Improving Reservoir Characterization and Modelling of Aeolian Systems using Process-based Geometrical Models and Multiple Point Statistics

Aeolian reservoirs are commonly considered to be homogeneous “tanks of sand”. However, under certain conditions (e.g. viscous oil, deep burial, low permeability) they will typically contain significant spatial anisotropy that is commonly not captured during conventional subsurface reservoir modelling. Significant permeability contrasts and spatial heterogeneity controlled by bedform arrangements are present.

Current geostatistical algorithms enable the reproduction of spatial statistics, but are unable to integrate sedimentological rules that govern the spatial arrangement of deposition and stratification of principle facies and therefore fluid flow.

Process-based geometry models produce detailed and geologically realistic models that honour depositional rules, but are difficult to condition to observed well data. To address this issue a new workflow in which the results of process based models are used as “Training images” for Multi-Point Statistics This methodology has been tested on the high-quality outcrops of the Page Sandstone Formation, Arizona, which provide a unique opportunity to study bedform geometry and dune architecture on multiple scales.

The work flow involved characterising the aeolian succession from vertical sections (analogous to wells), then creating process based models of the dunes bodies. These were then used to condition the MPS models and finally compared back to drone derived, photogrammetric Virtual Outcrops of the Page. Reservoir models were created at reservoir (km) scale that accurately captures the individual bedform geometries and heterogeneities observed at the outcrop. These models are sufficiently generic to be applicable at an inter-well scale while remaining sufficiently specific to depict the expected architectural complexity contained within aeolian depositional environments.

This methodology can be applied to differing aeolian systems, with different parent dunes, as well as other depositional environments including deepwater reservoirs, where similar architectural elements can be modelled accurately using only a few simple geobodies.