Peter J. Vrolijk

On the mechanical role of smectite in subduction zones

Smectites, a group of hydrated phyllosilicate minerals, are often a main component of sediments entering a subduction zone. They are common in hemipelagic sediments and consist of an authigenetic component that arises from the alteration of volcanic ash intercalated in the sediment and a detrital component. Smectite-rich strata are structurally weak, retain water, and are highly porous. A review of sediment lithology in subduction zones around the world suggests that decollement zones may preferentially form in smectite-rich horizons. One consequence of this observation is that the assignment of high frictional resistance to materials in a decollement may be unwarranted. At deeper structural levels smectite is transformed to illite, a mechanically stronger mineral. Estimates of the depth of this transition coincide with the depth at which earthquakes are first detected along plate interfaces. I speculate that the smectite-illite transition in subduction zones may be an important component of sediment consolidation processes that trigger interplate seismicity.

Fault dating in the Canadian Rocky Mountains: Evidence for late Cretaceous and early Eocene orogenic pulses

Fault rocks from the classic Rocky Mountain foreland fold-and-thrust belt in southwestern Canada were dated by Ar analysis of clay grain-size fractions. Using X-ray diffraction quantification of the detrital and authigenic component of each fraction, these determinations give ages for individual faults in the area (illite age analysis). The resulting ages cluster around 72 and 52 Ma (here called the Rundle and McConnell pulses, respectively), challenging the traditional view of gradual forward progression of faulting and thrust-belt history of the area. The recognition of spatially and temporally restricted deformation episodes offers field support for theoretical models of critically stressed wedges, which result in geologically reasonable strain rates for the area. In addition to regional considerations, this study highlights the potential of direct dating of shallow fault rocks for our understanding of upper-crustal kinematics and regional tectonic analysis of ancient orogens.

Geochemical and geothermal evidence for fluid migration in the Barbados Accretionary Prism (ODP leg 110)

Measured geochemical and geothermal effects of fluid migration in the northern Barbados accretionary prism indicate that: (1) fluid flows laterally along low-angle active faults; (2) the entire length of the Leg 110 transect is a fluid discharge zone; (3) migration results in heating of sediments across the transect but most strongly at the deformation front; and (4) migration results in dilution of pore water-Cl and {sup 18}O enrichment more in the arcward than the seaward side of the prism, suggesting that the Cl-poor, {sup 18}O-rich fluids that intrude the prism are derived from smectite-illite reactions occurring arcward. These conclusions have strengthened those derived from the initial appraisal of shipboard data, modified ideas about where fluids are heated most and where the geochemistry is strongly altered, and provided new ideas about the sources of fluids and their relative contributions.

Gas geochemistry of the Mobile Bay Jurassic Norphlet Formation: Thermal controls and implications for reservoir connectivity

The Mobile Bay gas field is located offshore Alabama in the northern Gulf of Mexico. Production is from eolian dunes of the Jurassic Norphlet sandstone at depths exceeding 6100 m (20,000 ft) and temperatures greater than 200C. Reservoir connectivity and compositional variation, including the distribution of nonhydrocarbon gases (H2S and CO2), are critical factors in production strategy. To evaluate the controls on compositional variation and connectivity, detailed molecular and isotopic analyses were conducted for 29 wells. Analysis of volatiles in fluid inclusions suggests that the field was originally filled with oil that subsequently cracked to gas. In addition to the thermal destruction (cracking) of oil, the process of thermochemical sulfate reduction (TSR) continues to destroy the remaining hydrocarbons through oxidation of gas and reduction of sulfate to form H2S and CO2. The variable extent of the TSR process at Mobile Bay results in a wide range of hydrocarbon and H2S compositions. Condensates are almost exclusively composed of diamondoids whose composition appears controlled by H2S concentrations. In contrast to hydrocarbon and H2S contents, CO2 concentrations are relatively constant throughout the field. Carbon isotopic ratios for CO2 correlate positively with those for wet-gas hydrocarbons but are heavier than expected for CO2 originating from hydrocarbon oxidation via TSR. The narrow range of CO2 contents and heavy isotope ratios suggests that CO2 is regulated by water-rock equilibration and carbonate precipitation. The destruction of the hydrocarbon gas and mineralization of the carbon dioxide product create a volume reduction and an associated drop in reservoir pressure. This process creates several internal sinks (or exits) that may control the spill direction for gas in the field.

