flooding and may even become enlarged by dissolution and filled with mud. This preserves the cracks, and comparable features have been found in ancient halite deposits (Dellwig, 1968; RichterBernburg, 1980; Tucker, 1981). When dissolution by flooding does not completely remove the polygon ridges but simply lowers and rounds them, the pan is left with an uneven surface of dished polygons. This influences the deposition of the subsequent mud layer to produce a discontinuous, lenticular bedding, such as can be seen in the Death Valley salt pan at Devil’s Golf Course. In Saline Valley, the old polygonal cracks have determined the location of the new cracks in the overlying surface halite crust and stacked polygons result. Ancient examples of these can be seen in the Permian Salado evaporites of Texas and New Mexico (Lowenstein, 1983, pp. 150-151) and in the Triassic halite of Cheshire, England (Tucker, 1981). zyxwvuts Fig. 18. Thin section photomicrograph of a gypsum sand layer cemented by halite from beneath (<0.5 m) the pan of Salina Omotepec, Baja California, Mexico. The large clear crystals of halite (light-coloured areas in the upper right and near scale bar) poikilotopically enclose gypsum prisms, mud and insect larval cases. Uncemented areas appear dark grey. Scale bar is 2 mm. polygons rimmed by pressure ridges that override each other like tectonic thrusts (Fig. 19). It is possible that initially the crack pattern is produced by thermal contraction as suggested by Tucker ( I 98 I), but even the tips of newly propagated cracks are overthrust so that compression must be a dominant force from the earliest stages of rupture (Fig. 19). Preferential evaporative pumping of subsurface brine takes place along the cracks between the polygons and leads to precipitation of a spongy efflorescent halite (Fig. 19). This promotes further elevation and distortion of the zyxwvuts zyxwv zy Recognizing salt-pan evaporites One final note about diagenetic growth of halite concerns the occurrence of meniscus cements. In laboratory experiments well-defined halite meniscus cements have been produced during the last stages of desiccation. These consisted of fine-grained mosaics which lined the intergranular voids between bottom accumulated halite cubes and bridged the voids at points of contact between the cubes. As a result the angular void spaces were converted into rounded pores typical of vadose cementation processes. Unequivocal evidence of the existence of this type of cement has not been found so far in ancient halite deposits. The rounded voids which are typical of salt pan halite (Figs 2, 4 and 8, for example) look very much like the bubble-shaped pores left by meniscus cements. However, close examination of the clear halite lining the walls of these voids reveals a patchy, irregular distribution that conforms more to the pattern expected for dissolution than for surface tension influenced cementation. This is especially obvious in Fig. 2(B) where the scalloped walls of the voids irregularly cut across both chevron hosts and clear overgrowths of halite. These rounded voids are undeniably solution vugs. This example is characteristic of all the salt pan halites examined in this study. Nevertheless, it remains likely that some meniscus cement forms during the initial stage of desiccation of the surface crust but it is probably modified or removed by dissolution or masked by a cover of phreatic cement. 639 zyxwvuts T H E P R O D U C T I O N OF HALITE ROCK I N A S A L I N E P A N Mature saline pan halite Ephemeral halite pan evaporites consist of dkontinuous centimetre-scale layers of porous halite that alternate with layers of mud, and in some cases, gypsum layers. As demonstrated above, ‘first cycle’ crusts of freshly deposited halite preserve a wealth of information on the processes by which halite layers form by settle-out of surface nucleated crystals and competitive syntaxial bottom growth (Fig. 12). However, it is the older, buried crusts, modified by many saline pan cycles of dissolution and diagenetic crystal growth, that are the most realistic analogues for comparison withancient halite deposits. Such ‘mature’ salt pan halite exhibits only patchy preservation of chevrons and cornets that are dissolutionally truncated, rounded or corroded (Fig. 20). Mud and fine zyxwvu Fig. 20. Thin