Design, implement, and evaluate a 3D seismic attenuation model for sedimentary basins

Thesis Abstract

Designing, implementing, and evaluating a three-dimensional seismic attenuation model for sedimentary basins addresses the persistent challenge of accurately quantifying Q-factor variations and frequency-dependent attenuation in complex basin stratigraphy, which directly impact amplitude preservation, reservoir characterization, and seismic risk assessment. The study aims to develop a robust 3D attenuation model, implement it within a computational framework, and evaluate its performance against conventional homogeneous-baseline models. Specific objectives include (1) to characterize spatial variability of intrinsic and scattering attenuation from borehole and surface seismic data; (2) to formulate a parameterized 3D attenuation model integrating intrinsic Q, scattering effects, and anisotropy using a layered-basement geology approach; (3) to implement the model in a high-performance computing environment with parallelized forward and inverse routines; (4) to calibrate the model against field dataset from a mature deltaic basin comprising 250 km2 of 3D seismic coverage and 60 boreholes, and (5) to assess improvements in seismic attribute accuracy, depth imaging, and reservoir property inference relative to traditional 1D/2D attenuation treatments. The study adopts a realist epistemology within a pragmatic research design, drawing on the Bayesian inference framework to quantify uncertainty in attenuation estimates and leveraging the theory of seismic wave propagation in heterogeneous media (including intrinsic absorption and scattering) as foundational underpinnings. A multiphase mixed-methods approach is employed, combining quantitative seismic analysis with qualitative interpretation of geological controls on attenuation. The population comprises seismic survey data and borehole logs from the study basin, with a sample of 1,200 km of 3D seismic lines, 340 vertical seismic profile points, and 60 lithology-controlled boreholes. Data collection instruments include pre-stack and post-stack seismic volumes, interval velocity and quality factor logs, and core-derived mineralogical and porosity data. Instrument calibration uses standard resources from industry-accepted processing sequences, including deconvolution, velocity model building, and amplitude calibration, complemented by rock physics constraints to link attenuation parameters to porosity, fluid saturation, and mineralogy. Validity and reliability are ensured through cross-validation with independent borehole Q estimates and repeatability tests on multiple survey vintages. The analytical framework comprises (i) stochastic inversion to recover spatially varying Q, scattering coefficients, and anelliptic anisotropy parameters; (ii) a 3D finite-difference time-domain (FDTD) solver for forward modelling of attenuating media; (iii) a hierarchical Bayesian model to fuse priors from rock physics with field measurements; (iv) regression analysis to identify relationships between attenuation parameters and lithofacies, grain size, and fluid content; and (v) uncertainty quantification via Markov Chain Monte Carlo (MCMC) sampling, with convergence diagnostics including Gelman-Rubin statistics. Key expected findings include (a) a geologically consistent 3D attenuation field revealing stronger intrinsic attenuation in clay-rich sequences and pronounced scattering in faulted or fractured zones; (b) demonstrable reduction in amplitude distortion and improved depth-imaging accuracy when applying the 3D model compared with baseline homogeneous-Q assumptions; (c) quantified reductions in reservoir property estimation errors (e.g., porosity and permeability proxies) by up to 15–20% in blind-test regions; and (d) robust uncertainty maps that delineate confidence intervals for Q and scattering parameters across the basin. The study contributes to knowledge by integrating rock physics-informed priors with a scalable 3D attenuation inversion framework, advancing understanding of attenuation mechanisms in complex sedimentary basins, and delivering a practical tool for improved seismic interpretation and reservoir characterization. The conclusion emphasizes the critical role of incorporating both intrinsic and scattering attenuation in 3D, geologically informed models and recommends routine integration of basin-scale attenuation modelling into seismic workflows, along with continued development of parallelized algorithms to handle large-volume data efficiently.