Statistical Methods of Predicting Source Rock Organic Richness From Open Hole Logs, Niobrara Formation, Denver Basin, CO
The Niobrara Formation in the Denver Basin is an unconventional oil and gas play composed of alternating chalk and marl units. Key characterization parameters that provide an understanding of the distribution of source potential within the Niobrara include: total organic carbon (TOC), maturity level, mineralogy, thickness and organic matter type. Sample based results combined with full log suites including high resolution density, resistivity, sonic, porosity and spectral gamma ray logs will help in fully characterizing the Niobrara. Two widely used empirical approaches developed to quantitatively estimate TOC from log data are the Schmoker density log technique and the combination sonic or density and resistivity log technique known as Δ log R. Other common methods for TOC estimation include using uranium or gamma ray logs as indicators of organic matter, although they often require a local calibration. This study compares the different methods of quantifying TOC from logs and how they apply to the Niobrara formation. Each method was analyzed and tested on the Niobrara Formation. The methods were then modified by recalibrating to core TOC data and applying new empirical relationships. Qualitative log indicators of elevated TOC include elevated neutron porosity, low bulk density, high sonic transit time, and high gamma ray or uranium. However, these measurements respond to more than just organic matter and therefore, interpretation of these logs in terms of organic matter requires accounting for the effect of mineralogy and fluids on log signatures in the Niobrara as well. The modified equations were used to quantify TOC in the Niobrara and a method was developed to classify organic-rich (high TOC) and organic-lean (low TOC) facies in an ordered manner. Statistical tools such as neural networks and decision tree analysis were then utilized to identify the best open hole log combinations that would be most predictive of those facies in wells that lack core TOC data.
AAPG Datapages/Search and Discovery Article #90216 ©2015 AAPG Annual Convention and Exhibition, Denver, CO., May 31 - June 3, 2015