New Mexico Bureau of Mines & Mineral Resources Bulletin 117 Geology of Carlsbad Cavern and other caves in the Guadalupe Mountains, New Mexico and Texas

Age of Guadalupe caves

The spar at the Big Room level, Carlsbad Cavern, has been ESR-dated at 879,000 ± 124,000 ybp (Table 24). This age probably exceeds (but not by much) the time when the water table was at the level of the Big Room. The approximate 600,000 ± 200,000 ybp date on the Texas Toothpick stalagmite in Lower Cave (Table 24) and the >730,000 ybp (but probably <0.9 my) date on the silt in Lower Cave (Table 25) put the time of water-table development at the Lower Cave level approximately at 750,000-800,000 ybp. The difference in elevation between the Big Room and Lower Cave is 33 m (Fig. 17), so the rate of water-table lowering in Carlsbad Cavern is estimated to have been approximately 0.03-0.07 cm/yr.

Based on these data, the lower levels of Carlsbad Cavern (Lower Cave and the Big Room) are approximately 750,000-850,000 yrs old, and the upper level (Bat Cave) is approximately 1.2 my old (using the 0.05 cm/yr value as the rate of water-table lowering). Caves in the southwestern, higher parts of the Guadalupe Mountains are older than Carlsbad (Fig. 86); if one assumes a constant rate of uplift and water-table lowering over time, then a cave like Cottonwood, with an entrance elevation of 2,074 m (6,819 ft), is roughly 3 my old. Thus, the caves of the Guadalupe Mountains are late Pliocene-Pleistocene in age, a judgment that agrees with King's (1948) contention that the major rise of the Guadalupe Mountains occurred in the late Pliocene to early Pleistocene.

Cave development in present erosion cycle

Based partly on what he thought was evidence of a past "stream" in Lower Cave, Carlsbad Cavern (i.e. the cobble gravel and silt), Bretz (1949) postulated that there had been two exhumations of the reef escarpment: a pre-Ogallala one, at which time the caves developed, and a present-day one, of canyon downcutting. Motts (1957, 1959) and Bachman (1976) disagreed with Bretz's idea that there had been a post-peneplain Pecos Valley deeper than the present one. Bachman cited drill-core data for thinning of the Ogallala Formation south and southwest of the Llano Estacado, and Motts presented evidence that cave formation is still going on today, even though the potentiometric surface in the Capitan reef aquifer is over 100 m lower than the potentiometric surface in the Gypsum Plain.

This study supports Motts' and Bachman's thesis of only one exhumation of the reef and contends that: (1) the cobble gravel in Lower Cave may be a debris flow rather than a stream deposit, and (2) the silt overlying the cobbles is autochthonous residue derived from the dissolution of bedrock. Thus, an ancestral Pecos Valley deeper than the present one need not be invoked to explain the cobble and silt deposits in Lower Cave.

The <1,000,000 yr dates on the spar, travertine, silt, and calcified siltstone-cave rafts confirm that the caves date from the present erosion cycle (Table 24). Cave rafts always deposit on a water surface; if the rafts of the siltstone-raft sequence formed at the surface of the water table during an earlier erosion cycle, then this sequence should be covered with other subaqueous deposits rather than having air space or subaerial travertine overlying them.

Rate of canyon downcutting

The bat-guano deposits of New Cave are unusual in that they have been water-washed, mixed with silt, and bat-bone fragments in them have been stratified with respect to different types of bone pieces. The most likely explanation for this distribution is that flood water entered New Cave from the stream bed of Slaughter Canyon when the stream bed was at the level of the New Cave entrance, instead of being 174 m below the entrance as it is today. If this is the correct explanation for these unusual deposits, then the washed guano-silt can be related to the rate of downcutting in Slaughter Canyon. R. H. Brown (in a letter to Carlsbad Cavern National Park, 1981), reported a date of >32,500 ybp for the bat guano in New Cave (Table 24). Using this as a minimum date for the guano, the rate of downcutting in Slaughter Canyon can be calculated at <4.5 cm/yr.

