GSA Annual Meeting in Indianapolis, Indiana, USA - 2018

Paper No. 208-14
Presentation Time: 5:15 PM


HOLLENBACH, Andrew Michael, BIDGOLI, Tandis S., ANSARI, Esmail and NOLTE, Alex, Kansas Geological Survey, University of Kansas, Lawrence, KS 66047

The inferred cause of a more than 40-fold increase of seismicity in Kansas and Oklahoma is increased brine disposal into the Cambrian-Ordovician Arbuckle Group aquifer. Despite numerous studies illustrating the correlation between injection activity and seismicity in the region, a simple remedial tool has remained elusive. Here we investigate spatial and temporal relationships between injection and seismic activity and seismicity mitigation strategies through geostatistical analysis in R™. The analysis uses well locations and industry-reported injection-volumes for 4415 wells Class I and Class II Arbuckle injection wells in Kansas and Oklahoma during the period of increased injection/seismicity from 2010 through 2016. We also use a catalog of 4215 USGS-reported earthquakes over the same timeframe (de-clustered for this study). A kernel density estimation method is used to represent densities of discrete and non-continuous point data (injection volumes at well locations and earthquake epicenters) as a smoothed continuous contour. The method estimates the spatially-averaged density of any given variable V, at any given location. A grid lattice is configured across the study area to create n points i1…in, and a circle with radius r around each point defines an area. At each i1 through in the sum value of injection volumes, the number of wells, and the number of earthquakes are counted, summed, and averaged within the circular area. The lattice spacing and radius of investigation can be adjusted to investigate the robustness of the spatial relationships, with a coarser lattice mesh and large radius smoothing results. Preliminary results suggest that there is a spatio-temporally constrained volumetric threshold for inducing earthquakes. Several areas are observed that remain in excess of the threshold and effective earthquake mitigation can be achieved by reducing injection volume density through reduction of Arbuckle disposal volumes and addition of lower volume Arbuckle disposal wells. The simple but elegant method of analysis offers a new approach for evaluating the relationship between injection volumes, well density, and earthquake counts, over time that may be of considerable value to well operators and regulators working on the mitigation of induced seismicity.