Extracting Coseismic Signals from Groundwater Level
- PDF / 685,370 Bytes
- 19 Pages / 439.37 x 666.142 pts Page_size
- 38 Downloads / 195 Views
Extracting Coseismic Signals from Groundwater Level Ting Wang · Mark Bebbington · David Harte
Received: 16 December 2010 / Accepted: 5 July 2011 / Published online: 20 September 2011 © International Association for Mathematical Geosciences 2011
Abstract There has been no statistical examination of the possible link between earthquakes and the transient hydrological response of a well with potentially hundreds of responses. We propose an algorithm for automatically detecting anomalous transient changes in groundwater level (signals) based on a moving average of the variance of the first differences in water level. This is tested on 4 years of groundwater level data sampled at 1-minute intervals (approximately 2 million data) from the Tangshan Well, ∼100 km southeast of Beijing, using the global catalog of earthquakes of minimum magnitude 6.0 during the same period (600 events). We investigate the relationship between the extracted signals and the arrival times of various seismic waves, in particular, the earliest P phase, S phase, Love wave, and Rayleigh wave arrivals. The detection probability is estimated as a function of the earthquake characteristics magnitude, well-epicenter distance, depth, and azimuth from the well. Keywords Point process · Mutual information · Regression · Groundwater level · Statistical seismology
1 Introduction Observed hydrological anomalies in response to large earthquakes include sustained changes (or steps) in groundwater levels in wells, and groundwater level oscillations (Montgomery and Manga 2003 and references therein). Physical mechanisms T. Wang () · M. Bebbington Volcanic Risk Solutions, Massey University, Private Bag 11222, Palmerston North 4442, New Zealand e-mail: [email protected] D. Harte Statistics Research Associates, PO Box 12 649, Wellington 6144, New Zealand
800
Math Geosci (2011) 43:799–817
proposed to explain such coseismic changes include suggestion of strong ground motion induced dilatancy (Bower and Heaton 1978); poroelastic response to volumetric strain (Roeloffs et al. 1989); local permeability changes near the wells (Rojstaczer et al. 1995; King et al. 1999; Brodsky et al. 2003); and mobilization of gas bubbles (Roeloffs 1998). While sustained changes in groundwater level appear to be rare (Matsumoto 1992; Kitagawa and Matsumoto 1996; Matsumoto et al. 2003), transient oscillations of groundwater level are considered much more common. The amplitude, and hence detection, of groundwater oscillations appears to be a function of the characteristics of the well-aquifer system (Cooper et al. 1965; Kunugi et al. 2000). Earthquakes generate two kinds of seismic waves; those that travel through the solid body of the earth, such as P- and S-waves, and slower (surface) waves that travel along the surface of the earth, such as Rayleigh and Love waves. Love waves are normally faster than Rayleigh waves, and P-waves are the fastest of all. Body waves travel through multiple layers of the earth, generating different phases of the wave that arrive at different times, de
Data Loading...
