AAPG Annual Convention and Exhibition

Datapages, Inc.Print this page

Using Markov Chain Analysis to Determine Relative Avulsion Rate and its Impact on Reservoir Architecture in Basin-Floor Lobe Deposits in the Neoproterozoic Upper Kaza Group, Windermere Supergroup, Canada


Recent advances in deep-water research have shown that the seemingly simple “sheet-like” lobe deposits of the mid-fan are made-up of smaller architectural elements, and the spatial arrangement of these elements creates complex reservoir architectures. A major control on the arrangement of lobes and their constituent architectural elements is channel avulsion. Rates of lobe avulsion have been determined by dating lobe deposits in several modern fans, but the factors determining the rate of avulsion and the consequences for reservoir architecture are less well understood. This study aims to address these issues with a combined sedimentological and statistical approach. Data for this study were collected at Castle Creek, southern Canadian Cordillera from basin-floor lobe deposits of the Neoproterozoic Upper Kaza Group, Windermere Supergroup. A single ∼1 km wide and ∼100 m thick outcrop was logged and mapped in detail. The Upper Kaza Group is interpreted to be proximal mid-fan deposits comprising depositional lobes, which in turn are made up of five architectural elements: scours (and their fills), avulsion splays, distributary channels, terminal splays and fine-grained deposits. These elements are interpreted to be systematically arranged, both spatially (along a single depositional transect) and temporally (stratigraphically upward). Lobe evolution starts with the formation of isolated scours and avulsion splays in the proximal to mid-lobe. As the lobe matures distributary channels develop in proximal areas of the lobe, which more distally shallow and widen until they merge into terminal splays. Fine-grained deposits accumulate at all times in low energy inter- and intra-lobe areas. Markov chains analysis is used to describe the vertical succession of architectural elements and to test if the vertical succession of these elements exhibits non-random behaviour. The input data is derived from ten ∼90 m long stratigraphic logs and in total 369 stratal elements are included in the analysis. The results indicate that the vertical arrangement of architectural elements is in fact non-random and significantly more heterogeneous than expected from a random distribution of architectural elements. This indicates that the depositional system is highly avulsive, most probably reflecting rates of sedimentation that outpaced subsidence, and consequent backfilling of feeder channels. These conditions, therefore, form a highly heterogeneous reservoir architecture.