Guadalupe caves and Pleistocene climate

Bachman (1974) assigned the most humid climate and greatest erosion in the Delaware Basin to the middle Pleistocene (approximately 600,000 yrs ago), when the Gatuña Formation was being deposited. This assignment correlates with the carbon-oxygen-isotope data on the Texas Toothpick stalagmite in Lower Cave, Carlsbad Cavern, which is approximately 600,000 yrs old in its center (Fig. 75). The carbon-oxygen composition of the Texas Toothpick stalagmite is low for the period >350,000 yrs to the center of the core (600,000 yrs), suggesting that this period may have been dominated by forest types in a wet and humid glacial stage.

Later in the Pleistocene, approximately 140,000-170,000 ybp. there was another wet stage (the penultimate glaciation), during which a large amount of travertine deposited in the Georgia Giant stalagmite, Carlsbad Cavern (Fig. 74). Then, 120,000-130,000 ybp (the last interglacial) there was a cessation of growth which corresponded to a semiarid grass and sedge environment (Fig. 77). Harmon and Curl (1978) found a similar trend of late Pleistocene growth in the travertine of Ogle Cave. Around 200,000 ybp, after the initiation of pluvial glaciation, there was a maximum amount of growth; then, at 125,000 ybp, growth ceased as a change from pluvial to arid climate ensued.

Wet and dry climatic conditions can also be deduced from the amount of travertine deposited during certain intervals. Maximum growth of the Georgia Giant stalagmite on Iceberg Rock, Main Corridor, occurred during a glacial maximum at 140,000-150,000 ybp, and all growth ceased from 120,000-130,000 ybp, during the well-documented warmest phase of the last interglacial (Fig. 77).

Relationship of Guadalupe caves to oil and gas fields of Delaware Basin

According to the findings of this study, the genesis of Guadalupe Mountain caves is related to the oil and gas fields of the Delaware Basin. Where hydrocarbons migrated updip in the basin and mixed with overlying Castile anhydrites, H₂S and CO₂ were produced. These gases moved further updip along the base of the impermeable halite beds in the Castile Formation until they intersected north-south joints, whereupon the gas moved into the reef to form the caves of the Guadalupe Mountains. Alternatively, the gas generated at the base of the Castile Formation moved up into the reef along the Bell Canyon Formation (Fig. 87). Where halite beds remained intact in the basin, H₂S and CO₂ continued to migrate into the reef to form caves; however, where the margin of halite dissolution moved past a cave location in the reef, development of that cave stopped.

The movement of the halite margin past specific cave locations may be the reason why the caves in the Guadalupe Mountains seemingly "die with depth." The lowest cave passage in the Guadalupe Mountains is the Lake of the Clouds, Carlsbad Cavern; the halite margin in the Gypsum Plain has moved approximately 0.5 km past the north-south extension of this passage (Fig. 85). According to Bachman and Johnson (1973), the horizontal rate of dissolution of halite in the Gypsum Plain has been 10-13 km/1,000,000 yrs (this is a minimum rate according to G. Bachman, pers. comm. 1986). Using this rate and the distance of 0.5 km, it can be calculated that the halite margin moved past Carlsbad Cavern roughly about 50,000 yrs ago. The age of the upper part of the cave-raft cones on the Balcony of the Lake of the Clouds Passage, which probably signifies the last episode of significant CO₂ degassing at the water table in Carlsbad Cavern, is 50,000 ybp (Table 24, sample 30). In the same area of the cave (the Christmas Tree Room), sulfur crystals overlie cave rafts and other speleothems. Thus, the last injection of CO₂ (and presumably H₂S) into Carlsbad Cavern roughly correlates in time with the movement of the halite margin past Carlsbad Cavern.

Implications for evolution of intracratonic basins

Mississippi Valley-type ore deposits

The evolution of intracratonic basins, where carbonate rocks (very often reef masses) host ore deposits along basin margins, has been the topic of much discussion among ore geologists. The results of this study are compatible with the thesis that some sulfuric-acid-formed caves (like those in the Guadalupe Mountains surrounding the Delaware Basin) may be a manifestation of the evolution of intracratonic basins, as are hydrocarbon deposits and Mississippi Valley Type (MVT) sulfide-ore deposits.

