Applying basin modeling technology to predict high-resolution fluid distribution and properties, taking into account the local high resolution of the sediment properties in fields and prospects has been a growing need for the past 10 yr. To minimize simulation time, local grid refinement (LGR) techniques have been introduced. The main interest of LGR is to gain computing time and memory with respect to classical methods, such are Tartan gridding. With LGR, it is possible to define local areas with high resolution in a regional model. The LGR approach gives a more detailed picture of individual fields or prospects while using models of reasonable size. The models incorporate various regional elements of the petroleum system that include source rock and seal, for instance, to obtain a detailed understanding of local processes such as trap filling history.

To validate the LGR approach, a benchmark is performed. It aims at comparing the different refinement methods: (1) a high-resolution grid, (2) an LGR grid, (3) a Tartan grid, and (4) windowing. To test the behavior of and the results produced by LGR, this method is applied to a real case study from northern Kuwait. It illustrates the coupling between LGR and compositional three-dimensional Darcy flow modeling to predict the distribution of hydrocarbon composition and properties in local reservoir rock areas where accumulations are predicted. The LGR approach efficiently fills the gap between conventional basin modeling and reservoir modeling. Its application in northern Kuwait provides useful guidelines to predict API gravities, gas-oil ratio, and oil-water contact depth estimates in new prospects.