Seismic Data Conditioning for Quantitative Interpretation of Unconventional Reservoirs

Abstract

Shale resource plays are fairly new to the petroleum industry, but they have reinvigorated oil and gas production in North America. Brittleness and TOC are the two most important parameters for shale resource characterization. Ideally, of the multilinear and non-linear regression can be used to correlate TOC and brittleness measured on core to well logs forming a proxy for TOC and brittleness with in the seismic survey. In turn seismic attributes correlated to TOC and brittleness predictions from well logs. The success of such integration depends on data quality. In Texas and the mid-continent much of our seismic data have been merged and reprocessed using modern technology. I will expose one pitfall on merged seismic surveys due to offset range variation. Other pitfalls are best addressed by seismic modeling. Legacy seismic data acquired in the mid-continent region have low fold, resulting in a rise to low signal to noise ratio. Such data often exhibit a strong acquisition footprint, which can be caused by the presence of aliased ground roll. Conventional processing techniques cannot suppress such groundroll without damaging the signal. I developed and applied a coherence-based technique to remove highly aliased ground roll present in a survey of North Central Texas Mississippi Lime play. The predicted TOC and brittleness volumes showed a fair correlation with production in the Barnett Shale of Fort Worth Basin. The areas of good production are associated with high brittleness in the vicinity of high TOC.