Minimal Clustering of Injection-Induced Earthquakes Observed with a Large-n Seismic Array

Abstract

ABSTRACT The clustering behavior of injection-induced earthquakes is examined using one month of data recorded by the LArge-n Seismic Survey in Oklahoma (LASSO) array. The 1829-node seismic array was deployed in a 25  km×32  km area of active saltwater disposal in northern Oklahoma between 14 April and 10 May 2016. Injection rates in the study area are nearly constant around the time of the deployment. We develop a local magnitude (ML) equation for the region and estimate magnitudes for 1104 earthquakes recorded by the deployment. The determined earthquake magnitudes range from ML 0.01 to 3.0. The majority of earthquakes occurred between 1.5 and 5.5 km depth, and the shallowest earthquake depths overlap with the base of injection wells at depths between 1.5 and 2.5 km. We compute focal mechanisms of the largest events (ML>2.0), and find a mix of normal- and strike-slip-faulting types. Earthquakes occur regularly in time during the deployment, but are not evenly distributed in space across the study area, that is, they are spatially clustered. Analysis of the nearest-neighbor distances in the space–time–magnitude domain shows the seismicity is dominated by single-event clusters (i.e., independent events). This high proportion of single-event clusters compared with multievent clusters has been previously noted for induced events at geothermal sites. When clustering occurs, the number of events in a cluster is typically small. We observe only four clusters with 10 or more events. For these larger clusters, we find equivalent numbers of foreshocks and aftershocks; however, the foreshock sequences are significantly longer in duration lasting days to tens of days, while aftershock sequences are observed only on the order of one day. The minimal clustering observed for events in the LASSO array suggests that the majority of events are being directly driven by stress changes due to local saltwater disposal.