oa Reconciling well test permeability with geological heterogeneities in a giant carbonate field, northern Oman

A significant development target in this giant carbonate reservoir is the Natih E unit, which is an approximately 150-m-thick, third-order, shallowing upward succession of intra-shelf basinal to lagoonal depositional settings. Plug data show mainly low to moderate matrix permeabilities. However, a subset of the plug data and a number of well tests suggested that permeability may be an order of magnitude higher in some cases. This may be related to under-sampling of plugs due to sample failure in highly leached (friable) layers. or to fracturing or a combination of both. The implications for development designs are considered significant. Thief zones can be related to either layers with very high matrix permeability or to dense layers with highly connected, bedbound<br>fracture systems. Both types have been observed and their occurrence can be related to a combination of depositional setting, depositional cyclicity and a complex diagenetic overprint. Layers crossing larger fractures superimposed on the complex matrix system are a significant additional heterogeneity, which can give rise to well tests with high permeability. A key uncertainty is the connectivity of the fracture system. Previous modelling efforts indicated limited lateral connectivity between fracture sets and highlighted shortcomings in using stochastic methods to place fractures in areas of extensive well control. In our new models, fracture corridors have been model-constrained by high-resolution seismic attributes. Bed-bound, distributed fractures were represented by matrix permeability scalars with limited vertical transmissivity. An improved understanding of the field’s kinematic history, combined with field data (evidence of short-cutting, tracer tests, PLTs), was used to constrain those components of the fracture system associated with higher permeabilities. Combining static models that contain various combinations of thief zones and layer-crossing fracture systems resulted in arrays of models, which – after history match – were used to optimize field development.