Identifying Areas at Risk for Injection-Induced Seismicity Through Subsurface Analysis: An Example From Southern Kansas

Abstract

The U.S. midcontinent has experienced an unprecedented increase in seismicity linked to large-scale injection of wastewater. A major challenge for stakeholders is that subsurface faults and stresses are not well known, making it difficult to properly site new disposal wells or make decisions regarding operations of existing wells. This study focuses on using publicly available well data to map subsurface faults and lineaments and present-day stresses in southern Kansas, with the ultimate goal of identifying specific structures and regions that may be at higher risk for injection-induced seismicity. A database of >500,000 stratigraphic well tops from the Kansas Geological Survey was used to create new structure contour maps of 18 stratigraphic boundaries between the Precambrian basement and present-day surface. Most of the boundaries have between 10,000 to 40,000 well picks associated with them; however, the spatial density of the picks is variable. Outliers in the tops data were visually identified and removed, including from high data density areas where outliers were often subtle. Structure contour maps were created using kriging, convergent interpolation, moving average, and minimum curvature data interpolation methods. These maps were then compared against each other and data confidence maps to select the most reasonable data interpolation technique, convergent interpolation. A range of 3D surface analysis techniques were then employed to tease out discontinuities and lineaments within each unit, including slope, aspect, curvature, edge detection, and residual analysis. The mapped lineaments were compared to known faults and lineaments from analysis of 3D seismic data, potential field data, satellite data, LiDAR data, and air photos. These analyses show two groups of structures across the region. NNE-trending structures are likely related to the Nemaha Ridge-Humboldt fault zone and earlier Midcontinent Rift System, while NW-trending structures are probably associated with the Central Kansas Uplift. Subsurface stresses were examined using available image, caliper, and dipmeter logs from a large inventory of well data across the state and through inversion of earthquake moment tensor solutions, where available. The findings are being used to assess the reactivation potential of the mapped faults, and will be evaluated in conjunction with brine disposal data, to identify areas at risk for injection-induced earthquakes.

AAPG Datapages/Search and Discovery Article #90291 ©2017 AAPG Annual Convention and Exhibition, Houston, Texas, April 2-5, 2017