Three-dimensional (3D) seismic data have insufficient resolution to image faults with throws less than ca. 20 m. Despite their potential impact on reservoir performance, the true 3D structure of sub-seismic scale fault networks has only been determined in exceptional circumstances, for example by cutting serial sections through faults in unconsolidated sediments, or within active opencast mines. A further difficulty has been that most structural datasets from onshore analogues have been collected using traditional mapping techniques, which require the 3D geology and surface topography to be projected onto a 2-D plane (or along a 1-D scan line). Terrestrial laser scanning (TLS) now enables structural geologists to produce 3D representations of geological outcrops (digital outcrop models), but faults are generally recorded as intersections on the outcrop surface, rather than planes.
We use a digital outcrop model of sub-seismic scale, post-depositional normal faults from SE Scotland to illustrate a methodology for extrapolating fault surface traces to create a fully 3D fault model. The faults are exposed on the foreshore and in cliffs behind the beach. We created a pseudo-3D seismic grid across the digital outcrop model and extrapolated fault sticks from the surface intersections using geologically-driven rules. The cliff section provides constraints on the range of permissible fault dips, fault heights, and the impact of host rock stratigraphy on fault bifurcation. The geometries of larger-scale post-depositional normal faults observed in 3D seismic datasets have been used to guide our interpretations of fault tip- and branch-lines. These geological rules provide a conceptual framework to generate multiple 3D realisations from a single digital outcrop model, which could be used to further test the implications of small faults on production flow.