Published in

Society for Sedimentary Geology (SEPM), Journal of Sedimentary Research, 6(71), p. 868-879, 2001

DOI: 10.1306/051501710868

Links

Tools

Export citation

Search in Google Scholar

Conditioning a process-based model of sedimentary architecture to well data

Journal article published in 2001 by D. J. Karssenberg ORCID, T. E. T{̈o}rnqvist, Te E. Tornqvist ORCID, Js S. Bridge
This paper is available in a repository.
This paper is available in a repository.

Full text: Download

Red circle
Preprint: archiving forbidden
Green circle
Postprint: archiving allowed
Red circle
Published version: archiving forbidden
Data provided by SHERPA/RoMEO

Abstract

Prediction of sedimentary architecture for modeling of fluid flow in hydrocarbon reservoirs and aquifers is accomplished mainly using stochastic, structure-imitating models, because these can be conditioned to data from wells, seismic profiles, and outcrop analogs. Conditioning implies that modeled architecture fits all available observations. However, the sedimentary architecture simulated by such models is commonly unrealistic. Process-based (forward) models potentially provide more realistic prediction and understanding of sedimentary architecture, but these models are not widely used because conditioning to well, seismic, or outcrop data is considered to be very difficult. We show here that conditioning of process-based models to well data is possible in principle, using a 3D alluvial-architecture model as an example. This model considers the formation of alluvial deposits as a predominantly deterministic process, with a single channel belt moving by avulsion over an aggrading floodplain. However, the initial floodplain topography is simulated by a random field, thus yielding different model output for each run. Monte Carlo simulation was used to produce model realizations that fit five hypothetical vertical wells within predetermined tolerance bands. Such simulation allows calculation of the probability of occurrence of channel-belt deposits for each 3D cell in the 3D block of sediments generated by the model, as well as the probability distributions of volumes of channel-belt deposits and connectedness ratios. Adding more conditioning wells increases the precision of model predictions. Application of this approach in practice will require a major effort, particularly in overcoming the anticipated large amounts of computing time.