THE USE OF COMPUTER WORKSTATIONS IN THE STUDY OF ENVIRONMENTAL GEOLOGY: INTEGRATION OF GEOPHYSICAL AND GEOLOGIC DATA A. Martinez1, T. Carr1, R. Black2, H.R. Feldman1, J.F. Hopkins1, A.J. Feltz1, D.R. Collins1, J. Doveton1 and N.L. Anderson3, 1Kansas Geological Survey, Lawrence, KS 66044; 2Department of Geology, University of Kansas, Lawrence, KS; and 3Department of Geology & Geophysics, University of Missouri-Rolla, Rolla, MO

ABSTRACT Geophysical techniques coupled with the modern computer workstation have not been widely used in environmental geology. The Kansas Geological Survey is investigating the application of the computer workstation in a number of projects underway that involve environmental questions and problems associated with salt dissolution. The computer workstation provides an efficient approach to integrate geologic and geophysical interpretations of the shallow subsurface. In addition to modification of available techniques originally developed for computer-aided exploration and development of hydrocarbons, several new applications unique to shallow subsurface characterization are being developed. All applications are capable of addressing environmentally-related questions at various scales from regional aquifer mapping to site specific characterization.

Various geophysical and geologic methods have been used to image the shallow subsurface (0-100 m). These methods used in conjunction with a modified computerized interpretation system include high-resolution seismic reflection (2-D and 3-D), vertical seismic profiling (VSP), ground-penetrating radar (GPR), and color image transformation and treatment of the transformed wireline log data as "seismic" traces (pseudo-seismic). The computer workstation approach allows efficient, detailed and integrated studies to be performed at these sites.

Examples from throughout Kansas involving a variety of environmental questions and problems associated with salt dissolution are used to illustrate the utility of using the computer workstation in the study of environmental geology.

KEY WORDS computer workstations, environmental geology, Permian salts, geophysical data

INTRODUCTION digital databases, have radically changed the petroleum industry's approach to expl o- Computer workstation-based interpret a- ration and development. Emphasis within tional software has revolutionized how the petroleum industry is now on powerful geophysical, geologic and engineering data interpretational software that minimizes r e- are interpreted by the petroleum industry. petitive tasks, integrates technologic disc i- Major computer technologies, including plines, maximizes analysis and simulation, near super-computing capabilities on inte r- and increases the number of mistakes (dry active desktop workstations, complex ne t- holes) that are drilled on the desk instead working, high capacity mass storage d e- of in the field. These powerful computer- vices, 3-D visualization software, and large aided interpretational systems have yet to

Proceedings of the 10th Annual Conference on Hazardous Waste Research 341 entire aquifers to localized dissolution co l- lapse features. Causation is a complex i n- teraction of natural processes and anthr o- pogenic activities (e.g., irrigation and oil field activities). Examples from Kansas illu s- trate the utility of the computer workstation to perform integrated studies of enviro n- mental problems at scales ranging from small-scale site investigation to regional aquifer studies (Figure 1). Data were co l- lected as part of ongoing studies of ground water aquifers in Kansas by the Kansas Figure 1. Location of the Siefkes Study Site and the Geological Survey. selected pseudo-seismic line from the Dakota Study Area in Kansas. Various geophysical methods were used at several site investigations to image the g e- be widely used by environmental profe s- ology of the shallow subsurface (0-100 m). sionals. These methods include high-resolution seismic reflection (2-D and 3-D), vertical Adequate understanding of heterogeneities seismic profiling (VSP), and ground- of near subsurface geology is an important penetrating radar (GPR). These geophys i- component to improved site investigations, cal methods along with subsurface geologic avoidance of excess spending and ineffe c- data are used in conjunction with a co m- tive remediation. Environmental investig a- puterized interpretation system. The sy s- tions have focused on the point data o b- tem allows efficient, detailed and integrated tained from the drill rig. Point data are often studies to be performed at these sites. inadequate for understanding preferential flow pathways and variations in relative In order to undertake a regional aquifer permeability. Near-surface computer-aided study, a new application of computer geophysical and geologic methods are a n- workstation-based interpretational software other way to image the subsurface. Appl i- was developed. This application treats cation of computer-aided visualization and transformed wireline log data as "seismic" analysis systems in conjunction with the traces for the purposes of processing, i n- use of near-surface geophysical and ge o- terpretation and display. A color image logic methods can provide detailed info r- transformation can combine data from s e- mation of subsurface structures that affect lected wireline logging tools to generate a fluid flow direction and rate of fluid mov e- color coded "crossplot log" for each well. A ment and enhance containment and rem e- well-designed transformation can provide diation procedures. an image of the spatial distribution of su b- surface lithology or fluids. The transformed This paper presents projects underway in image, in either 2-D or 3-D, can be treated Kansas that involve the application of ge o- on the workstation as "seismic" data, ea s- physical and geological methods and the ing the data handling burdens through use computer workstation to a variety of env i- of computerized techniques designed for ronmental questions and problems assoc i- interpretation of seismic data. ated with salt dissolution. Permian salts present in the shallow subsurface of the SMALL- TO MEDIUM-SCALE central and southern parts of Kansas result in a number of environmental problems APPLICATIONS: THE SIEFKES ranging from large-scale contamination of STUDY SITE