Fault-seal analysis using a stochastic multifault approach

We have developed a stochastic multifault method for analysis of the impact of stratigraphic uncertainty on cross-fault leakage at sand-sand juxtapositions. This method assumes that all sand-sand juxtapositions leak across the fault. Stratigraphic uncertainty is modeled by stochastic variation of stratigraphic stacking. Structural uncertainty is addressed through variation of the input. Our objectives were to quantitatively predict the impact of uncertainties in stratigraphic and structural input and to simulate the complex system of structural spills and juxtaposition leak points that control hydrocarbon contact levels in traps with stacked reservoir systems and many faults.Three examples demonstrate how this stochastic multifault method has helped us evaluate uncertainty and understand complex leak fill-and-spill controls. The Ling Gu prospect demonstrates that widespread cross-fault leakage on two crestal faults with throw changes that exceed seal thickness causes only a single hydrocarbon column to accumulate in multiple-stacked reservoirs. This column is controlled by a juxtaposition leak point on a third, deeper fault. We have learned from examples like Ling Gu that the relative size of throw change and seal thickness is a fundamental control on the probability of cross-fault juxtapositions. An example at prospect A demonstrates the sensitivity of hydrocarbon entrapment to small faults in a sand-prone interval with thin seals. The prospect A analysis shows that if seals are thin, faults or channel incisions below seismic resolution can leak hydrocarbons out of stacked reservoirs that are interpreted as unfaulted on seismic data. This introduced significant predrill uncertainty and risk. Guntong field demonstrates that a thin sand in a juxtaposed seal interval can introduce large uncertainty in the prediction of hydrocarbon columns.These examples and many other analyses using the method demonstrate how small changes in stratigraphic and structural input to a fault-seal analysis can introduce significant uncertainty in the predicted range of hydrocarbon volumes. Such uncertainties need to be directly and systematically accounted for in a fault-seal analysis.

Illite polytype quantification for accurate K-Ar age determination

Abstract PolyQuant, a new method to quantify illite polytypes in physical mixtures, has been coupled with Illite Age Analysis (IAA, Pevear 1992) to determine accurate K-Ar ages of the individual polytypes. K-Ar radioisotope dates from illite in shales and sandstones are typically meaningless because these rocks are mixtures of diagenetic and detrital components in unknown proportions. IAA uses linear extrapolation to unmix measured K-Ar ages and establish the end-member ages. This procedure requires samples to have a diagenetic Illite-Smectite (I-S) component mixed with a detrital discrete illite component because they are separable and quantifiable using X-ray diffraction (XRD) of oriented aggregate samples. We extend IAA for use in mixtures of diagenetic and detrital illite with no I-S component by coupling a genetic algorithm (GA) with the WILDFIRE programs (1993) of Reynolds in a new computer program, PolyQuant, which quantifies illite polytypes in a natural sample. Based on first principles of XRD, WILDFIRE is a forward model that calculates hkl reflections by varying parameters such as rotational disorder, octahedral cation occupancy (i.e., cis- and transvacancies), chemical substitutions (K and Fe), orientation, and crystallite thickness. PolyQuant executes several forward models, using the GA to vary the parameters and weight percents of each polytype to automatically achieve a “best fit” with an experimental XRD pattern. Sensitivity tests reveal that changes in the weight percent of an illite polytype, disorder (P0, Pcv), and the orientation function affect the pattern more than other parameters. In addition, PolyQuant successfully quantified illite components in prepared mixtures of 1M illite and 2M1 muscovite and 1Md illite and 2M1 muscovite. We show examples to validate IAA from prepared mixtures of end-members with known ages and also to obtain end-member ages of highly illitic clay fault gouge. Using K-Ar ages and PolyQuant to quantify the 1M and 2M1 components from mixtures of RM-30 (25 ± 1 Ma) and a 2M1 muscovite from Wards (428 ± 9 Ma), IAA determined end-member ages of 19.9 Ma (0-44 Ma) and 429 Ma (408-459 Ma)-in reasonable agreement with the measured results. In addition, we use clay gouge from the Canadian Rockies to date movement along thrust faults. Regional geologic data indicate movement along these faults occurred between 73 and 48 Ma. IAA determined the ages of faulting to be from 79.2 to 53.5 Ma, in good agreement with the known range of fault dates. Our results indicate that neoformation of illite in the fault zone must have modified the ages to younger dates.

