Printer Friendly

Utilizing ground penetrating radar to characterize gypsum karst features in Eddy County, New Mexico and Culberson County, Texas.

Abstract-Ground penetrating radar (GPR) has become a popular geophysical tool for locating subsurface karst features such as cavities, conduits and solutionally enlarged fractures. This study examines the usefulness of GPR for identifying buried sinkholes in gypsum and characterizing the sinkhole origins as either solutional or collapse. The GPR data were collected from multiple sites in the Castile Formation outcrop area in Culberson County, Texas and Eddy County, New Mexico. A Pulse EKKO 100 system manufactured by Sensors and Software Inc. with an antenna center frequency 100 MHz in the common offset technique mode was used to conduct GPR profiling. Resulting profiles showed a pattern of radar reflections which reveal a possible series of filled sinkholes. Analyses of the collapse sinkholes show that they are the result of upward stoping subsurface voids, which is characterized by large electrical contrast between the unbroken host rock and the infilled collapse. The sinkholes thought to have originated due to solution show surface disturbance on radargrams. The solutional sinkholes are formed by epigene processes and collapse structures represent failure into an underlying void, which may have formed by hypogene or epigene processes. Hypogene karst processes dominate the area, thus most collapse sinkholes are likely associated with surficial breaching of pre-existing hypogene cave by surface denudation.

Ground Penetrating Radar is a high resolution geophysical technique that uses electromagnetic (EM) waves in frequencies ranging from 50 to 2000 MHz to locate shallow structures and geological features (Davis & Annan 1989). In recent years, ground penetrating radar (GPR) has become a popular geophysical tool for locating subsurface karst features such as caves, cavities, conduits and solutionally enlarged fractures (Pueyo-Anchuela et al. 2008, Brown et al. 2011, Margiotta et al. 2012).

The Castile Formation, a highly soluble, gypsum-dominated unit, crops out over an area of 1800 [km.sup.2] in Eddy County, New Mexico and Culberson County, Texas (Stafford 2006). Currently, there is a limited number of geophysical studies being conducted to better understand void development in the Castile Formation. Stafford et al. (2008b) indicate that more than 10,000 surficial manifestations of karst features occur across the outcrop area; the majority of these karst features form shallow closed depressions indicating karst recharge areas that have been filled with sediment. Because these features are filled with sediments, they cannot be studied directly without excavation; therefore, non-invasive geophysical techniques should provide a means of characterizing the extent and origin of these features.

In addition to improving the understanding of karst processes, this study provides a thorough analysis of the usefulness of GPR in locating, analysing and characterizing voids in the Castile Formation. The results also provide additional information on the speleogenetic evolution of the region as well as potential fluid transport.

Speleogenesis.--The Castile outcrop area of the western Delaware Basin in Eddy Co., NM and Culberson Co., TX is dominated by surficial karren, sinkholes and associated caves. Cave and karst development within the Castile Formation is widespread due to the fact that evaporite rocks are highly soluble (~2.53 g [L.sup.-1]) and near-linear solution kinetics of these rocks encourages the rapid development of a karst landscape (Stafford et al. 2008b). Speleogenesis in continental settings occurs in two diagenetic settings, which is based on the characteristics of the soluble fluids: 1) epigene and 2) hypogene (e.g. White 1988; Ford & Williams 2007). Due to the fact that geologic systems constantly evolve through time, multiple episodes of different types of speleogenesis may occur within one system, overprinting earlier phases with latter phases of karst development (e.g. Klimchouk et al. 2000).

In the Delaware Basin, hypogene processes appear to dominate the speleogenetic evolution of the area. Numerous examples of the dominance of hypogene speleogenesis occur throughout the greater Delaware Basin region, though the specific composition of the fluids and host rocks differ; however, epigene karst is also common in the Castile Formation, and other evaporite strata in the region, because of the high solubility of gypsum (Stafford et al. 2008a; 2008b).

Epigene speleogenesis occurs in unconfined settings and it is directly associated with meteoric precipitation (Palmer 1991). Gravity drives dissolutional patterns in the unsaturated, vadose zone, while hydraulic potential dominates in the saturated, phreatic zone. Epigene secondary deposits may form either subaerially in the vadose zone, commonly due to variations in void microclimate, or subaqueously in the phreatic zone, generally associated with changes in water chemistry (Ford & Williams 2007).

