Depth structure maps were estimated from both their time structure maps and well top data using a geostatistical approach, and seismic attribute time volumes were converted to their depth volumes in the oil-sands reservoir in Alberta, Canada. The objective of the study was to provide basic information in the depth domain for constructing three-dimensional (3D) geological depth models of the oil-sands reservoir.
Well log information is precise vertically, but sparse sampled data horizontally. Seismic interpretation time data are less precise than well log data vertically, but dense, well-sampled data horizontally. The seismic interpreted horizons and the well depth information of the reservoir top and bottom were geostatistically integrated to estimate the depth structure maps of them. Time volumes were also converted to their depth volumes using simple relationships between the time structure maps and the depth structure maps of the reservoir top and bottom on the assumption that the seismic interval velocities in the reservoir interval vary horizontally, but they were vertically constant.
After the study, five new wells were planned and drilled in the study area. Depth prediction errors of the reservoir top and bottom were less than two meters at the wells. The geostatistical approach continued to be applied to the other four 3D seismic surveys acquired later in the adjacent oil-sands areas, and depth structure maps and depth seismic attribute volumes estimated by the method were helpful for geological interpretation studies in those area.
View full abstract