Alan P. Byrnes

Improving ground-penetrating radar data in sedimentary rocks using deterministic deconvolution

Abstract Resolution is key to confidently identifying unique geologic features using ground-penetrating radar (GPR) data. Source wavelet “ringing” (related to bandwidth) in a GPR section limits resolution because of wavelet interference, and can smear reflections in time and/or space. The resultant potential for misinterpretation limits the usefulness of GPR. Deconvolution offers the ability to compress the source wavelet and improve temporal resolution. Unlike statistical deconvolution, deterministic deconvolution is mathematically simple and stable while providing the highest possible resolution because it uses the source wavelet unique to the specific radar equipment. Source wavelets generated in, transmitted through and acquired from air allow successful application of deterministic approaches to wavelet suppression. We demonstrate the validity of using a source wavelet acquired in air as the operator for deterministic deconvolution in a field application using “400-MHz” antennas at a quarry site characterized by interbedded carbonates with shale partings. We collected GPR data on a bench adjacent to cleanly exposed quarry faces in which we placed conductive rods to provide conclusive groundtruth for this approach to deconvolution. The best deconvolution results, which are confirmed by the conductive rods for the 400-MHz antenna tests, were observed for wavelets acquired when the transmitter and receiver were separated by 0.3 m. Applying deterministic deconvolution to GPR data collected in sedimentary strata at our study site resulted in an improvement in resolution (50%) and improved spatial location (0.10–0.15 m) of geologic features compared to the same data processed without deterministic deconvolution. The effectiveness of deterministic deconvolution for increased resolution and spatial accuracy of specific geologic features is further demonstrated by comparing results of deconvolved data with nondeconvolved data acquired along a 30-m transect immediately adjacent to a fresh quarry face. The results at this site support using deterministic deconvolution, which incorporates the GPR instruments unique source wavelet, as a standard part of routine GPR data processing.

Characterization of the Mississippian chat in south-central Kansas

To understand production from low resistivity-high porosity Mississippian chat reservoirs in south-central Kansas it is necessary to understand the nature of deposition and diagenesis, how tectonics is a factor, the lithofacies controls on petrophysical properties, and log response to these properties. The initial mudstones to sponge-spicule wacke-packstones were deposited in transgressive-regressive (T-R) cycles on a shelf to shelf margin setting, resulting in a series of shallowing-upward cycles. Sponge-spicule content appears to increase upward with increasing cycle thickness. After early silicification, inter- and post-Mississippian subaerial exposure resulted in further diagenesis, including sponge-spicule dissolution, vuggy porosity development in moldic-rich rocks, and autobrecciation. Meteoric water infiltration is limited in depth below the exposure surface and in distance downdip into unaltered, cherty Cowley Formation facies. Areas of thicker preserved chat and increased diagenesis can be correlated with structural lineaments and, in some areas, with recurrent basement block movement. Combination of folding or block fault movement prior to or during development of the basal Pennsylvanian unconformity, sponge-spicule concentration, and possibly thickness of overlying bioclastic wacke-grainstones resulted in variable reservoir properties and the creation of pods of production separated by nonproductive cherty dolomite mudstones. These events also resulted in alteration of the depositional cycles to produce a series of lithofacies that exhibit unique petrophysical properties. From bottom to top in a complete cycle seven lithofacies are present: (1) argillaceous dolomite mudstone, (2) argillaceous dolomite mudstone that has chert nodules, (3) clean dolomite mudstone that has nodular chert, (4) nodular to bedded chert, (5) autoclastic chert, (6) autoclastic chert that has clay infill, and (7) bioclastic wacke-grainstone. The uppermost cycle was terminated by another lithofacies, a chert conglomerate of Mississippian and/or Pennsylvanian age. The chert facies exhibit porosities ranging from 25 to 50% and permeabilities greater than 5 md. The (Begin page 86) cherty dolomite mudstones, argillaceous dolomite mudstones, and bioclastic wacke-grainstones exhibit nonreservoir properties. Reservoir production, numerical simulation, and whole core data indicate fracturing can be present in chat reservoirs and can enhance permeability by as much as an order of magnitude. Capillary pressure data indicate the presence of microporosity and can explain high water saturations and low resistivity observed in wire-line logs. Relative permeabilities to oil decrease rapidly for saturations greater than 60% and may be influenced by dual pore systems. Archie cementation exponents increase from 1.8 for mudstones to more than 2.5 in the cherts that have increasing sponge-spicule mold and vug content. Detailed modified Pickett plot analysis of logs reveals critical aspects of chat character and can provide reliable indices of reservoir properties and pay delineation. Models developed provide additional insight into the chat of south-central Kansas and understanding of the nature of controls on shallow-shelf chert reservoir properties.

