C. Childs

Fault overlap zones within developing normal fault systems

Overlap zones between normal faults have been studied using a variety of 2D and 3D seismic reflection datasets. The overlaps are of two types, (i) relay zones in which displacement is transferred between the overlapping faults and (ii) non-relay overlaps in which displacement is not transferred. Overlap zones are continually formed and destroyed during the growth of a fault system. Overlap zones are formed either by interference between initially isolated faults or as a result of bifurcation of a single fault. The mode of overlap formation is reflected in the 3D geometry of the overlapping faults which may be either unconnected or linked at a branch-line or branch-point. Seismic reflection data from regions of growth faulting, and also sandbox analogue data, allow analysis of fault development through time. Reconstructions of the displacement distribution on some faults with sharp bends and associated hanging-wall splays, show that the bends originated as overlap zones which were later breached to form through-going faults. Depending on the displacements of relay-bounding faults, the effect of relay zones on hydrocarbon reservoirs may be to (a) provide structural closure, (b) form gaps in otherwise sealing faults or (c) increase reservoir connectivity.

Formation of segmented normal faults: a 3-D perspective

The interpretation of fault kinematics from geometric data is an essential step in developing an understanding of the growth of fault systems. Constraints on fault geometry are, however, often restricted to 2-D maps or cross-sections. In this article we consider the extent to which kinematic interpretations of faulting benefit from a 3-D, rather than 2-D geometrical perspective. Concentrating on relay zones and segmented normal fault arrays, we suggest that very different interpretations of their evolution arise from the recognition that the propagation directions of faults, and fault segments, will rarely be contained within the inspection plane of 2-D data. A 3-D perspective favours an interpretation in which the segments of a fault array are kinematically interrelated from their initiation. Individual segments in such systems may link into a single fault surface out of the plane of inspection or may be unconnected in 3-D. We argue that this interpretation, which conflicts with the often suggested model of incidental overlap of originally isolated faults, should be the preferred model for the generation and growth of segmented normal fault arrays.

The shapes, major axis orientations and displacement patterns of fault surfaces

Abstract Displacement contour diagrams constructed using seismic reflection data and coal-mine plans are analysed to establish the factors determining the dimensions, shapes and displacement patterns of normal faults. For blind isolated normal faults in layered sequences the average aspect ratio is 2.15, with sub-horizontal major axes. Earthquake slip-surface aspect ratios range from 0.5 to 3.5 and are independent of slip orientation. The principal control on the shape of blind isolated faults is mechanical anisotropy associated with rock layering, resulting in layer-parallel elongation of fault surface ellipses. Faults that intersect the free surface and/or interact with nearby faults have aspect ratios ranging from 0.5 to 8.4, and are referred to as restricted. Restriction of fault growth has various effects including: (i) reduced curvature of the tip-line and of displacement contours; and (ii) increased displacement gradients in the restricted region. Many faults are restricted at more than one place on their tip-line loop and so have highly irregular shapes and displacement patterns. Subsequent linkage of interacting faults produces combined faults with aspect ratios within the normal range for unrestricted faults. Lateral interaction between faults does not necessarily lead to a change in the power-law exponent of the fault population.

An alternative model for the growth of faults

Conventional growth models suggest that faults become larger due to systematic increases in both maximum displacement and length. We propose an alternative growth model where fault lengths are near-constant from an early stage and growth is achieved mainly by increase in cumulative displacement. The model reconciles the scaling properties of faults and earthquakes and predicts a progressive increase in fault displacement to length ratios as a fault system matures. This growth scheme is directly applicable to reactivated fault systems in which fault lengths were inherited from underlying structure and established rapidly; the model may also apply to some non-reactivated fault systems. Near-constant fault lengths during subsequent growth are attributed to retardation of lateral propagation by interaction between fault tips. The model is validated using kinematic constraints from growth strata, which are displaced by a system of reactivated normal faults in the Timor Sea, NW Australia.

Growth of vertically segmented normal faults

The geometry and evolution of vertically segmented normal faults, with dip separations of < ca 11.5 m have been studied in a coastal outcrop of finely bedded Cretaceous chalk at Flamborough Head, U.K. Fault trace segments are separated by both contractional and extensional offsets which have step, overlap or bend geometries. The location of fault trace offsets is strongly controlled by lithology occurring at either thin (ca 1 mm-8 cm) and mechanically weak marl layers or partings between chalk units. Fault segmentation occurred during either fault nucleation within, or propagation through, the strongly anisotropic lithological sequence. An inverse relationship between fault displacement and number of offsets per length of fault trace reflects the progressive destruction of offsets during fault growth. The preservation of fault offsets is therefore dependent on offset width and fault displacement. Fault rock, comprising gouge and chalk breccia, may vary in thickness by 1.5–2.0 orders of magnitude on individual fault traces. Strongly heterogeneous fault rock distributions are most common on small faults (< 10 cm displacement) and are produced mainly by destruction of fault offsets. Shearing of fault rock with increasing displacement gives rise to a more homogeneous fault rock distribution on large faults at the outcrop scale.