342 Proceedings of the 10th Annual Conference on Hazardous Waste Research Figure 2. Detail map of the geophysical study at the Siefkes Study Site. The locations of the ground-penetrating radar (GPR) profile, 2-D seismic reflection profiles, 3-D seismic reflection patch and vertical seismic profiles (VSP) are shown.

The Siefkes Study Site is located in central severely limited site characterization and Kansas (Section 27, T21S, R12W, Stafford subsequent ground water mode ling. County). The site is contaminated with a region of salt water intrusion into the fresh In order to improve site characterization, a water aquifer. Salt water contamination of series of geophysical investigations were subsurface fresh water supplies is a serious performed (Figure 2). The geophysical data concern within this area of central Kansas collected include a ground-penetrating r a- [1]. The main source of fresh water in the dar profile imaging the very-near surface region is a near-surface aquifer contained above the water table, 2-D and 3-D near- within unconsolidated Quaternary alluvium. surface seismic reflection data imaging the Salt-rich waters are upwelling from the Permian bedrock surface and deeper rock Permian bedrock into this overlying all u- units, and near-surface vertical seismic vium. At the Siefkes Study Site there is a profile data used to depth-tie the seismic noticeable upwelling of salt-rich waters data with lithological information. around an irrigation well late in the pumping season. Salt water contamination reaches Ground-penetrating radar (GPR) a high enough level that crop damage can A short line (76 m) of GPR data was a c- result. quired in the vicinity of two wells at the Siefkes Study Site. Data was acquired u s- The Siefkes Study Site is part of the Mi n- ing a GSSI SIR System 8 GPR unit with a eral Intrusion Study Area, a region of i n- 500-MHz transducer. The data were tensive study by the Geohydrology Section downloaded from the unit and converted of the Kansas Geological Survey [2, 3]. Site into SEGY format for data integration and characterization and ground water mode l- interpretation within the workstation env i- ing of the Siefkes Study Site were unde r- ronment. The vertical (time) and horizontal taken in order to improve understanding of (distance) scales for GPR data differ co n- ground water flow within the area. Two siderably from seismic reflection data. B e- monitoring wells were placed within the cause the GPR data length is in nanose c- proximity of the irrigation well to monitor onds, rather than milliseconds, it was ne c- subsurface salinity levels and provide essary to time-scale the sample interval of lithological information. Ground water the data by a factor of 1x106 for viewing modeling of the area used information pr o- purposes. The horizontal spacing (CDP) for vided by these ground-water monitoring this data set is approximately 1.2 cm per wells. However, the limited control points CDP point.

Proceedings of the 10th Annual Conference on Hazardous Waste Research 343 Figure 3. Detailed image of ground-penetrating radar profile across a feature interpreted as a paleo-stream cha n- nel. The cut bank, as well as several lateral accretionary bars, are hig hlighted.

Once time-scaled, the GPR data were subsurface and should be incorporated to loaded into the computer workstation and develop accurate models of aquifer r e- viewed. The GPR data were treated as charge. stacked seismic data, opening up the po s- sibility of using conventional post-stack 2-D and 3-D near-surface seismic seismic reflection tools to process the data. reflection and vertical seismic Trace balancing and gains were applied in profiling order to give the profile an even appea r- ance and to accentuate anomalies. Other Near-surface 2-D seismic reflection profiles, digital processing techniques included a 3-D seismic reflection patch, and two ve r- bandpass filtering (to remove high- tical seismic profiles were collected at the frequency multiple events caused by the Siefkes Study Site. Seismic Line One i m- water table) and f-k filtering (to reduce hor i- ages the subsurface between the two zontal linear events and accentuate ground-water monitoring wells (wells P and anomalies). DA on Figure 2). Seismic Line Two is pe r- pendicular to Line One and is located near Several features of interest were found well P. The single-fold 3-D seismic patch is along the GPR profile. A portion of the pr o- centered on the two seismic lines in order file over a subsurface feature is interpreted to maximize velocity information for normal as a portion of a paleo-stream channel, move-out (NMO) corrections. The two vert i- with the interpreted point bar and cut bank cal seismic profiles (VSP) were collected at highlighted (Figure 3). Also seen is a di f- wells P and DA. The VSP data provided fraction, believed to be caused by the r e- basic velocity functions for the seismic data mains of a shot point station from the interpretation process and, more impo r- seismic data collected in the area. The tantly, time-depth information to depth-tie GPR data indicates that the very-near su r- the seismic reflection data to known su b- face is relatively laterally and vertically he t- surface stratigraphic units. The vertical erogeneous in nature (Figure 3). This he t- seismic profile trace from well P in relation erogeneity can greatly affect the infiltration to gamma-ray and conductivity logs is of fresh water from the surface into the shown in Figure 4.