Epigenic dissolution, involves surficial features that act as interface points for descending waters that may recharge local ground water or form integrated cave networks that function as subsurface bypass features for overland flow (White 1988). Epigenic karst forms surface features that are easily recognized near the land surface, while hypogenic speleogenesis is not as easily recognised because it forms without a direct surface connection, usually in confined or semi-confined settings. Hypogenic karst is often only exposed by surface denudation and incidental breaching of voids (Palmer 1991; Klimchouk et al. 2000).

Hypogene speleogenesis develops where dissolution is aided by a mixed convection hydrologic flow system, including significant components of free and forced convection, where soluble fluids are delivered from underlying or adjacent zones. Hypogene systems do not form direct connections with surface process, but they are associated with regional and basin-scale fluid patterns (Klimchouk et al. 2000).

Hypogene speleogenesis is often associated with hydrothermal or sulfuric acid systems, but includes most geologic systems where fluids originating from lower depths or distal margins migrate vertically and laterally through overlying or adjacent soluble rocks. This process can form complex solution features but also extensive, economic secondary mineral deposits (Ford & Williams 2007).

Sinkholes in the study area occur as filled and open forms, where open sinkholes are connected directly to solutional conduits. These sinkholes are formed by two basic mechanisms; solutional incision of descending waters (epigene) or collapse of upward stopping subsurface voids (hypogene or epigene). Morphology of sinkholes is used to determine if features are solutional voids resulting from epigene processes or if these voids are the result of collapse into a hypogene or epigene void. When formed by solutional processes, sinkholes are expected to exhibit a more elongate shape as opposed to sinkholes formed by collapse processes, which have a more elliptical, near circular shape (Stafford et al. 2008b). They are both evidenced on the surface by changes in topography and vegetation infilling sinks.

The main objectives of this research were: 1) test the usefulness of GPR for characterizing sediment-filled sinkholes in the Castile Formation; 2) interpretation of subsurface hydrologically active karst features; 3) characterization of sinkhole origins as either solutional or collapse features; and 4) evaluation of the sedimentary fills in karst features.


Study Area.--The study area is located in the semi-arid, northern edge of the Chihuahuan Desert, which lies within the Delaware Basin of south-eastern New Mexico and West Texas (Fig. 1). The annual precipitation averages 26.7 cm, with the greatest rainfall occurring as monsoonal storms in late summer from July to September. Annual temperature averages 17.3 degrees Celsius, with an average annual minimum of 9.2 degrees Celsius and a maximum of 25.2 degrees Celsius (Stafford et al. 2008b).

The gypsum facies of the Castile Formation crop out over an area of 1,800 [km.sup.2] in Eddy Coounty, New Mexico and Culberson County, Texas on the western edge of the Delaware Basin (Fig. 1) and dip into the subsurface to the east where it is mantled by later Permian and later strata (Stafford et al. 2008b). It was deposited as restricted, basin-filling evaporites that include interbedded sulfates, carbonates and halides. The Castile Formation is conformably underlain by siliciclastics strata of the Delaware Mountain Group, which includes the Cherry and Bell Canyon Formations. It is conformably overlain by the Salado and Rustler Formations (Hill 1996).


Data collection.--Ground Penetrating Radar (GPR) data were collected at four sites, chosen based on accessibility and surficial morphology of karstic features. GPR technique was chosen for this study due to its ability to non-invasively image shallow subsurface features. One site (Site 1) is located in southern Eddy Co., NM and three other sites (Sites 2, 3 and 4) are located in central Culbertson Co., TX (Fig. 1).

GPR profiling was carried out using a Pulse EKKO 100 system manufactured by Sensors and Software Inc. (2003). Data acquisition parameters were: antenna center frequency 100 MHz; step mode; 64 samples per scan; 400 ns time window; antenna separation 1 m; and 0.25 m trace spacing. The common offset technique was used where the pair of transmitter and receiver antennae separated by 1 m was advanced simultaneously along the profile.

