This paper presents the use of geostatistical inversion for the delineation of Mauddud reservoirs on the flanks and saddle parts between Raudhatain and Sabiriyah fields of North Kuwait. Mauddud formation is sub-divided into ten reservoirs from MaA to MaJ and of those, the most prospective zones are from MaB to MaE, consisting of thick porous carbonate reservoir, with porosity ranging from 15 to 30%. MaE is low quality reservoir, which shows drop in resistivity and increase in water saturation and appears to be tight. MaG shows high impedance and low porosity and in MaH, MaI & MaJ observed presence of shale. Most of the wells were drilled concentrated on the main structures whereas few wells were drilled on the flanks and saddle parts. The plan is to take up the drilling on these parts, but before carrying out any drilling activity, it has been decided to perform reservoir simulation study to reduce the uncertainty.
The challenge is to characterize the reservoir property (porosity) of Mauddud formation on the saddle part and flanks as well as to improve the vertical resolution. Hence, geostatistical inversion was carried out to improve the reservoir description of its internal heterogeneity, increase the vertical resolution of individual reservoir units and map porosity on flanks and the saddle parts. Geostatistical inversion used the Bayesian inference technique to integrate the seismic, well logs, and geostatistics giving global probability of the reservoir and sampling it using Markov Chain Monte Carlo (MCMC). Multiple plausible realizations of geostatistical inversion were generated to output highly detailed P-impedance and lithotype of the reservoir, which were then used to co-simulate the porosity. The result of porosity of this study was validated through recently drilled wells as blind well tests, which gave reasonably good match. Further, the ranking process was carried out to estimate the porosity for the wells planned for drilling on the saddle part.