section photomicrograph of a ‘mature’ halite crust from just beneath the surface of the salt pan of Salina Omotepec, Baja California, Mexico, showing characteristic patchy remnantsof chevronsand cornets (C), cleardiagenetic halite (D) and voids (V). Note the halite crystal rafts (R) and the fine gypsum and mud internal sediment (arrows). Scale bar is 10 mm. grained gypsum prisms occur as internal sediment in vugs between halite crystals. Surface nucleated cumulates of fine halite crystals are only rarely preserved, but rafts seem to be able to survive dissolution and crystal overgrowth and some remnants of the flattened raft texture are usually preserved (Fig. 20). However, the bulk of a mature crust consists of dissolution cavities, and clear diagenetic halite in the form of overgrowths and void-lining cement which together account for well over half the rock volume (Fig. 20). In ancient halite evaporites porosities are commonly very low, and it remains to be shown therefore how such porous surface crusts become tightly crystallized. Compaction is probably not the major force responsible for eliminating voids in buried halite evaporites because there is a general lack of pressure solution boundaries and stylolites, and primary textures are preserved in an undeformed condition. Clear halite cements, however, occur in the halite of the Permian Salado Formation and in shallow buried (up to more than 200 m) Quaternary halite deposits from arid basins in the western U S . It is concluded, therefore, 640 zyxwvuts zyxwv zyxwvutsrqp zyxwvutsr zyxwvut T. K . Lowenstein and L. A . Hurdie that pore spaces in salt pan halite are filled with halite cement, but it is not yet possible todistinguish cements precipitated during the desiccation stage of the saline pan cycle from those formed during burial. Relationship of clear and cloudy halite in ancient halite rock The patchy distribution of cloudy halite separated by clear halite is a characteristic feature of ancient halite (Fig. 21; Roedder & Belkin, 1979, figs 1 and 3, Lowenstein, 1982, figs 6c, d ; Roedder, 1984, figs 9 and 12 and Hardie, Lowenstein & Spencer, 1985, fig. 8 for the Permian Salado Formation; Wardlaw & Schwerdtner, 1966, plate 2, fig. 4, and plate 3, fig. 3 for the Devonian Prairie Evaporites of Canada; and Dellwig, 1955, figs 7 and 11 for the Silurian Salina salts of Michigan). In these deposits, sharp and commonly curved boundaries divide clear halite from remnants of cloudy halite. This cloudy halite has been interpreted as syndepositional (surface nucleated hoppers and rafts, or bottom grown chevrons and cornets), but the origin of the clear halite has been a matter of debate. Some workers have considered that during burial, cloudy primary halite has been altered to clear halite by recrystallization or replacement (Wardlaw & Schwerdtner, 1966; Roedder & Belkin, 1979; Roedder, 1984). On the other hand, Dellwig (1955) interpreted both clear and cloudy halite as primary, but grown under different conditions controlled by variations in brine temperature. Shearman (1 970), however, noted that clear halite cement partially fills the void spaces in modern salt crusts, and that if these voids were filled completely by such cement, a texture identical to that developed in the Devonian Prairie Formation would be produced. Shearman interpreted truncation of cloudy halite against clear halite in the Prairie Formation as a syndepositional feature produced partly by dissolution and partly by diagenetic growth of clear halite. He thus pointed out that clear halite may be a syndepositional cement, formed during desiccation stage diagenesis and shallow burial, rather than the product of recrystallization. Without doubt burial recrystallization and replacement produce clear halite, but usually this completely eliminates cloudy halite textures from the rock (Kiihn, 1968, plate 2, fig. 2; Lowenstein, 1983; Roedder, 1984; Hardie et ul., 1985). The Permian Salado Formation, for example, contains halite rock with unequivocal annealing recrystallization textures (Fig. 22 ; see also zyxwvutsrqp Fig. 21. Thin section photomicrograph of halite from the Permian Salado Formation, New Mexico, showing fluid inclusion rich chevrons and cornets (dark areas) with truncated tops (T) and sides (S) separated by clear halite. Grey spots are air bubbles