344 Proceedings of the 10th Annual Conference on Hazardous Waste Research Figure 4. Vertical seismic profile trace from well P shown in relation to gamma-ray, conductivity and lithological information.

All of the 2-D and 3-D seismic data were computer workstation for interpretation after collected using a downhole 30.06 rifle as a stacking. source. The end of the rifle barrel was placed approximately 0.6 m beneath the The VSP data were collected using an a c- ground surface and a bullet was fired into celerated weight drop source. The receiver the ground. Two common shot records consisted of a three-component downhole were gathered and stacked at each shot geophone. The data were recorded using a location before digital archiving. Receivers Bison 9024 seismograph with a 0.25 msec consisted of three 40-Hz vertical ge o- sample rate. Five common shot records phones placed in an array, with receiver were gathered and stacked at each r e- take-out spacings of 1.2 m. The data were ceiver depth location before digital archi v- recorded using a Bison 9048 seismograph ing. Data processing for the VSP data co n- with a 0.5 msec sample rate. Data proces s- sisted of trace sorting, trace editing, wave- ing for the seismic reflection data was pe r- field separation, coherency filtering, deco n- formed with the SierraSEIS processing volution, static correcting and stacking. The package and consisted of standard trace resulting traces were used to depth-tie the editing, first arrival muting, air-wave and seismic reflection data at the two ground ground roll muting, bandpass filtering, f-k water monitoring wells. Reflections from the filtering, velocity analysis, NMO-corrections base of the aquifer (51 m), and tops of the and stacking. The seismic reflection data Salt Plain (69 m), Harper (104 m), and were trace-balanced and loaded into the Stone Corral (195 m) formations are seen

Proceedings of the 10th Annual Conference on Hazardous Waste Research 345 Figure 5. Greyscale display of Seismic Line One. Interpreted horizons include the base of the aquifer, the top of the Salt Plain Formation, the top of the Harper Formation, and the top of the Stone Corral Formation. on Seismic Line One (Figure 5). All depths size of 0.6 m x 1.2 m. Although the data refer to the depths of reflectors at well P as are only single fold and encompass a rel a- determined by the VSP. tively small area, useful information was obtained once the data were loaded into Seismic reflection imaging of the subsu r- the computer workstation and interpreted. face reveals previously unknown subsu r- Interpretation of the 3-D seismic reflection face structure at the Siefkes Study Site. data was greatly enhanced by the data i m- The area between wells P and DA appears aging capabilities of the workstation. Vert i- to be a Permian structural low in the su b- cal and horizontal data slices were taken of surface. In addition, the reflection from the the data (Figures 6 & 7). Horizon tops were contact between the base of the Salt Plain selected using these data slices. The resul t- Formation and the top of the Harper San d- ing horizon picks were displayed as a 3-D stone Formation disappears on the eastern volume for visualization purposes (Figure portion of Seismic Line One. This is inte r- 8). The reflections from the base of the preted as destructive interference between aquifer, the top of the Salt Plain Formation, the second peak of the Salt Plain reflection and the Stone Corral Formation are visible doublet and the first peak of the Harper on vertical slices. The reflection from the Sandstone doublet as a result of top of the Harper Sandstone is very faint stratigraphic thinning of the interval b e- and is visible on only a few vertical slices. tween the two units. The well data was i n- The time-slice through the Salt Plain Fo r- sufficient to image the structural and mation reflection shows that the reflection stratigraphic variations at the Siefkes site. has a slight dip to the south-east, sugges t- Such subsurface heterogeneities may be ing that the Salt Plain Formation may dip hydrologically significant in influencing the slightly to the south-east in the Siefkes movement and concentration of Permian area. The experimental 3-D patch could be salt-rich brines at the site. significantly improved with higher fold. The higher fold would provide greater velocity The single-fold near-surface 3-D seismic control and more reflection continuity for reflection data were collected in an attempt the 3-D data. to determine the feasibility of small-scale 3- D seismic methods in near-surface studies. The 3-D seismic reflection data have a bin

