Covering the risks

O&G. Why are geostatistics and seismic growing in importance in reservoir modeling?
EA. The ideal reservoir model is highly-detailed and matches production history. As plays become increasingly complex, it is critical that reservoir models move closer to that ideal. 3D Seismic data significantly improves lateral detail. Geostatistics generates statistically correct models that can be tested against history. The combination of seismic and geostatistics – geostatistical inversion – yields a highly-detailed and more accurate model.

O&G. How does adding full 3D seismic data improve geostatistical models?
EA. Geostatistical modeling has become the standard for analyzing and integrating data from various sources with different scales and for increasing the reliability and objectivity of uncertainty estimates. Early approaches collapsed seismic to 2D maps, leaving uncertainties inadequately quantified. Now full 3D seismic data volumes can be used to constrain simulations. By combining vertically dense well measurements with laterally dense seismic, highly-detailed models are constructed. New models combining geostatistics and 3D seismic volume data are accurate near and away from wells, have realistic detail beyond the seismic bandwidth, retain integrity even when wells are removed, and provide more accurate estimates of uncertainty and bias.

O&G. How does this technique affect uncertainty and bias?
EA. There will always be uncertainty and potential bias. The goal is to quantify the uncertainty and reduce bias as much as possible. For example, well data is almost always preferentially clustered in higher pay areas, introducing redundant data that must be handled. Full integration of well and 3D seismic data through geostatistical inversion produces far more quantitative and geologically-plausible models. Geostatistical inversion techniques are also well-suited for incorporating qualitative or ‘soft’ data, such as expert insight into the physical processes underlying the spatial distribution of the rock layers, without over-constraining the solution. Confidence levels and specialized algorithms combine in a probabilistic technique to balance and weight the importance of all inputs. The resulting numerical models integrate soft and hard data in the most unbiased and realistic form.

O&G. Is this practical for typical E&P operations?
EA. Any operation would benefit from having a highly-detailed, accurate model – the foundation for better drilling decisions, future well placement, and clearer risk assessment. For 20 years, Fugro-Jason has provided software and services that assist clients in creating accurate models. The latest software combines geostatistics and advanced statistical physics with innovative seismic inversion methods, producing reservoir models that reliably quantify uncertainty. The workflow is an expanding loop with built-in QCs that produces realizations which are then used in co-simulation to identify highly-detailed, realistic petrophysical properties such as porosity and water saturation. Based on this technique, one company working a carbonate reservoir switched from grid-based to targeted drilling and found the highest porosity ever seen in the field.

O&G. How can different grids/scales for seismic data and geological models be combined?
EA. To bring well-based models and seismic data together requires the correlation of a corner point grid (CPG) and seismic grid into a new stratigraphic model grid (SMG). Zone based adjustment algorithms manage the transform from the seismic grid into the SMG, and ensure structural and stratigraphic integrity when transformed from the SMG into the CPG. This includes management of detailed structural models with many faults and sub-parasequences. In this way, detailed Geostatistical Inversion results are reliably transferred into the geological domain, enabling better drilling decisions and reserves assessment. Individual realizations can be easily transferred to a flow simulator and compared to actual production data to determine which equi-probable reservoir model is the most likely.

Adams is Managing Director of Fugro-Jason, a leading provider of seismic inversion and reservoir characterization products and services. He joined the company in 1996 as Region Manager, North and South America. Adams has 30 years of experience in the industry, including previously work experience at Schlumberger and Veritas.