A GPR survey line was established across each of the four sinkholes with estimated depth of penetration approximately 12 m. An additional line was shot across a known cave at a depth of less than 5 m as a control for data analyses (Fig. 1). At the study sites, the maximum depth of recognizable radar reflections was approximately 12 m. As surficial expression of sinkhole features is shallow, this was considered sufficient depth to image the characteristics of these sinkholes.

GPR data were processed using EKKO_View Deluxe software, version 2 (Sensors & Software 2003) which allows for data plotting, editing and processing. The processing flow consisted of: filtering (Dewow) to remove unwanted low frequency signals; automatic gain control (AGC) application to amplify the strength of deeper signals (window width 1.5 and maximum gain 500); and time to depth conversion. The velocity used to convert time to depth was 0.104 m/ns obtained from hyperbola velocity calibration tool built into the software.


Survey line 1 (produced from survey data collected at site 1 in southern Eddy County, NM) was oriented in an east to west direction and shows two distinct voids labelled void 1 and void 2 (Fig. 2A). The observed difference in radar signature is due to differences in electrical properties between void spaces and the host rock. Voids created in the host rock may be filled with unconsolidated sediments of the host rock or brought in from elsewhere. In addition to difference in electrical property, the porosity and permeability of the unconsolidated zone is also different from the surrounding host rock and is revealed by the sharp electrical contrast of the waves on the radargram (Fig. 2A).

Both voids occur at depths of approximately 2 m with void 1 displaying a stoping upward shape. The morphological interpretation is shown in Figure 2B. The larger of the two voids observed in line 1 (Void 1) appears to have a surface depression interpreted to be a filled sinkhole. This depression is occupied by a cluster of vegetation which is likely extracting nutrients and moisture from the material infilling the subsidence feature. Void 2, located towards the western end of the radargram (Fig. 2A), shows reflections from a possible conduit, or a possible future sinkhole that is stoping upwards. The radar velocity of the sediments filling voids is approximately 0.092 m/ns which is consistent with the radar velocity of brecciated evaporites.

The features in line 1 (Fig. 2A) both display a stoping upward shape. These structures probably represent collapse into an underlying void, which could be either hypogenic or epigenic in origin. Collapse sinks generally form near-circular or elliptical features with steep walls but the sink morphology may not show this distinct shape with GPR techniques, as it may be obscured by sediment infilling.

A direct connection to the surface depression was not observed on the radargram. Figure 2A likely shows a feature that collapsed upward and surficial beds sagged due to loss of support from below, resulting in minor deflection of surface layers. The upward stoping shape could explain the limited surface disruption, as the upward stoping shape resulted in "sagging" near the surface and the shallow surface depression. As upward collapse occurs, breakdown will fill up the original void leaving air- or sediment-filled space between the broken blocks. Therefore, the collapse structure filled the void such that only minimal collapse, possibly only sagging, occurred near the surface, while greater disruption occured deeper in the strata. GPR technique was incapable of imaging the internal structure of the collapse feature, possibly because broken down blocks could be oriented at different angles and spaces between them may be partially filled with sediment.


Survey line 2 was oriented in a south to north direction (Fig. 3A). It shows two shallow hyperbolas at depths less than 2 m which is shallow enough to be either a surface or subsurface feature. These features were interpreted to be possible solutional holes and are located between 14-21 m and 28 - 34 m (Fig. 3A) on survey line 2. These diffractions probably originated from shallow epigene enlargement of fractures or joints by descending vadose waters (Fig. 3B).

There was a subtle electrical contrast between the observed solutional features and the surrounding rock. The radar velocity of the solutional features was 0.092 m/ns, indicating that they were infilled with sediment, very similar in nature and consistency to gypsum, which forms the host rock.

Solutional holes observed at the surface (Fig. 3B) were interpreted to represent features that were subsequently filled with sediment. They were likely formed by the dissolution of descending waters that dissolved pathways from the surface down, forming sinkholes. The fill possibly originated from erosion of surface soils that were subsequently washed into the solutional opening, filling the voids. There are no visible dipping beds or blocks that could deflect radar waves, suggesting that the voids are not collapse in origin, but are solutional sinkholes.