Flow unit modeling and fine-scale predicted permeability validation in Atokan sandstones: Norcan East field, Kansas

Characterizing the reservoir interval into flow units is an effective way to subdivide the net-pay zone into layers for reservoir simulation. Commonly used flow unit identification techniques require a reliable estimate of permeability in the net pay on a foot-by-foot basis. Most of the wells do not have cores, and the literature is replete with different kinds of correlations, transforms, and prediction methods for profiling permeability in pay. However, for robust flow unit determination, predicted permeability at noncored wells requires validation and, if necessary, refinement. This study outlines the use of a spreadsheet-based permeability validation technique to characterize flow units in wells from the Norcan East field, Clark County, Kansas, that produce from Atokan aged fine- to very fine-grained quartzarenite sandstones interpreted to have been deposited in brackish-water, tidally dominated restricted tidal-flat, tidal-channel, tidal-bar, and estuary bay environments within a small incised-valley-fill system. The methodology outlined enables the identification of fieldwide free-water level and validates and refines predicted permeability at 0.5-ft (0.15-m) intervals by iteratively reconciling differences in water saturation calculated from wire-line log and a capillary-pressure formulation that models fine- to very fine-grained sandstone with diagenetic clay and silt or shale laminae. The effectiveness of this methodology was confirmed by successfully matching primary and secondary production histories using a flow unit-based reservoir model of the Norcan East field without permeability modifications. The methodologies discussed should prove useful for robust flow unit characterization of different kinds of reservoirs.

Improving resolution and understanding controls on GPR response in carbonate strata: Implications for attribute analysis

For more than a decade, environmental, engineering, groundwater, and shallow stratigraphic studies have demonstrated and advanced the usefulness of ground-penetrating radar (GPR) in lithified and unconsolidated sedimentary deposits (e.g., see Neal, 2004 and references therein). Despite the advances, important questions still remain on factors that control the actual appearance and characteristics of GPR reflections and diffractions in any given setting.

Multiscale Geologic and Petrophysical Modeling of the Giant Hugoton Gas Field (Permian), Kansas and Oklahoma, U.S.A.

Reservoir characterization and modeling from pore to field scale of the Hugoton field (central United States) provide a comprehensive view of a mature giant Permian gas system and aid in defining original gas in place and the nature and distribution of gas saturation and reservoir properties. Both the knowledge gained and the techniques and workflow employed have implications for understanding and modeling reservoir systems worldwide that have similar geologic age and reservoir architecture (e.g., Gwahar and North fields, Persian Gulf). The Kansas–Oklahoma part of the field has yielded 34 tcf (963 billion m3) gas throughout a 70-yr period from more than 12,000 wells. Most remaining gas is in lower permeability pay zones of the 557-ft (170-m)-thick, differentially depleted, layered reservoir system. The main pay zones represent 13 shoaling-upward, fourth-order marine-continental cycles comprising thin-bedded (6.6–33-ft; 2–10-m), marine carbonate mudstone to grainstone and siltstones to very fine sandstones and have remarkable lateral continuity. The pay zones are separated by eolian and/or sabkha red beds having low reservoir quality. Petrophysical properties vary among 11 major lithofacies classes. Neural network procedures, stochastic modeling, and automation facilitated building a detailed full-field three-dimensional (3-D) 108-million-cell cellular reservoir model of the 10,000-mi2 (26,000-km2) area using a four-step workflow: (1) define lithofacies in core and correlate to electric log curves (training set); (2) train a neural network and predict lithofacies at noncored wells; (3) populate a 3-D cellular model with lithofacies using stochastic methods; and (4) populate model with lithofacies-specific petrophysical properties and fluid saturations.

