An Innovative Method For Detecting Subtle Events And Patterns In Seismic Data

Abstract

Seismic surveys are a preferred method of exploration of the subsurface. Modern sensitive electronic sensors, combined with modern high speed computers and signal processing techniques allow for a high degree of resolution. However, there is still much more for the data to reveal hidden within the noise. Here we apply an automated template matching method to resolve subtle patterns from data that would otherwise fall below the typical threshold resolution. A template event or pattern is provided and cross correlated against the data stream of interest. The cross correlations for multiple channels and stations can then be stacked, further increasing the resolution of the method. The method as presented is largely automated and provides a natural match with artificial intelligence and soft computing techniques for near real time reservoir monitoring and characterization. Likewise, the approach can serve to enhance resolution for pre-existing data. In the case example provided, a potential induced seismic sequence in an area of sparse seismometer resolution is investigated. From a template event of interest, we use a matched filter approach to detect microearthquakes with similar waveform characteristics. Through multichannel and multistation stacking, very small magnitude events are detected in an automated fashion. In this way, we can track fractures as they propagate through the formation. This approach does not require additional temporary and more spatially dense sensor deployments, and thus reduces the time and cost of traditional surveys. In fact, we use only publicly accessible waveform data from existing seismic networks to achieve high resolution and catalog completeness without the need for spatially dense deployments.

AAPG Datapages/Search and Discovery © 2014 Pacific Section AAPG, SPE and SEPM Joint Technical Conference, Bakersfield, California, April 27-30, 2014