Survey line 3 was oriented in a west to east direction. The radiogram (Fig. 4A) displays a strong electrical contrast between the host rock and the void infill. The velocity of the void infill is 0.090m/ns which is consistent with unconsolidated gypsum. Voids occur at a depth of approximately 1 m depth and display stoping upward shapes. Similar to survey line 1, the stoping upward shape observed in line 3 (Fig. 4B), likely represents a sinkhole formed by collapse of a pre-existing void. The stoping upward trend seen in the solutional feature is indicative of a collapsed sinkhole, as the radargram shows no direct connection to the surface. The surficial sagging of the beds was probably caused by the loss of support from below.

Survey lines 4, oriented in a south to north direction (Fig. 5A), revealed two distinct solution features (Fig. 5B), which occur at 4 to 9 meters and 21 to 27 m on the radargram, both at shallow depths 2 m and less. These solutional holes show slight disturbances in the subsurface and a small electrical contrast in the subsurface with surrounding rocks. The velocity of 0.092 m/ns indicates that these solutional holes are infilled with material consistent with the host rock, which is gypsum. These solutional holes also appear to have resulted from dissolution by descending waters that dissolved out pathways from the surface down forming a solutional sinkhole that was subsequently filled with sediments from the surface (most likely gypsic soil). This is, evidenced on the surface by a depression and vegetation infilling.

Figure 6A represents a survey line over a known cave which was used as a control for this study. This cave was observed at the surface and can be seen on the radargram between 2 meters and 3.5 m, as a hyperbolic shaped subsurface feature. The cave occurs between 1 meter and 3.5 m depth. This cave contains void spaces partially filled by sediments that have been washed in, as evidenced by the observed surface conduit and sediments lining the conduit floor. The feature appears to be solutional in origin based on field observations; however, distinct reflections representing the cave were observed, suggesting that this cave may be a small part of a larger system (Fig. 6A & 6B).

Ground Penetrating Radar works by injecting radio signal or series of signals into the subsurface. The reflected radar signals provide spatial information regarding anomalous subsurface features. Also, GPR works best when there are well-defined differences in the electromagnetic properties of the subsurface materials being surveyed, gradual changes are more difficult to detect, such as those between gypsum bedrock, gypsum collapse blocks and gypsic soils. This was observed when viewing radargrams where solutional features such as sinkholes were filled with sediment similar in nature and consistency to the host rock.

The main objectives of this research were achieved using GPR to image the subsurface and characterize the filled sinkholes in the Castile Formation. Also characterizing sinkhole origins as either solutional or collapse features was done using GPR images. Calculating the velocity and analysing the different electrical properties of the features seen, made it possible to evaluate the sedimentary fills in the karst features.



Cave and karst development within the Castile Formation is widespread and diverse, because karst development ranges from surficial (solutional) to subsurface features (collapse sinkholes) of both epigene and hypogene speleogenetic origins. The collapse sinkholes seen in the study area most likely represent the surficial expression of hypogene caves that have stoped to the surface, based on the abundance of hypogene features previously reported from the study area (e.g. Stafford et al. 2008b). They were likley formed by the collapse of a void and upward stoping of materials towards the surface in conjuction with high rates of surface denudation. Thus most collapse sinks are likely the result of loss of ceiling support and collapse into pre-existing hypogene caves.

The solutional sinkholes resulted from dissolution of descending waters that dissolved out pathways as surficial interface features, forming a sinkhole that was subsequently filled with surface-derived sediments. Solutional sinks are the result of epigenic processes and are likely associated with laterally limited cave networks.

In conclusion, the surficial landscape of the Castile Formation is dominated by epigenic karst (due to the highly soluble gypsum which is exposed on the surface), while cave development and subsurface voids are formed mainly due to hypogene processes. The dominant process, which is carving the landscape, in the study area is the degradation and overprinting of original features formed through hypogene speleogenesis, thus all collapse features into hypogene voids are currently being overprinted by epigene processes further complicating the characterization of karst development within the Castile outcrop area. Also extensive field research conducted by Stafford et al. (2008b) has provided additional evidence for this hypothesis. His research concluded that most caves exhibit hypogene origins in the Castile Formation and more than 55% of sinkholes in the area are collapse features, which is further supported by number of collapsed features detected in the Castile outcrop. Additionally, Stafford et al. (2008b) documented the widespread occurrence of evaporite calcitization, sulfur ore mineralization, and secondary gypsum (selenitic textures) crystallization that indicate that karst development in the region is dominated by hypogene processes where ascending, transverse speleogenesis is associated with both water and light hydrocarbons that have created complex secondary diagenetic alterations within the Castile Formation.