4D seismic monitoring of the miscible CO2 flood of Hall-Gurney Field, Kansas, U.S.

Time-lapse seismic monitoring of enhanced oil recovery (EOR) programs has been described by several authors over the last 15 years. Advances in equalization techniques have allowed legacy data (in some cases 15 years prior to a monitoring survey) to be used as a baseline. Sensitivity of seismic data to subtle changes in pore composition has been notably improved with the development of dozens of new attribute analysis techniques in the last several years. Unique to our application is the use of nonamplitude, noninversion attributes for monitoring the effectiveness of EOR in thin, shallow (less than 1000 m) carbonates. Considering the very thin (< 5 m) reservoir interval, this first-time, high-resolution 4D survey needed to be highly repeatable, low-cost, high signal-to-noise, and include several repeat surveys prior to breakthrough. Devel-opment of a highly accurate map of CO2 progression through this field was complicated by difficulties propagating high-frequency signal in this near-surface setting, maintaining uniform fold coverage throughout the optimum offset range, vulnerability of signal to contamination by ground roll and air- coupled wave within the noise cone, weak fluid effects due to high carbonate stiffness, and complexity in porosity distribution. Primary and secondary production phases of the 70-year-old Hall-Gurney Field are nearly done. The Pennsylvan-ian Lansing-Kansas City (L-KC) groups have yielded 90 million bls of the 155 million bls cumulative production in the multipay field. Primary production was begun in 1931 and was followed by extensive waterflooding in the 1950s–60s. Waterfloods reached their economic limits in the 1970s–80s but bypassed oil represents a significant resource for CO2 miscible flooding, a third development phase. Reservoir rock of the Hall-Gurney oil field consists of shallow (about 900 m), thin (3.5-5 m), oomoldic limestone layers in L-KC with largely moldic porosities that decrease downward from around 35% to 12%. This reservoir is …

Comparison of ground-penetrating radar response and rock properties in a sandstone-dominated incised valley-fill deposit

Alex Martinez*, Exxon Exploration Company, Houston, TX (formerly Kansas Geological Survey, Lawrence, KS); Alan P. Byrnes, Kansas Geological Survey, Lawrence, KS; D. Scott Beaty, Chevron USA, Midland, TX formerly Kansas Geological Survey, Lawrence, KS); Timothy R. Carr, Kansas Geological Survey, Lawrence, KS; and James M. Stiles, Department of Electrical Engineering and Computer Science, University of Kansas, Lawrence, KS

Analysis of Critical Permeabilty, Capillary Pressure and Electrical Properties for Mesaverde Tight Gas Sandstones from Western U.S. Basins

Although prediction of future natural gas supply is complicated by uncertainty in such variables as demand, liquefied natural gas supply price and availability, coalbed methane and gas shale development rate, and pipeline availability, all U.S. Energy Information Administration gas supply estimates to date have predicted that Unconventional gas sources will be the dominant source of U.S. natural gas supply for at least the next two decades (Fig. 1.1; the period of estimation). Among the Unconventional gas supply sources, Tight Gas Sandstones (TGS) will represent 50-70% of the Unconventional gas supply in this time period (Fig. 1.2). Rocky Mountain TGS are estimated to be approximately 70% of the total TGS resource base (USEIA, 2005) and the Mesaverde Group (Mesaverde) sandstones represent the principal gas productive sandstone unit in the largest Western U.S. TGS basins including the basins that are the focus of this study (Washakie, Uinta, Piceance, northern Greater Green River, Wind River, Powder River). Industry assessment of the regional gas resource, projection of future gas supply, and exploration programs require an understanding of reservoir properties and accurate tools for formation evaluation. The goal of this study is to provide petrophysical formation evaluation tools related to relative permeability, capillary pressure, electrical properties and algorithms for wireline log analysis. Detailed and accurate moveable gas-in-place resource assessment is most critical in marginal gas plays and there is need for quantitative tools for definition of limits on gas producibility due to technology and rock physics and for defining water saturation. The results of this study address fundamental questions concerning: (1) gas storage; (2) gas flow; (3) capillary pressure; (4) electrical properties; (5) facies and upscaling issues; (6) wireline log interpretation algorithms; and (7) providing a web-accessible database of advanced rock properties. The following text briefly discusses the nature of these questions. Section I.2 briefly discusses the objective of the study with respect to the problems reviewed.

