AAPG Annual Convention and Exhibition

Datapages, Inc.Print this page

Understanding the Correlation Between Induced Seismicity and Water Injection in the Fort Worth Basin

Abstract

Starting in the mid-2000s, there has been an increase in seismic activity around areas where fluid injection was expanding because of shale development. As the injection rate increased, so did occurrences of earthquakes in the surrounding the area. Extensive studies have been done regarding the correlation between injection wells and induced seismicity (Frohlich 2012, Davis 1995). However, many injectors don't cause earthquakes, and the boundaries between safe and high risk practice have yet to be defined. Also, there is often a time lag between the onset of injection and the occurrence of seismic activities – what controls that timescale? In order to encompass areas of injection with and without seismic activity, a reservoir simulation model was built for most of the Fort Worth Basin (FWB), including 374 wells with available relevant data located in the following counties: Denton, Ellis, Erath, Hill, Hood, Jack, Johnson, Palo Pinto, Parker, Somervell, Tarrant and Wise. The data needed for the simulation include minimum and maximum injection depths, monthly injection pressures and monthly volumes. The locations of major faults in the basin are being worked into the model to include the effects of transsmissive versus sealing faults on flow patterns. Preliminary simulation results show that where earthquakes occur, there is some spatial correlation with injection well locations. Furthermore, the modeling shows the quake areas have substantial increases in pore pressure due to injection. However, not all areas of increased pore pressure have induced earthquakes. Preliminary analysis suggests absence of induced seismicity in areas of elevated pore pressure might be attributed to shallow depth and lack of large pre-existing faults. Similar correlation difficulties are seen between the timing of injection and earthquake occurrence. The distance between the injector and the fault as well as the permeability-thickness (kh) of the injection formation are being investigated as controls on this time lag effect.