Relay zone geometry and displacement transfer between normal faults recorded in coal-mine plans

Abstract Overlap lengths, separations and throw gradients were measured on 132 relay zones recorded on coal-mine plans. Throws on the relay-bounding fault traces are usually ≤ 2 m and individual structures are recorded on only one seam. Throw gradients associated with relay zones are not always higher than on single faults, but asymmetry of throw profiles is diagnostic of relay zones. Bed geometries around larger faults in opencast mines are used to assess the displacement accommodated by shear in the vertical plane normal to the faults and displacement transfer accommodated by shear in the fault-parallel plane. Three-dimensional structure is defined for two relay zones, each recorded on five seam plans. These relay zones are effectively holes through the fault surfaces and overlap occurs between salients or lobes of the parent fault surfaces. Lobes initially terminated at tip-lines but, as the faults grew, gradually rejoined the main fault surfaces along branch lines. This type of relay zone originates by bifurcation of a single fault surface at a locally retarded tip-line and is an almost inevitable result of a tip-line irregularity.

Fault relays, bends and branch-lines

Abstract Branch-lines between normal faults and their sub-parallel splays mapped from 3D seismic reflection data show a range of forms from straight lines to closed loops. The different geometries are interpreted as representing stages in the failure of relay zones and in the progressive replacement of fault tip-lines with fault branch-lines. The geometries of these normal fault branch-lines are similar to those for thrusts previously inferred from limited two-dimensional data. The orientation of the axis of a relay and its associated bends relative to a fault slip direction is identified as an important control on the structures developed within the relay. Neutral restraining and releasing bends can each occur on any fault type (normal, reverse and strike-slip), but data bias is a major factor in determining which bend geometry is most often observed with each fault type. On normal faults the initial relay zone geometry controls the dominant branch-line orientation and the same control is likely on branch-lines associated with the other modes of faulting. A review of the relay geometries and strains occurring with all three modes of faulting highlights the role of the orientation of the mechanical anisotropy of a bedded sequence relative to the orientations of fault surface and slip directions. This relative orientation determines how the relay strain is accommodated and hence the degree of hard-linkage and development of branch-lines.

Kinematic analysis of faults in a physical model of growth faulting above a viscous salt analogue

Abstract Syn-sedimentary faults overlying an extended ductile layer within a sand☐ model provide an analogue for studying some aspects of the geometry and kinematics of basin scale extensional listric faults associated with salt rollers. The displacement analysis techniques used to examine displacement variations over fault surfaces within the sand☐ model, show that displacement varies in a regular and predictable manner. Deviations from simple fault displacement patterns are due to transfers of displacement onto adjacent synthetic faults. Aggregation of the displacements on adjacent synthetic faults produces more regular patterns of displacement distribution and demonstrates the geometric coherence of the aggregated faults. Rapid upward decreases in displacement on the model faults are due to fault growth during sedimentation. Elsewhere in the model, post-sedimentary faulting is associated with very low upward displacement gradients and pre- and syn-faulting sedimentation are readily distinguished on cross-sectional fault traces. In three dimensions a non-horizontal line separating the pre-faulting and syn-faulting sequences tracks the lateral extent of a fault through time. Backstripping of fault displacements, by subtracting the fault displacement of a sedimentary horizon from the displacements of underlying horizons, gives the displacement distribution on a fault surface at the time of deposition of the backstripped horizon. Backstripped fault displacement distributions also allow examination of the general fault structure through time. Backstripping of fault displacements on one of the larger faults within the model and on an associated hanging wall synthetic splay, shows that the two structures originated as a relay zone which was breached by lateral propagation of one of the overlapping faults which form the relay zone. The backstripping technique allows determination of the geometrical changes during growth of a fault system.

A Method for Estimation of the Density of Fault Displacements below the Limits of Seismic Resolution in Reservoir Formations

The fracture density per unit volume in a reservoir formation can be measured directly from seismic data for fractures with displacements greater than the limit of seismic resolution. Plots of displacement size vs. cumulative number are produced from seismic interpretations digitized and processed using software developed for the analysis of fault geometries and displacements.

Complexity in fault zone structure and implications for fault seal prediction

In their simplest form, brittle faults consist of a single zone of intense deformation which macroscopically is seen as a slip surface and/or a zone of fault rock. More generally, fault zones have complex geometries with multiple slip surfaces and/or deformation zones. The most common pattern in complex fault zones observed at outcrop is a fault zone bounded by a pair of sub-parallel slip surfaces. In three dimensions, fault zones bounded by paired slip surfaces alternate both laterally and up/down dip with areas of only one slip surface. Within this overall framework, a range of fault rocks is irregularly distributed as spatially impersistent sheets and lenses. Due to seismically irresolvable complexities of fault zone structure, the juxtapositions of footwall and hangingwall rocks predicted from seismic data will in most cases be different from those actually present. The importance of such differences to the prediction of across-fault connectivity, of both hydraulically passive and hydraulically active fault zones, is strongly dependent on the reservoir sequence. Connectivities are calculated for hydraulically passive and active faults offsetting an Upper Brent Reservoir sequence. Shaley fault rocks within brittle fault zones often represent a spatially persistent, although variable thickness, component of the zones and provide a basis for the application of empirical methods of fault seal prediction to brittle faults. The distribution of fault rocks cannot be characterised from well data, raising the question of whether purely deterministic methods for fault seal prediction can ever be successful. The way forward is refinement of current empirical methods by achieving a more detailed characterisation of sub-surface faults, allowing more quantitative comparisons of target faults with those of known sealing behaviour.