GPR analyses can provide significant insight into the morphology and speleogenetic origin of sinkholes and buried karst features in gypsum terrains. These results can be used to non-invasively evaluate karst features as potential geohazards within the region and within other evaporite facies, where upward stoping collapse structures could cause significant geohazards threats.


Brown, W. A., K. Stafford, M. Shaw-Faulkner & A. Grubb. 2011. A Comparative Integrated Geophysical Study of the Horse Shoe Chimney Cave, Colorado Bend State Park, Texas, International Journal of Speleology, 40(1): 9-16.

Davis J.L. & A. P. Annan. 1989. Ground penetrating radar for high resolution mapping of oil and rock stratigraphy. Geophys. Prospect., 37: 531-551.

Ford, D. & P. Williams. 2007. Karst Hydrogeology and Geomorphology. John Wiley & Sons, 562 pp.

Hill, C. A. 1996. Origin of barite and sulphur deposits in the Delaware Basin discussion: West Texas Geol.Soc, Bull. 35(6): 5-7.

Klimchouk, A., D. C. Ford, A. N. Palmer & W. Dreybrodt. eds. 2000. Speleogenesis of Karst Aquifers. National Speleological Society, Inc., Huntsville AL, 527, pp.

Margiotta, S., S. Negri, M. Parise & R. Valloni. 2012. Mapping the susceptibility to sinkholes in coastal areas, based on stratigraphy, geomorphology and geophysics: Nat. Hazards, 62: 657-676.

Palmer, A.N., 1991. Origin and morphology of limestone caves: Geological Society of America Bull., 103: 1-21.

Pueyo Anchela O., A. M. Casas-Sainz, M. A. Siriano, & A. Pocovi-Juan. 2008. Mapping subsurface karst features with GPR: results and limitations: Environ Geol. 58: 391-399.

Sensors & Software Inc. 2003. Ground Penetrating Radar technology, Ekko_View Enhanced user's guide: (July 2009).

Stafford, K. W. 2006, Gypsum karst of the Chosa Draw area. Pp 82-83 in L. Land, V.W. Lueth, W. Raatz, P. Boston, and D. Love, eds., Caves and Karst of Southeastern New Mexico. New Mexico Geological Society, Socorro, NM.

Stafford, K.W., F. Behnken, & J. White, 2008a, Hypogene speleogenesis within the Central Basin Platform: Karst Porosity in the Yates Field, Pecos County, Texas, U.S.A., in Sasowsky, I., C. Feazel, J. Mylroie, A. Palmer & M. Palmer. (eds), Karst from Recent to Reservoirs, Special Publication 14. Karst Waters Institute, Inc., Leesburg, VA. 174-178. pp

Stafford, K. W., R. Nance, L. Rosales-Lagarde & P. J. Boston. 2008b. Epigene and hypogene karst manifestations of the Castile Formation: Eddy County, New Mexico and Culberson County, Texas, USA. International Journal of Speleology. 37(2): 83-98.

White, W.B., 1988, Geomorphology and Hydrology of Karst Terrains. Oxford University Press, New York, NY, 464 pp.

WAB at:

Wesley A. Brown (1), Trina K. Melville (2) and Kevin W. Stafford (1)

(1) Department of Geology, Stephen F. Austin State University Nacogdoches, TX 75962

(2) Shell Oil, 701 Poydras St., New Orleans, LA 70139
COPYRIGHT 2017 Texas Academy of Science
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Brown, Wesley A.; Melville, Trina K.; Stafford, Kevin W.
Publication:The Texas Journal of Science
Article Type:Report
Geographic Code:1U8NM
Date:Jun 1, 2017
Previous Article:Transgenic cell lines inducing the RNA interference pathway are resistant to EIAV infection.
Next Article:Lesser yellow-headed vulture mandibular ecomorphology and feeding interactions at an established feeding site in Belize.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters