Data from a large-scale outcrop analogue (Upper Cretaceous Blackhawk Formation, Wasatch Plateau, central
Utah, USA) were used to construct three-dimensional, object-based reservoir models of low to moderate net-to-gross (NTG)
ratios (11–32%). Two descriptive spatial statistical measures, lacunarity and Ripley’s K function, were used to characterize
sandbody distribution patterns in the different models. Lacunarity is sensitive to sandbody abundance and NTG ratio, while
Ripley’s K function identifies clustered, random and regular spacing of sandbodies. The object-based modelling algorithm
reproduces sandbody dimensions and abundances, but patterns of sandbody distribution generated by river avulsion are
poorly replicated because pseudo-well spacing provides only limited constraint on sandbody positions.
In common with previous studies, the connected sand fraction in the reservoir models increases with increasing NTG
ratio and increasing range of sandbody orientations, but there is significant stochastic variation around both of these trends.
In addition, low NTG reservoir models in which sandbodies exhibit strong clustering may also have a low connected sand
fraction across the model volume because the sandbody clusters are widely spaced and, thus, tend to be isolated from each
other. Consequently, connected sand fraction could be overestimated if avulsion-generated sandbody clusters are not identified
and replicated in models of such reservoirs.