Arbuckle Group Platform Strata in Kansas: A Synthesis

Cambrian–Ordovician Arbuckle Group rocks in Kansas occur entirely in the subsurface and are absent only in areas of northeastern and northwestern Kansas and over ancient uplifts and buried Precambrian highs. During Arbuckle deposition, Kansas was located approximately between 20 and 30 south of the equator and south of the Transcontinental arch. Because of the lack of biostratigraphic data and a chronostratigraphic framework, correlation of Arbuckle Group subunits has relied predominantly on lithologic character and insoluble residues. Core studies reveal shallow-water, carbonate-dominated depositional facies that are stacked in vertical cycles (ranging from less than 1 m [3.3 ft] to several meters thick) and cycle sets. Eight depositional facies predominate: (1) clotted algal boundstone (subtidal conditions) with porosities less than 6% and permeabilities less than 0.1 md; (2) muddy to grainy laminated algal boundstones (subtidal to peritidal conditions); muddy textures exhibit porosities generally less than 6% and permeabilities less than 0.1 md, and grainy textures represent some of the best reservoir rock ranging in porosity up to 32% and permeability up to 1500 md; (3) peloidal packstone-grainstone (subtidal to peritidal conditions) with porosities from 0 to 4% and permeabilities generally below 0.005 md; (4) mixed packstone-grainstone (subtidal to peritidal conditions) with porosities from 6 to 18% and permeabilities from 0.1 to 50 md; (5) ooid packstone-grainstone (subtidal to peritidal conditions) with porosities from 11 to 30% and permeabilities from 10 to 1500 md; (6) wackestone-mudstone (restricted subtidal to peritidal conditions) with porosities from 0 to 17% and permeabilities from less than 0.0001 to 1000 md; (7) intra-Arbuckle shale (low-energy subtidal to peritidal and, perhaps, supratidal conditions); and (8) intraclastic conglomerate and breccia, fracture-fill shale, and chert in variable abundances. The abundance of intercrystalline, moldic, fenestral, and vuggy porosity is related to depositional facies, early diagenesis, and dolomitization and not necessarily to karst influence from the upper super-Sauk subaerial exposure surface. Arbuckle reservoirs historically have been viewed as fracture-controlled karstic reservoirs with porosity and permeability influenced by basement structural patterns and subaerial exposure. Although fractures and karst influence production in some Arbuckle reservoirs, the presence of reservoirs where water drive is minimal or absent indicates the dominance of matrix porosity. The Arbuckle in Kansas can be characterized using three end-member reservoir architectures, representing fracture-, karst-, and matrix-dominated architectural systems. Lithofacies and stratal packaging of reservoir and nonreservoir strata exert an important influence in all three reservoir architectures.

Schaben field, Kansas: Improving performance in a Mississippian shallow-shelf carbonate

Schaben field (Kansas), located along the northeastern shelf of the Hugoton embayment, produces from Mississippian carbonates in erosional highs immediately beneath a regional unconformity. Production comes from depths of around 4400 ft (1342 m) in partially dolomitized shelf deposits. A detailed reservoir characterization/simulation study, recently performed as part of a Department of Energy Reservoir Class Oil Field Demonstration Project, has led to important revision in explanations for observed patterns of production. Cores recovered from three new data wells identify three main facies: spicule-rich wackestone-packstone, echinoderm wackestone/packstone/grainstone, and dolomitic mudstone-wackestone. Reservoir quality is highest in spicule-rich wackestone/packstones but is subject to a very high degree of vertical heterogeneity due to facies interbedding, silicification, and variable natural fracturing. The oil reservoir is underlain by an active aquifer, which helps maintain reservoir pressure but supports significant water production. Reservoir simulation, using public-domain, PC-based software, suggests that infill drilling is an efficient approach to enhanced recovery. Recent drilling directed by simulation results has shown considerable success in improving field production rates. Results from the Schaben field demonstration project are likely to have wide application for independent oil and exploration companies in western Kansas.