Modelling a Mid-Slope Channel Complex from Outcrop Data: Example from Skeiding, the Laingsburg Karoo, South Africa

Brunt, Rufus L.¹, Jamie K. Pringle¹, David M. Hodgson¹, Christopher J. Haigh¹, Stephen S. Flint¹ (1) University of Liverpool, Liverpool, United Kingdom

Examination of deepwater systems in outcrop supplies a wealth of small-scale facies and spatial information unattainable in subsurface investigations. Collated outcrop data may be used to build detailed digital outcrop models which are directly transferable to commercial reservoir modelling packages. The use of outcrop collected subseismic scale data enhances the value of the resultant reservoir models, the details of which may be transferable to similarly complex subsurface systems.

Skeiding provides an oblique section through an 80 meter thick mid-slope channel complex continuously exposed over more than 1.3 km. A combined investigative methodology at Skeiding utilised photomontages, logged sections, and direct digital capture through ground-based LiDAR, dGPS and GPR. Key erosional surfaces and facies boundaries were walked out and transferred to a correlation panel built from 63 logged sections. Through synthesis of the data eight individual channels were identified, which under further investigation were divided into lower and upper channel complexes. Channels in the lower high-aspect-ratio complex are stacked vertically, whereas the upper low-aspect-ratio channels are stacked laterally and therefore are amalgamated to a greater degree. Superimposition of 1D and 2D outcrop data onto 3D dGPS and LiDAR field data gave spatial accuracy to the correlated facies intervals and architectural bounding surfaces characterised in the photomontage and correlation panels. Larger scale features such as channels and infilling packages could then be extended into 3D by assuming trends associated with measured palaeoflow data. End-member reservoir models could then be constructed, exploring the effects of upstream and downstream erosiveness of the component channels.