Investigations of fluid-strain interaction using Plate Boundary Observatory borehole data.

摘要

The aim of this thesis is to explore poroelastic properties near the surface of crust, on and off fault zones, using pore-pressure and strain data from the borehole strainmeter (BSM) network operated by the Plate Boundary Observatory (PBO) component of Earthscope. In particular, fluid diffusivities provide critical insight into the relationship between fluid-pressures and strain, and we use pore-pressure and borehole strain (BSM) data from the PBO to better understand this relationship. In this thesis we focus on PBO stations in the Anza region of the active San Jacinto fault, which has a seismic slip gap (and observed seismicity gap), and accommodates significant plate boundary strain. Anza is also well-instrumented: it ranks second in density of broadband seismometers, next to Parkfield in central California.;Chapter 1 outlines the PBO borehole instrumentation, and offers some requisite background information. The remaining chapters of this thesis can be grouped into two basic categories: (I) characterization of the instruments (Chapters 2, 3, and 4), and (II) focused studies of fluid-strain interaction (Chapters 5 and 6).;Chapter 2 establishes the statistics of noise in the seismic band for data from PBO borehole strainmeters BSMs and seismometers, including cyclic (daily, weekly, or annual) variations, since a full statistical description is important in finding anomalous behaviors, and in characterizing physical sources themselves. This statistical description has been reported in Barbour and Agnew (2011); a related paper (Barbour and Parker, 2014) documents the primary spectral-analysis tool used for this characterization (reprinted in Appendix A).;Chapter 3 builds upon Chapter 2: we used data from the same instruments to compare the seismic-wave detection capabilities of the different sensors. To make this comparison, we use the instruments' noise spectra to determine the relative signal-to-noise ratio on different sensors, as a function of the phase velocity and frequency of a signal. The BSM is less sensitive to seismic waves than surficial broadband instruments are, but more sensitive than colocated short-period geophones in the surface-wave frequency band. We reported these results in Barbour and Agnew (2012).;Chapter 4 is a systematic study of the nature of apparent coseismic strains observed by the BSM using a probabilistic detection method. Rarely do the observed strains agree with predictions based on elastic dislocation modeling, and independent observations from longbase strainmeters. Surprisingly, we find no statistical evidence suggesting the effect is controlled by seismic energy density, or poroelastic effects, which suggests a localized effect or an instrumental hysteresis.;Chapter 5 presents the results from a semi-controlled experiment in fluid-extraction. We collected multiple years of water-well pump activity near a pair of PBO strainmeters, and find remarkable agreement with the calculated extraction volumes, and the strain and pore-pressure observations. We are able to fit the borehole observations by simulating withdrawl from a poroelastic halfspace with relatively high values of hydraulic diffusivity, and low values of elastic shear modulus. Chapter 6 is a systematic study of the pore-pressure response to seismic waves. We find strong correlations between strain and pressure at each station, and in southern California we observe a clear reduction in effect-size (scaling) at stations near the San Jacinto fault (compared to much further away). We show that this reduction is directly linked to crustal shear strain rates.