Quantitative reservoir characterization using rock physics, seismic and geological constraints – examples from Semliki Basin in Albertine Graben, Western Uganda

Abstract

The Albertine Graben located in the Western arm of the East African Rift System holds Uganda’s oil and gas discoveries. The Semliki basin has not yet yielded the government of Uganda with commercial quantities of oil and gas. In this basin, there is lack of a clear understanding of the lateral distribution of the reservoir properties (porosity, saturation and lithology). This study utilized rock physics modeling together with rock physics templates and inverse rock physics modeling to delineate reservoir properties. The two rock physics models namely; the patchy constant cemented model and the constant cement model described the Upper Pliocene Formation and Upper Miocene Formation respectively. The analysis revealed high sandstone bulk modulus values attributed to the presence of feldspars, micas, and calcareous clays in the reservoir units. The rock physics templates of Vp/Vs ratio versus acoustic impedance indicate 5% gas for the Upper Miocene Formation and 50% gas for the Upper Pliocene Formation. In both Formations, three lithological zones interpreted as gas sand, brine sand, and shaly sand were observed. Also, the Upper Miocene Formation data were compared with selected regions of Turaco 2 and Turaco 3 that indicated a perfect match within the plots of Vp data with Vp modelled Kanywataba well data. This indicates that, the three wells are in the same or related environment and a compaction regime. The Upper Miocene Formation indicates the presence of a chemical compaction regime which implies, that it has undergone diagenesis. The evidence from logs indicates higher velocities, lower porosity, and some amount of cementation associated with this Formation. The Inverse rock physics modelling constrained by seismic inversion data takes into account non-uniqueness and data error propagation issues. The procedures are designed to obtain the most likely estimate mean, weighted mean and posterior mean of the reservoir properties and there is a good match between measured and modelled porosity data. Fluid saturation data were less successfully predicted, but this was most probably a result of lack of real saturation logs for use in the calibration of the rock physics model. A misfit between observed and predicted lithology is attributed to the uncertainties in defining the mineral properties. The integrated approach reveals that a high fraction of porosities correlated with the low fraction clay volumes and this indicates two distinct reservoir units interpreted as Oluka and Kakara Formations.