trapped during sample preparation. Scale bar is 2 mm. Fig. 22. Thin section photomicrographof recrystallized halite with 'foam' texture from the Permian Salado Formation, New Mexico. Note the equigranular mosaic texture and the curved crystal boundaries meeting at triple junctions with angles approaching 120". Mud impurities in most cases have been purged to the grain boundaries as is typical of annealling recrystallization. Small bubbles scattered throughout were artificially trapped during thin section preparation. Scale bar is 5 mm. zyxw zy zy zyxwvu zyxw Recognizing salt-pan evaporites Roedder, 1984, fig. I I ) composed of equigranular mosaics of crystals with boundaries that meet at triple junctions of about 1 2 0 angles. Also, fluid and solid inclusions are purged to grain boundaries (Hardie et al., 1985), so that the depositional texture is completely eradicated. These features must almost certainlyresult from annealing due to increased temperatures and pressures of burial whereby grains ‘optimize’ their size, shape and orientation in order to minimize energy in the manner of bubbles in foam (Stanton & Gorman, 1968). The features described here from a variety of modern salt pans fully support Shearman’s views of cloudy versus clear halite relations. As indicated above, dissolution is an important process in the modification of the primary crystalline framework of cloudy chevrons, cornets, cubes, plates and rafts, and is critical for differentiating between a cementation or recrystallization origin of the clear halite in ancient salt pan evaporites. ‘Pristine’ cumulates of cubes, plates and raftsshow anopen, grain-supporteddetrital texture with typical point- and long-contacts and a very high primary porosity (Figs 11 and 12). ‘Pristine’ bottom growth of chevrons and cornets results in a crystalline intergrowth texture with compromise boundaries that are planar or stepped (Figs 12, 13 and 14). These grain boundaries are characteristic of primary competitive growth into a void (cf. Bathurst’s 1975 criteria for cements). Some interpenetrating boundaries between cloudy crystals in these bottom grown layers appear curved (Fig. 12) but close examination shows that the contacts in fact consist of a series of small, rectilinear steps. These must result from competitive growth between neighbouring crystals that underwent simultaneously enlargement in successive pulses, as recorded by the ‘stratigraphy’ of the internal banding that can be correlated across grain boundaries (Fig. 12). Consequently, if the grain boundaries of primary growth crystals are rectilinear, it follows that smoothly curved boundaries between cloudy and clear halite must be formed by some secondary process. In modern salt pans this process is preferential dissolution along primary grain boundaries, resulting in new intergranular voids and enlargement of existing ones, followed by growth of clear halite cement into these voids. As a result cloudy halite is brought into sharp contact with clear halite along scalloped boundaries (either smoothly rounded or irregularly rounded) that cut across the internal banding of the cloudy crystals (Fig. 2). This boundary criterion then, can, be used to interpret the origin of clear and cloudy halite in the 641 geological record. For example, in the Permian Salado halite shown in Fig. 21, the cloudy chevron crystal on the right is penetrated by fingers of clear halite (labelled S in the photo) that have smooth rounded boundaries cutting across the primary internal banding. Such boundaries are identical to those in modern salt pan halite, and give strong weight to the interpretationof the clear patches in the Salado cloudy halite as cements that fill dissolution cavities in salt pan crusts. A compilation of the many other similarities between the Salado and modern salt pan halite that support this view are given below. Backed by such evidence, the problem of the Salado halite reduces to whether the clear halite was emplaced as an early syndepositional cement or as a later burial cement, or both. Other ancient deposits with cloudy and clear halite can be fruitfully subjected to the same petrographic testing. Application to the rock record The Permian Salado Formation of West Texas and New Mexico is a good example of ancient salt pan halite and can be used to demonstrate