Cutting Costs by Locating High Production Wells: A Test of the Volcano seismic Approach to Finding ''Blind'' Resources

Shalev, Eylon ; Malin, Peter E. ; McCausland, Wendy

Office of Research Support Duke University (United States)

关键词: Amplification;    Power Generation;    Microearthquakes;    Attenuation;    15 Geothermal Energy;   

DOI  :  10.2172/796234
RP-ID  :  DOE/ID/13863
RP-ID  :  FG07-00ID13863
RP-ID  :  796234

美国|英语

来源: UNT Digital Library

【 摘 要 】

In the summer of 2000, Duke University and the Kenyan power generation company, KenGen, conducted a microearthquake monitoring experiment at Longonot volcano in Kenya. Longonot is one of several major late Quaternary trachyte volcanoes in the Kenya Rift. They study was aimed at developing seismic methods for locating buried hydrothermal areas in the Rift on the basis of their microearthquake activity and wave propagation effects. A comparison of microearthquake records from 4.5 Hz, 2 Hz, and broadband seismometers revealed strong high-frequency site and wave-propagation effects. The lower frequency seismometers were needed to detect and record individual phases. Two-dozen 3-component 2- Hz L22 seismographs and PASSCAL loggers were then distributed around Longonot. Recordings from this network located one seismically active area on Longonot's southwest flank. The events from this area were emergent, shallow (<3 km), small (M<1), and spatially restricted. Evidently, the hydrothermal system in this area is not currently very extensive or active. To establish the nature of the site effects, the data were analyzed using three spectral techniques that reduce source effects. The data were also compared to a simple forward model. The results show that, in certain frequency ranges, the technique of dividing the horizontal motion by the vertical motion (H/V) to remove the source fails because of non-uniform vertical amplification. Outside these frequencies, the three methods resolve the same, dominant, harmonic frequencies at a given site. In a few cases, the spectra can be fit with forward models containing low velocity surface layers. The analysis suggests that the emergent, low frequency character of the microearthquake signals is due to attenuation and scattering in the near surface ash deposits.