346 Proceedings of the 10th Annual Conference on Hazardous Waste Research Figure 6. Vertical data slice from the northern face of the 3-D seismic reflection data volume. Figure 7. Horizontal data slice taken at 95 msec from the 3D seismic reflection data volume. The r e- flection from the Salt Plains Formation is intersected A LARGE-SCALE APPLICA- by this data slice. TION: THE DAKOTA STUDY AND THE PSEUDO- Pseudo-seismic profiling SEISMIC APPROACH Stratigraphic interpretation from wireline As near-surface ground water sources b e- logs is typically drawn from multiple log come more scarce, the Cretaceous Dakota traces or from crossplots of log data. Both aquifer is becoming an important potential techniques can readily depict vertical source of ground water for western Kansas. changes in lithology or reservoir quality, but A study of the Dakota aquifer is underway lateral relationships are not easily visua l- in order to develop a better assessment of ized. Significant improvement in the ge o- water-resource potential and planning logic interpretation of wireline log data can needs of the aquifer [4]. The Dakota aquifer be achieved through color image transfo r- is composed of irregularly distributed mation and treatment of the transformed sandstones of the Dakota and Kiowa fo r- data as "seismic" traces for the purposed of mations and Cheyenne Sandstone. An u n- processing, interpretation and display [6]. derstanding of the aquifer's lateral extent, Such transforms can combine data from vertical thickness and overall capacity porosity, gamma and density tools genera t- across western Kansas is important for d e- ing a color coded "crossplot log" for each veloping well-spacing requirements and well. A well-designed color transformation sound resource management. The Dakota of wireline log data from multiple wells aquifer study mainly involves using existing maximizes both spatial and compositional well log information to achieve these goals. information content and provides a readily Because the study encompasses a large interpretable image of the subsurface geo l- area (approx. 125,000 km 2) and a large ogy. The transformed image, in either 2-D number of wells, alternative methods of or 3-D, can be treated as "seismic" data, macro- scale well log analysis were exa m- easing the data-handling burdens through ined and a method designated as pseudo- the use of computerized techniques d e- seismic profiling was developed [5]. signed for interpretation of seismic data.

The pseudo-seismic approach to stratigraphic interpretation is based on the observation that wireline well logs resemble

Proceedings of the 10th Annual Conference on Hazardous Waste Research 347 increases the power of the visualization as compared to traditional computer-assisted wireline log cross-sections. Features such as truncation of individual beds within the Permian by the Cretaceous unconformity, the effect of the dissolution of Permian salt on localizing deposition of overlying Che y- enne-Dakota fluvial sands, and normal faulting on the western margin of the Ce n- tral Kansas Uplift are readily apparent (Figure 9). The zoom and datum capabil i- Figure 8. Perspective view of the selected reflector ties of the workstation can be used to map horizons from the 3-D seismic data cube. The truncation, onlap and downlap of individual ground roll was left in the display to illustrate the difficulties it causes for near-surface 3-D seismic "beds" within specific stratigraphic units. In data collection methods. terms of vertical resolution, pseudo-seismic data can surpass seismic data, as vertical resolution is limited only by the resolution of seismic traces in many respects. Both the logging tool and the digital data. seismic data and wireline log data are si m- ple x-y series, one in the amplitude-time domain and the other in the ampl itude- CONCLUSIONS depth domain. Wireline well logging tools In order to improve site investigations, env i- record various rock properties and output ronmental professionals need to consider these data as a depth series. Ultimately, application of near-surface geophysical and the goal of seismic-style processing is to geologic methods coupled with the visual i- approach on a trace-by-trace basis the zation and analysis power of computer resolution of geophysical well logs. workstation-based interpretational systems. Improved understanding of the subsurface A pseudo-seismic profile provides a simple geology is an important consideration for representation of Permian through Cret a- development of adequate sampling strat e- ceous rocks of western Kansas based gies and the design of workable remedi a- solely on gamma-ray logs (Figure 9). More tion systems. than 150 wells were converted to SEGY format and loaded into the workstation as Computer workstation-based interpret a- seismic traces. Trace spacing was selected tional software provides the tools to handle to be approximately 1.6 km, and wells were mass quantities of data for improved site assigned to the nearest trace. Areas of characterization. Changes in computer sparse well control are represented by technologies have allowed integration of blank traces. The section can be inte r- once discrete professional skills and tec h- preted and displayed using all the tools nologies. Major computer technologies i n- available within the workstation enviro n- clude near super computing capabilities on ment. Using just rudimentary workstation interactive desktop workstations, complex tools results in a startling increase in inte r- networking, high capacity mass storage pretation speed. Using the zoom, datum, devices, 3-D visualization software, and and manual and auto-picking tools, the large digital databases. Engineers, geol o- profile was interpreted in a few hours. I n- gists, geophysicists and other enviro n- terpretations were subject to continuous mental professionals will more likely meet verification by examination of the data set at a computer terminal than at a confe r- as a whole. The display capabilities of ence table and communicate over the ne t- workstation-based interpretational software work rather than in formal meetings. Wor k-

