Characterizing urban seismicity using a dense array in Long Beach, CA

Microearthquakes only produce detectible ground motion over a small area and are often overlooked by sparse seismic networks. However, these earthquakes illuminate active faults and provide data for seismic hazard assessment. This insight from small earthquakes is especially important in urban areas underlain by fault zones, but anthropogenic noise lowers the signal-to-noise ratio and makes it difficult to identify microseismicity in these areas. The goal of this study is to find systematic ways to discriminate small tectonic events from the ambient urban background noise with a large seismic data set. We use 6 months of data from a dense array of over 5200 sensors in a 7-by-10-km area of Long Beach, CA. Several active faults traverse the city, including the Newport-Inglewood Fault Zone. It also has many sources of anthropogenic noise, including two major freeways, a railroad, port, commercial airport, and producing oilfields. Additionally, during the study, active source surveys were performed by the company that deployed the sensors. Using a standard short-term over long-term event detector and then associating the individual picks into events, we find over 4000 events. There is a very strong correlation between the number of events and the time of day and day of the week. This temporal variation suggests a significant human influence on the detectability of seismic activity. Additionally, some of the detected events are not tectonic but direct human activity. The most prevalent examples of this are the active source surveys. We try several methods to separate earthquakes from man-made noise. We can use the location and timing of individual picks spread across the array to determine if their relationship is plausibly caused by an earthquake. Another method overcomes the poor signal-to-noise ratio by stacking data from multiple sub-arrays and correlating the wavelet transform of these stacks. Earthquakes are identified with a characteristic spectral decay that is not present in man-made noise. We also perform continuous analysis of the array stacks by back-projecting them onto a volume beneath the city.