Discussion of “Extracting K-Ar ages from shales: a theoretical test”

Abstract Illite age analysis (Pevear, 1992) and similar mixing methods demonstrate nearlinear behaviour when corrected for smectite content in illite-smectite and substitutions for K+ in illite layers. Srodon (1999) found that highly curved mixing lines result from theoretical calculations and challenged the validity of IAA. However, our theoretical calculations indicate near-linear mixing lines result when corrected for the amount of smectite and known K+ substitutions (Na+ or NH4+). We conclude that perfect linearity is unattainable because IAA attempts to unmix a system linearly, but the system is based on an exponential function (K-Ar equation). In practice, we believe IAA can successfully determine the age range of illite diagenesis, even though perfect linearity is not attainable because: (1) significant deviations from linearity occur only if >500 Ma separate the diagenetic and detrital ages; (2) IAA analytical uncertainties result in diagenetic and detrital ages that span a range of ages.

Evolution of structure and permeability of normal faults with clay smear: Insights from water-saturated sandbox models and numerical simulations

Clay smear is difficult to predict for subsurface flow applications and would benefit from an improved understanding of the processes controlling clay smear. We present water-saturated sandbox experiments with large clay smear surfaces (~500 cm2) that couple cross-fault fluid flow measurements with structural analysis of excavated clay smears. We compare measured flow data to numerical flow simulations to develop a tool to evaluate the evolving fault structure. Results show diagnostic relationships between fault structures and cross-fault flow. In experiments with one or two clay layers and a cumulative thickness of 10 mm at 100 mm displacement, normally consolidated clay in a structural domain of graben faulting initially yields hybrid brittle/ductile failure with early breaching of the clay layer and increased cross-fault flux. This is followed by fault backstepping, formation of clay smears, and reworking of clay fragments within the fault. Early formed holes remain open during the evolution of the faults. Fault zones are segmented by fault lenses, breached relays, and clay smears in which sand and clay mix by deformation. Experiments with two clay layers show that holes rarely form at the same position on the fault plane, producing a layered sand-clay fault rock with greater flow path tortuosity and lower permeability than in one-layer experiments. We compare our results with observations of faults in nature and discuss progress toward models with sufficient detail and understanding to allow prediction of flow across evolving faults, first in laboratory models and then in the subsurface.

Relating Structural Elements to Cross-fault Flow - Implications for Fault Permeability Estimation

The evolution of cross-fault fluid flow in sand-clay sequences, especially the role of structural elements is not well understood. We present observations from analogue experiments using an underwater sandbox setup that allows dynamic cross-fault flux measurements. We combine the results of flux measurements with observations of the evolving fault zone in map view as well as with structural information of the clay smear after completing deformation. Carefully excavating the clay smear allows finding holes in the clay as well as relays, horses and other structures. Clay veneers of ~0.1 mm remain intact during the excavation process. We present results of eight experiments with the same total clay volume, of which three experiments had two clay layers with each half the thickness of the other experiments. The results show a more linear evolution of flux with smaller total flux. In experiments with a single clay layer we interpret the initial deformation to be hybrid failure with sudden increase of flux while at later stages fault segmentation and shearing of clay fragments causes less distinct increases of flux. Finally, we observe a process potentially capable of resealing discontinuous clay smears.

Experimental Investigation of Clay Smear Processes in a Geotechnical Direct Shear Apparatus

Clay smear is an important process in siliciclastic sequences, but the details of the evolution of fault zones in sand-clay layered sequences are poorly understood. This project presents the results of an experimental investigation of clay smear processes, using a geotechnical direct shear apparatus.