Thermal-Maturity and Expulsion-Retention Modeling in the Utica Shale Play

Steven G. Crews and John M. Guthrie
Hess E&P Technology, Houston

A regional-scale 3D petroleum systems model was constructed to help predict maturity and hydrocarbon fluid properties in Ordovician (Pt. Pleasant) and Devonian (Marcellus) unconventional reservoirs in eastern Ohio and western Pennsylvania. The model covers an area of approximately 146,000 km² (36 million acres) and comprises 10 stratigraphic surfaces, from the present-day topography to the top of the Precambrian. These structure maps were based largely on well control, along with some published maps and public elevation data.

Two distinct levels of results emerged from the study: (1) the first-order results are source rock maturity and hydrocarbon generation history; (2) the second order results are related to specific fluid properties of the retained hydrocarbons; these results were obtained by fine-tuning the expulsion model to calibrate GOR and condensate-yield (CGR) predictions to production data. In essence, determination of source rock maturity is a necessary but not sufficient condition for predicting fluid properties such as CGR in the wet gas parts of the play and GOR in the oil-dominated region. More details on these results follow:

(1) Maturity and hydrocarbon-generation history. The prediction of source rock maturity was obtained by simulating the burial and thermal history and coupling it with a source-rock description comprising kinetics, richness and thickness. The key variable for this part of the study was the amount of erosion on the surface unconformity that characterizes the region. Bedrock strata that outcrop or are thinly covered in our study area range from Silurian to Upper Pennsylvanian, indicating a long gap in stratigraphic record. Moreover, the rock exposed at the surface is compacted and thermally mature, pointing to significant paleo burial. In particular, upper Pennsylvanian coals in the study area yield Ro values of 0.59-0.96%. We built the model initially with wedge of missing section whose shape and thickness were based on the preserved portion of the Appalachian basin; we then iteratively modified the shape and thickness of the missing section wedge until we achieved a good match between predicted and observed values for several calibration data sets. These data sets included: (1) surface coal vitrinite reflectance; (2) thermal maturity derived from aromatic biomarkers of oil from the overlying Clinton Sandstone conventional reservoir; (3) present-day hydrogen index values in the Point Pleasant; and (4) measured vitrinite reflectance in the Marcellus formation in Pennsylvania. These data sets only partially overlap in space; the Marcellus data set was used to extend our model eastward into Pennsylvania and West Virginia. It required modifications to the shape of our missing section “wedge”. The maximum estimated thickness of missing section in the dry gas portion of the play exceeds 10,000 feet.

(2) Fluid-property predictions. Predictions of GOR/CGR were obtained by combining the source rock maturity and generation history results with an expulsion model that distinguishes expelled from retained hydrocarbons and predicts the composition of each. Our expulsion-retention simulation considers sorption and pore saturation thresholds, inorganic (matrix) and organic (kerogen) porosity, cracking kinetics and other factors. However, all of these parameters are difficult to constrain a priori. We therefore take an empirical approach, iteratively adjusting them to calibrate the GOR predictions of our model to observed gas and liquids production data. Our initial, very small, calibration data set was easy to match; however, the wave of new well results that became available in 2013 revealed a large amount of scatter in the data. We have recalibrated to match the central tendency of the data cloud.

The integration of a calibrated burial/thermal history model with a separately calibrated expulsion-retention model thus allows us to model both the maturity and the expulsion-retention history of the source rock; and thereby predict the basic composition of fluids retained. The “bottom-up” nature of the underlying maturity model enabled us to make reasonable predictions (specifically regional CGR/GOR maps) during the early appraisal stage of the Point Pleasant play.

AAPG Search and Discovery Article #90186 © AAPG Geoscience Technology Workshop, Hydrocarbon Charge Considerations in Liquid-Rich Unconventional Petroleum Systems, November 5, 2013, Vancouver, BC, Canada