AAPG Europe Regional Conference, Global Analogues of the Atlantic Margin

Datapages, Inc.Print this page

Combining Seismic Inversion and Geostatistics to Predict Reservoir Proprieties

Abstract

The main purpose of using inversion is to convert the seismic section in order to more accurately represent the properties of the Earth's subsurface. Often in exploration areas we have data of high uncertainty due to the lack of drilled wells so we have to rely completely on seismic reflection data. To overcome this uncertainty, a geostatistical framework was developed that doesn't require any well-log data within the inversion area. In seismic inverse modelling, the result of reflection seismology, the seismic trace, is reversed. Geophysical inversion involves mapping the physical structure and properties of the subsurface of the Earth using measurements made on the surface and possibly constrained by well log measurements. The first step is de-convelving the trace, the result being the reflectivity series. This reflectivity series is then displayed side-by-side as a set of pseudo-acoustic logs, which we can then interpret as a cross-section of the subsurface in terms of its acoustic impedance distribution. Basically, inversion is the process of extracting, from the seismic data, the underlying geology which gave that particular seismic response. Traditionally, inversion has been applied to post-stack seismic data, with the aim of extracting acoustic impedance volumes. However, we used inversion results to directly predict lithology parameters such as porosity and water saturation. We can estimate from seismic numerous proprieties such as: the lithofacies, the porosity, depth, age, diagenesis, pressure, fluid type (oil, gas or water) etc. The main benefit, amongst many others, of seismic inversion is that it improves exploration and reservoir management success, producing more hydrocarbons with fewer, more highly productive wells. On the other hand, geostatistics provides a toolbox for the geologist to use in analysing data and transferring such analysis and interpretation to the task of reservoir forecasting. Another very important benefit of geostatistical methods is the availability to assess uncertainty associated with kriging and cokringing using stochastic methods. The main conclusion of this paper is that the interpretation of secondary data sets, such as Acoustic Impedance from Seismic data, can significantly reduce inter-well estimation uncertainty and that this method is suitable for reservoir characterization and similar studies in petroleum engineering. Also, by combing seismic inversion with geostatistics leads to significant results. The Kringing model is smooth and we have observed that for the studied area Kringing with external drift (KED) is more accurate that cokringing when estimating the depth of the horizon when there is insufficient data. In conclusion, this new method provides a reliable model of reservoir characterization.