Two basic depositional models have been proposed for MVT mineralization: (1) a "non-mixing" model, where metals and reduced sulfur (H₂S) move together in brines from basins into carbonate margins; and (2) a "mixing" model, where metals and reduced sulfur move separately (Anderson and Macqueen, 1982). MVT deposits are known to occur worldwide and are usually karst-related. Considerable variation exists among different MVT deposits, especially in the amount of mineralization and carbonate solution (karst development), and also in the relative timing of gas migration, ore mobilization, sulfide precipitation, and carbonate solution.

Carbonate rocks surrounding the Delaware Basin of southeastern New Mexico contain small deposits of MVT sulfides. Mazzullo (1986) reported Pb-Zn MVT mineralization on the southeastern side of the basin, and barite has been found in the southwestern (Apache Mountains) part of the basin (McAnulty, 1980). Iron (as pyrite, in distinct large crystals) and other metals are known to occur in the Guadalupe Mountains, on the northwestern side of the basin. The pyrite occurs in shelf rocks (primarily the Yates Formation). One piece of pyrite, collected from Guadalupe Ridge near Dark Canyon Lookout, exhibited concretionary structure in thin section, but no detailed petrographic analyses have been done to determine if the pyrite is of epigenetic or syngenetic origin. Anomalous concentrations of arsenic, barium, cadmium, copper, lead, molybdenum, silver, and zinc have been found at localities near the center of Guadalupe Ridge and on Lonesome Ridge. These metals are concentrated in an iron-stained sandstone unit (the lower Yates?) near the top of the Seven Rivers Formation; this sandstone represents a zone of high permeability that controlled the migration of weakly mineralized epigenetic fluids (Light et al., 1985).

The caves of the Guadalupe Mountains may relate to the problem of MVT sulfides, especially according to the "mixing" model, where metals and hydrogen sulfide move into host carbonate rocks separately, at different times, or from different sources, Using the results of this study and the Delaware Basin as a model, the following interpretation of MVT ore deposits is suggested.

(1) Hydrogen sulfide ascends from the basin in gas-phase transport (as indicated by late-stage, subaerial, sulfur deposits in Guadalupe caves). This mechanism eliminates the problems involved with compaction-driven-flow models such as discussed by Bethke (1985).

(2) Basinal rocks supply the hydrogen sulfide necessary for MVT mineralization, whereas backreef carbonate and evaporite rocks supply the metal. Chloride-rich brines have the ability to leach trace quantities of metals from rocks through which they flow. The metals move downdip along backreef beds by gravity flow until they reach the host reef rocks.

(3) Flow of gas from basin to reef is through shallow carbonate horizons at the basin edge. Hydrogen sulfide, according to the sulfuric-acid mechanism proposed in this study, is responsible for dissolving out cave voids along the reef margin of the basin. These sulfuric-acid reactions prepare the rock for epigenetic sulfide-mineral emplacement in karstic voids along basin margins, and also provide the low pH's necessary for the concentration of metals such as lead and zinc. Stratiform metal sulfides (and cave levels) form within a narrow range of elevations corresponding to water-table base levels where oxidation is the most pronounced.

(4) Ascending hydrogen sulfide mixes with descending metal-bearing solutions; where H₂S and carbonate rock are encountered, metal sulfides precipitate. The metal sulfides thus deposited typically display hopper, colloform, stalactitic, rhythmically banded, and sometimes liesegang-ring type structures where they fill karstic voids.

(5) Cave dissolution (and ore mineralization) do not necessarily occur in a hydrothermal regime (as suggested by carbon-oxygen signatures of bedrock; this study). Homogenization temperatures (50-200°C) of fluid inclusions in MVT ore minerals may possibly be caused by exothermic reactions or other unknown factors.

(6) According to some models of MVT mineralization, low pH and high total dissolved CO₂ are required for ore-forming fluids. According to the model of speleogenesis proposed in this study, both carbon dioxide and hydrogen sulfide ascend from basin to reef and could fulfill these requirements. Sulfuric-acid solutions could be responsible for periods of dissolution and etching of MVT sulfide crystals.

(7) The release of gas in the basin occurs episodically, so that hydrogen sulfide enters the caves in "pulses" or "gasps" (as indicated by possibly more than one episode of sulfur mineralization in the Christmas Tree Room area, Carlsbad Cavern). This episodic infiltration of gas may account for the zoning so commonly seen in MVT sulfide minerals. During each "pulse" the gas diffuses slowly through the wall rock of karstic voids, resulting in well-formed and large sulfide crystals.