348 Proceedings of the 10th Annual Conference on Hazardous Waste Research Figure 9. Structurally-datumed pseudo-seismic profile of gamma-ray logs showing the Dakota aquifer. Light gray tones correspond to low gamma-ray values associated with sandstones, limestones and evaporites. Darker shades correspond to high gamma-ray values from shales. This profile was constructed using gamma-ray logs from over 150 wells. (modified from [4]) stations can radically change the approach greatly enhance data interpretation, allo w- to environmental geology by integrating ing small-scale to large-scale lateral near- technology across disciplines (Figure 10). surface heterogeneities of the alluvium and Teams will become more common-place. deeper bedrock structures to be succes s- No longer can each discipline work on its fully imaged and interpreted. piece of the problem use a colored pencil to put the results on a piece of paper or map, A copy of this paper along with color ve r- pass the results along to the next discipline, sions of the figures are available on the and be done with it. Large volumes of data Kansas Geological Survey's web site should be accessible to and used by all the (http://www.kgs.ukans.edu) on the Petr o- members of a team, permitting reiteration in leum Research Section's recent public a- the constant process of site characteriz a- tions page (http://crude1.kgs.ukans.edu tion, sample design, remediation and policy /cgi-bin/wwwwais.pl). development. ACKNOWLEDGMENT As illustrated above, near-surface ge o- physical methods are capable of imaging Landmark Graphics Corporation provided the shallow subsurface and determining the the interpretation and analysis software u n- geometries of subsurface structures that der an educational grant to the Kansas affect direction and rate contaminant flow. Geological Survey and the University of Various two-dimensional profiles and three- Kansas. dimensional volumes of environmental- oriented geophysical data along with ge o- REFERENCES logic point data from wells can be brought 1. R.W. Buddemeier, R.S. Sawind, D.O. together for an integrated interpretation Whittemore and D.P. Young, Salt Co n- within the computer workstation enviro n- tamination of Ground Water in South- ment. The color display and post-stack Central Kansas, Kansas Geological processing capabilities of the workstation

Proceedings of the 10th Annual Conference on Hazardous Waste Research 349 Technical Newsletter, Landmark Graphic Corporation, Houston, TX, 1995.

6. D. Collins and J. Doveton, Color I m- ages of Kansas Subsurface Geology from Well Logs, Computers and Ge o- sciences, v. 12, 1986.

Figure 10. Integrated team approach to enviro n- mental geology and the central role of computer technology in technology integration.

Survey, Public Information Circular 2, 1995.

2. D. Young, G.H. Garneau, R.W. Bu d- demeier, D. Zehr and J. Lanterman, Saltwater Elevation and Variability in the Great Bend Prairie Aquifer, South- Central Kansas, Kansas Geological Survey Open-File Report 93-55, 1993.

3. R.W. Buddemeier, G. Garneau, J.S. Healey, M.A. Sophocleous, D.O. Whi t- temore, D. Young and D. Zehr, The Mineral Intrusion Project: Report of Progress During the Fiscal Year 1993, Kansas Geological Survey Open-File Report 93-23, 1993.

4. P.A. Macfarlane, J.H. Doveton, H.R. Feldman, J.J. Butler, Jr., J.M. Combes and D.R. Collins, Aquifer/Aquitard Units of the Dakota Aquifer System in Ka n- sas: Methods of Delineation and Sed i- mentary Architecture Effects on Ground Water Flow and Flow Properties, Jou r- nal of Sedimentary Research, B64 (1994) 464-480.

5. T.R. Carr, J. Hopkins, H. Feldman, A. Feltz, J. Doveton and D. Collins, Transforms of Wireline Logs as Pseudo-Seismic Traces, UserNet

350 Proceedings of the 10th Annual Conference on Hazardous Waste Research