the similarity between modern and ancient deposits. The formation extends over an area of about 150,000km2 and is up to 700 m in thickness (Jones, 1972). Observation of mine slope sections and borehole cores from the Carlsbad Potash District has shown ‘that the Salado consists of metre scale stacked cyclical sequences (Jones, 1972). Complete cycles display the following sequence from bottom to top : mud -+ laminated anhydrite and polyhalite -+ mud-free halite muddy halite (Lowenstein, 1982,1983). Haliterock comprises almost 90% of the succession and is interpreted as having formed in a shallow lake/lagoon environment that evolved in each cycle to a salt pan and saline mudflat setting (Lowenstein, 1982,1983). These halite beds contain features that are strikingly similar to modern salt pan halite deposits. These are listed below in order of their appearance in the saline pan cycle: -+ Flood stage; dissolution (1) Smooth horizontal dissolution surfaces that truncate chevron halite crystals (Lowenstein, 1982, fig. 6d). (2) Internal rounding of chevron and cornet crystals caused by dissolution (Hardie et al., 1985, fig. 8). (3) Cloudy chevron and cornet crystals penetrated on their margins by fingers of clear halite (Fig. 21). Smooth dissolution boundaries truncate cloudy bands and separate them from clear halite cement. 642 zy zyxwvutsr zyxwvutsrq zyxwvutsrq T. K . Lowenstein and L. A . Hardie Saline lake stage: crystal growth (1) Cumulates of randomly oriented hopper cubes and laterally linked rafts (Fig. 23; Lowenstein, 1982, figs 7b, c, d). (2) Vertically oriented and elongated chevron and cornet crystals (Figs 21 and 23; Lowenstein, 1982, fig. 6c; Hardie et al., 1985, fig. 6). Desiccation stage : syndepositional diagenetic growth (1) Displacive halite cubes in muddy layers (Lowen- stein, 1982, fig. 8a, b; Hardieetal., 1985, fig. 12). (2) Syntaxial rehealing of rounded, partially dissolved chevron halite (Fig. 23). (3) Metre-scale, dish-shaped structures in thin bedded halite with mud-rich layers which appear in crosssection as buckled polygons. (4) Clear overgrowths of halite (Fig. 23). These features clearly demonstrate the similarities between modern and ancient halite and they indicate that modem salt pans are realistic models for the formation of ancient halite. They also confirm that primary salt pan features can be expected to be preserved in halite rocks after burial. C O N C L U D I N G REMARKS Ephemeral saline pans and their associated fringing evaporitic environments are by far the most common setting for the deposition of Holocene evaporites. Whether closed continental or marginal marine in origin, these barren, salt-encrusted surfaces undergo a unique cycle of events that is recorded in the accumulated salts and interlayered sediments (Fig. IB). Most modern saline pans are small compared to the great expanses underlain by the ancient ‘saline giants’, with areas covering hundreds of thousands of square kilometres. However, some modern deposits cover areas of thousands of square kilometres, such as Lake Uyuni in Bolivia and the Great Desert of Iran. Lake Eyre in Australia contained a flood stage ephemeral lake that extended over an area of 8000 km2 before drying out (Bonython, 1956). Finally, though not strictly a saline pan environment, the salt encrusted mudflats of the Ranns of Kutch, India, are flooded annually by monsoons over an area of 30,000 km2 (Glennie & Evans, 1976). Perhaps more impressive is the vertical buildup of saline pan deposits, which, in the Quaternary alone, reach thicknesses of over 300m and thus rival the thickness of the Permian evaporites such as the Salado Formation or the Zechstein salts (Hardie, 1984, fig. 2). Both Death Valley and Bristol Dry Lake in California contain a halite dominated succession.with interbedded mud that is over 300 m in thickness (Hunt et a/., 1966; Bassett, Kupfer & Barstow, 1959). Furthermore, in Searles Lake, California, a bore hole reached quartz monzonite bedrock at over 900 m and contains Quaternary evaporites to a depth of over 500 m which rest on lacustrine muds and alluvial fan gravel (Smith et al., 1983). Thus, modern saline pans are clearly realistic models for deposition of ancient evaporites in process and scale. This conclusion is supported by comparison of the features of the halite of the Permian Salado Formation with those observed on modern saline pans. The possibility that many other ancient ‘giant’ halite evaporites may be palaeosalt-pan deposits is worth testing using the criteria presented in this study. zyxwvutsrq Fig. 23. Thin section photomicrograph of Permian Salado Formation halite showing two vertically oriented halite chevrons (bottom) overlain by a mosaic of small halite cubes. The chevrons, dark with fluid inclusions, have been truncated by a dissolution surface and overgrown syntaxially by a thin band of clear halite (arrows). The mosaic of small cubes settled as a cumulate layer on top of the clear halite substrate. Scale bar I S 2 mm. ACKNOWLEDGMENTS We would like to express our gratitude to our field colleagues Bobbi Castens-Seidell, Joe Smoot, Ron Spencer, Chris Haley, Bob Demicco and Ray Mitchell for their help and advice on numerous field trips to zyxwvu zyxwv zy Recognizing sult-pan rvaporites the desert basins of the 1J.S. We would also like to thank Van 'Treat of Conoco for help w i t h the photography, Kevin Benedict of Conoco for drafting Fig. l(B), and Kate Francis for her patient managementofthe typingchores. Dave Clark'scritical review and editing of an earlier draft led to major improvement in the style of prescntation. Finally, special acknowledgment goes to Doug Shearman o f Imperial College, London, for his pioneering work on salt-pan halite which provided so clear a path for us to follow. 643 GLENNIE,K.W. &EVANS,G . (1976) A reconnaissanceof the Recent sediment of the Ranns of Kutch, India. Sedimentology, 23,625-647. HANDFORD, C.R. (1982) Sedimentology and evaporite genesis in a Holocene continental-sabkha playa basin--Bristol Dry Lake, California. Sedimentology, 29,239- 253. HARDIE,L.A. (1968) The origin of the Recent non-marine evaporite deposit of Saline Valley, Inyo County, California. G'euchim. Cosmochim. Acta. 32, 1279-1301. HARDIE,L.A. (1984) Evaporites: marine or non-marine? Am. J . Sci. 284, 193-240. HARDIE,L.A., SMOOT,J.P. & EUGSTER,H.P. (1978) Saline lakes and their deposits: a sedimentological approach. In: Modern und Ancient Lake Sediments (Ed. by A. Matter & M. E . Tucker), pp. 7 ~ 4 1Spec. . Publs int. Ass. Sediment. 2. BlackRell Scientific Publications, Oxford. HARDIE,L A ., LOWENSTEIN, T.K. & SPENCER,R.J. (1 985) The problem of distinguishing between primary and secondary features in evaporites. In : Sixth Symposium on Sult, 'Toronto, Canada, May 1983. Northern Ohio Geological Society, Cleveland, Ohio (in press). HOLSER,W.T. (1979) Mineralogy of evaporites. In: Marine Minerals (Ed. by R. G . Burns), pp. 21 I -294. Mineral. Soc. Am. Short Course Notes, Vol. 6, Washington, D.C. C. (1969) Preliminary experiHSU, K . J . & SIEGENTHALER, ments on hydrodynamic movement induced by evaporation and their bearing on the dolomite problem. Sedimentologv, 12, 1 1 -~25. HUNT,C.B., ROBINSON, T. W., BOWLES,W.A. & WASHBURN, A.L.. ( I 966) Hydrologic basin, Death Valley, California. Prof. Pup. U . S .geol. Suro. 4Y4-B, I38 pp. JOHNS,R.K. & LUDBROOK,N.H. (1963) Investigation of Lake Eyre. Rep. Inrest. geol Sum. S . Aust. 24, 104 pp. JONES,C.L. (1972) Permian basin potash deposits, southwestern United States. In: Geology of Saline Deposits (Ed. by G . Richter-Bernburg), pp. 191 -201. UNESCO, Paris. KUHN, R. (1968) Geochemistry of the German potash deposits. Spec. Pup. g e d . Soc. Am. 88, 427-504. LLEWELLYN, P.G. (1968) Dendritic halite psuedomorphs from the Keuper Marl of Leicestershire, England. Sedimentology, 11,293 -297. LOWENSTEIN, T . K . (1982) Primary features in a potash evaporite deposit, the Permian Salado Formation of West Texas and New Mexico: In : Depositionul und Diugenetic Spectru cd'Et:uporites- -u <'ore Workshop (Ed. by C. R. Handford, R. G. Loucks & G. R. Davies), pp. 276-304. Soc. econ. Puleont. Miner. Core Workshop N o . 3, Calgary, Canada, 1982. LOWENSTEIN. T.K. ( I 983) Deposits undulteration of'un uncient potash etvporite : {he Permian Saludo Formution [if' New Mexico untl West Te.uus. Ph.D. dissertation, The Johns Hopkins University, Baltimore, MD. PHILLIPS,F.G. (1956) An Ititrocfucrion to Crystullography. Longmans, London. RETTIG,S.L., JONES,B.F. & RISACHER, F. (1980) Geochemical evolution of brines in the Salar of Uyuni, Bolivia. ('hem. Geol. 30, 57-79. RICHTER-BERNBURG, G . (1980) Aberrant vertical structures in well-bedded halite deposits. I n : F@h Symposium on Sult Vol. I (Ed. by A . H. Coogan & L. Hauber), pp. 159--166. Northern Ohio Geological Society, Cleveland, Ohio. ROEDDER,E. (1984) The fluids in salt. Am. Miner. 69, 413-439. zyxwvutsrq zyxwvutsrq zyxwvuts zyxwvutsrq REFERENCES ARTHURTON, R.S. (1973) Experimentally produced halite compared with Triassic layered halite-rock from Cheshire, England. Sedinientology, 20. 145- 160. BASSETT, A.M., KUPFER,D.H. & BARSTOW, F.C. (1959) Core logs from Bristol, Cadir and Danby Dry Lakes, San Bernardino County, California. Bull. U . S .geol. Surv. 1045D,97 138. BATHURST, R.G.C. ( 1 975) Carbonate Sediments and their Diugeneyis. 2nd ed. Elsevier, Amsterdam. BONYTHON, C.W. (1956)The salt of Lake Eyre: itsoccurrence in Madigan Gulf and its possible origin. Truns. R . Suc. S. Aust. 19, 66 92. CASTENS-SEIDELL, B. ( 1984) The tmu/omy of u moclern marine siliciclustic suhkhu in u rift valley setring: northwest Guij o j Cul(fnrniu tidul .fluis, Buju Califirniu, Me.Lico. Ph.D. dissertation, Johns Hopkins University, Baltimore, MD. CASTENS-SEIDELL, B. & HARDIE,L.A. (1983) Gypsumanhydrite deposition in sabkhas: new observations from the Holocene tidal flats of the N.W. Gulf' of California, Baja California. Ab.srr. geol. Sue. A m . Prog. 15, 540. CASTENS-SEIDELL, B. & HARDIE,L A. (1984) Anatomy of a modern sabkha in a rift valley setting, N . W. Gulf of California, Baja California, Mexico. Ahstr. Am. Ass. Petrol. Grid. 1984 Annual Convention, San Antonio, Texas. C'OOKE,E.G. (1966) The effect of additives on the crystal form of sodium chloride. In: Second Symposium on Sult Vol. 1 (Ed. by J . L. Rau), pp. 259-268. Northern Ohio Geological Society, Cleveland, Ohio. DELLWIG, L.F.(1955) Origin of the S a l ~ n aSalt of Michigan. J . srdim. Petrol. 25, 83 110. DELLWIG,L.F. (1968) Significant features of deposition in the Hutchinson Salt, Kansas, and their interpretation. Spec. Pup. grol. Soc. Am. 88, 41 9%426. EUGSTER,H.P. & HARDIE.L.A. (1978) Saline lakes. In: Lukes: Cherni.stry,Grologyurid Ph.vsics (Ed. by A. Lerman), pp. 237- 203. Springer-Verlag, New York. FRIEDMAN, I . , SMITH,G.I. & HARDCASTLE, K . G . (1976) Studies of Quaternary saline lakes---I]. Isotopic and compositional changes during desiccation of the brines in Owens Lake, California, 1969- 197 I . Ceochim. Cosmochim. Acki. 40,501 5 1 I . GAVISH, E. (1974) Geochemistry and mineralogy o f a recent sabkha along the coast of Sinai, Gulf of Suez. Sedirnentolo ~ , v , ~ I 397-414. , zyxwvutsrq 644 zyxwvutsrqp zyxwvuts zyxwv zyxwvutsrqp zyxwvut T. K . Lowenstein and L. A . Hurdic ROEDDER, E. & BELKIN, H.E. (1979) Application of studies of fluid inclusions in Permian Salado Salt, New Mexico, to problems of' siting the Waste Isolation Pilot Plant. I n : Scientific Bu.usis,JorNucleur Wustr Munuyement, Vol. 1 (Ed. by G. J. McCarthy), pp. 313 -321. Plenum, New York. SHEARMAN, D.J. (1970) Recent halite rock, Baja California, Mexico. Truns. Inst. Min. Metull. B, 19, 155-162. SHEARMAN, D.J. (1978) Halite in sabkha environments. In: Marine Eouporites (Ed. by W. E. Dean & B. (1. Schreiber), pp. 30. 42. Soc. econ. Pulcont. Minerul, Tulsu Short Course No. 4. SMITH,G.I.: BAKCZAK, V.J., MOULTON, G.F. & LIDDICOAT, J.C. (1983) Core KM-3, a surface-to-bedrock record of Late Cenozoic sedimentation in Searles Valley, California. Prof: Pup. U S . yeoi. Surv. 1256,24 pp. SOUTHGATE, P.N. (1Y82) Cambrian skeletal halite crystals and experimental analogues. Sedimentology, 29,39 I -407. STANTON, R.L. & GORMAN, H. (1968) A phenomenological study of grain boundary migration in some common sulfides. Econ. Geol. 63,907-923. zyxwvut zyxwv G. E. (1974) Geology of salars STOERTZ, G.E. & ERICKSEN, in Northern Chile. Prof: Pap. U.S. geol. Surv. 811,65 pp. TUCKER, R.M. (1981) Giant polygons in the Triassic salt of Cheshire, England : a thermal contraction model for their origin. J. sedim. Petrol. 51,119-786. WARDLAW, N.C. & SCHWERDTNER, W.M. (1966) Haliteanhydrite seasonal layers in the middle Devonian Prairie Evaporite Formation, Saskatchewan, Canada. Bull. geol. Soc. Am. 11,331~342. Y U&, KEQIN,C. (1983) Characteristics YUAN,J., C H E N ~ ~H. of salt deposits in the dry salt lake. Absfr. Si-xfhint. Symp. Suit, Northern Ohio Geological Society, Cleveland, Ohio. (Manuscript received 25 Muy I984 ;reoision received 20 March 1985)