(8) Gravity-driven flow may not be essential for hydrogen sulfide entering from basin into reef. According to Sares (1982) there is a hydrologic barrier between basin and reef in the Guadalupe Mountains, and according to Hiss (1980) the hydrologic gradient in the basin is parallel to the reef front, not perpendicular to it; thus, the desired gravity-driven "plumbing system" expected for fluid movement from basin to reef does not seem to exist in the Guadalupe Mountains area. Similar hydrologic conditions have probably been operative during the last half a million years or so (Bachman, 1984); yet, even possibly as late as 50,000 ybp, gas entered Guadalupe caves along beds dipping up from the basin and into the reef.

(9) If a basin has a limited source of metal but abundant hydrogen-sulfide generation within it (as has apparently been the case of the Delaware Basin), then only small amounts of ore will be deposited along basin margins, but large caves can develop there from H₂S—sulfuric-acid reactions. However, if a basin contains a sufficient source of metal (as has apparently been the case of the Illinois Basin), then the H₂S will react with the metal to form extensive sulfide-ore deposits along basin margins, but karst development will be limited.

(10) One thesis that has been central to MVT ore-mineralization theories is that ore-H₂S-bearing fluids are highly saline, Na-Ca-Cl brines (Anderson and Macqueen, 1982). However, the gypsum deposits in Guadalupe caves—which are the end product of H₂S migration from the Delaware Basin—are notably low in sodium and chloride (Na = 0.05-0.1 wt%, Cl=4.6 ppm; Table 26). Perhaps, since there seems to be a limited supply of metal associated with the development of the Delaware Basin, very little metal-chloride complexing took place. Or, perhaps, these sodium and chloride ions never became concentrated enough to be precipitated and were instead discharged by the Capitan aquifer system (Na = 1-30 ppm, Cl = 11-85 ppm; Table 2).

(11) The pyrite in the Yates Formation may have formed from the reaction of epigenetically introduced hydrogen sulfide (coming up from the basin) with iron (in the siltstone). Where this hydrogen sulfide reacted with oxygen in cave voids, it formed native sulfur.

(12) Anomalous uranium in the cave rafts of the Christmas Tree Room (U = 238 ppm) may also be the result of a MVT-type precipitation mechanism. Uranium is soluble in oxidized form, but where circulating ground water is reduced in an H₂S-rich environment (such as in roll-front deposits), uranium is precipitated. A high concentration of uranium may also be expressed in radon-daughter products and gamma radiation (Table 6); high radon associated with the grayish-green montmorillonite clay in Lower Cave may be especially significant, since uranium is readily taken up or "fixed" by montmorillonite-type clay under low pH conditions (U ≅ 320 ppm; Table 17).

Other cave systems fringing basins

Two other cave systems—Fiume Vento, Italy, and Akhali Atoni, USSR—fringe basins which are known to generate large quantities of hydrogen sulfide. Fiume Vento Cave (in the Apennine Mountains that rim the west side of the Adriatic Sea Basin) resembles Guadalupe caves in that it has large rooms, boneyard under large rooms, tubular pits, gypsum blocks and rinds, montmorillonite-endellite clay (with associated opal rather than chert), and condensation-corrosion features (Hill, in press). Hydrogen sulfide is detectable near the water table, but the ultimate source of this gas has not been determined (P. Forti, pers. comm. 1986). It is possible that H₂S derives from oil and natural-gas deposits in the Adriatic Sea Basin about 50 km to the east.

Akhali Atoni Cave in the Caucasus Mountains about 2 km from the edge of the Black Sea Basin may be another sulfuric-acid cave like those in the Guadalupe Mountains. Akhali Atoni is developed in thick Mesozoic carbonate rocks which dip southward, toward and underneath the Black Sea. It displays huge chambers and galleries (up to 95 m long, 60 m wide, and 90 m high) and has good phreatic form, as is typical of Guadalupe caves. The current thinking on the speleogenesis of Akhali Atoni is that it was opened by rising warm waters, enlarged by warm-water—meteoric-water mixing effects, and then progressively abandoned by a lateral and downward shift of spring points (D. C. Ford, pers. comm. 1986).