Comparative Seismic Attenuation in Crustal Media Across Regions

Thesis Abstract

Seismic attenuation variability in crustal media across regions influences ground motion prediction and seismic hazard assessments, yet comparative analyses that systematically quantify regional differences and their drivers remain limited. The study addresses the problem of inconsistent attenuation characterizations across geologically diverse crustal environments, which hampers the transferability of regional attenuation models to broader seismotectonic settings. The objective is to quantify regional variations in seismic Q, frequency-dependent attenuation, and anelastic strain energy dissipation, and to identify governing crustal parameters (temperature, composition, fracturing, porosity, pore-fluid effects) that explain observed differences. Specific objectives include (i) estimating site- and region-specific quality factors (Q) and their frequency dependence (Qf) from regional earthquake catalogs and borehole/ambient noise data; (ii) comparing attenuation proxies derived from P- and S-wave amplitudes, spectral decay, and coda wave methods across four tectonically distinct regions; (iii) examining the relationship between attenuation metrics and crustal properties using regression and multivariate analysis; (iv) testing the applicability of established attenuation theories (Anderson–Hildreth scattering, Biot poroelastic attenuation, and the Temperature-Dependent Q model) in each region; and (v) synthesizing regionally robust attenuation models suitable for regional hazard assessment. The methodology adopts a comparative cross-sectional design grounded in seismological data. The population comprises crustal seismic events (magnitude Mw ? 4.0) recorded over a 15-year window by dense regional networks in four geologically diverse regions. A stratified sampling framework selects approximately 350 events per region, ensuring uniform coverage of depth, focal mechanism, and back-azimuthal distribution. Data collection integrates high-quality waveform records from regional seismograph arrays, borehole instrumentation data where available, and ambient noise spectra. Instruments include broadband seismometers, strong-motion sensors, and microtremor arrays, coupled with published crustal property datasets (temperature gradients, mineralogy, porosity, fluid content). Attenuation parameters are estimated through multiple, cross-validated methods (a) spectral ratio and amplitude decay analyses to derive Qf across frequency bands (0.5–20 Hz), (b) coda wave analysis to extract Qc and its frequency dependence, and (c) scattering attenuation models to separate intrinsic and scattering losses. Statistical analyses utilize hierarchical linear models to account for regional nesting, generalized additive models for nonlinear effects, and multivariate regression to link attenuation metrics to crustal properties. Theoretical framing incorporates three named theories Biot’s poroelastic attenuation for fluid–solid interactions, fractal scattering theory for heterogeneous media, and the temperature-dependent Q model to capture mantle-crustal transitions, enabling comparison of their explanatory power across regions. Model validation employs bootstrap resampling, sensitivity analyses, and out-of-sample prediction. Anticipated findings include statistically significant regional differences in Qf and Qc, with higher intrinsic attenuation in regions with higher temperatures and pervasive porosity, and greater scattering attenuation in regions with pronounced fracturing and heterogeneity. The study expects that regression models will reveal crustal temperature, mineralogical composition, porosity, and fluid saturation as key predictors of attenuation variations, with regional coefficients reflecting local tectonic history. The research contributes to knowledge by providing a unified, cross-regional framework for crustal attenuation that reconciles disparate regional inventories, clarifies the dominance of intrinsic versus scattering mechanisms in different tectonic settings, and yields region-specific parameterizations for ground-motion models. The findings offer improved inputs for seismic hazard analyses, including regional ground-motion prediction equations and scenario simulations under varying crustal conditions. The main conclusion posits that crustal attenuation is regionally modulated by a combination of thermal state, poroelastic properties, and heterogeneity, in which no single mechanism universally dominates. Recommendations include integrating regionally calibrated attenuation models into seismic design codes, expanding borehole and ambient-noise campaigns to reduce parameter uncertainty, and pursuing targeted laboratory experiments to constrain Biot-porous and scattering parameters under crustal conditions.