Deep-Learning Geophysics for Earthquake Forecasting, Subsurface Characterization and Geological Carbon Storage
Year: 2026
Type: reference-entry
Abstract: Deep-learning geophysics connects heterogeneous observations to uncertain processes in the crust and subsurface. This monograph develops a research framework for earthquake forecasting, subsurface characterization and geological carbon storage using only relevant or transferable methods in the supplied source corpus. Ten themes link scale-aware inference, prospective seismic forecasting, inverse uncertainty, storage containment, spatial causality, physical constraints, graph structure, provenance, adaptive monitoring and public accountability. Five chapters progress from earth-system foundations through statistical explanation and deep learning to scalable data operations and applications. The framework distinguishes event detection from future-earthquake probabilities, a plausible subsurface image from a unique geological truth, and measured carbon outcomes from modelled storage expectations. Proposed research outputs include catalog audits, uncertainty budgets, causal ledgers, graph-construction records, physics-constrained surrogates, prospective forecast archives and storage monitoring plans. Validation uses chronological and spatial holdouts, numerical forward comparisons, proper scoring rules and sensitivity to alternative geology. Carbon accounting links subsurface, site and atmospheric observations without treating any one stream as conclusive. The supplied bibliography supports several methods directly, but earthquake forecasting is developed as a research design rather than presented as a result from those works. Applications are framed as staged tests with independent review, response thresholds and locally governed data access. The result is a bounded evidence standard for geophysical prediction and climate-related subsurface decisions. Keywords deep-learning geophysics, earthquake forecasting, seismicity, subsurface characterization, geological carbon storage, inverse problems, seismic waveforms, physics-informed learning, neural operators, graph neural networks, fault connectivity, probabilistic forecasts, spatiotemporal inference, remote sensing, monitoring verification, geospatial data, uncertainty quantification, digital twins, reservoir monitoring, model governance, climate integrity
Author: Murali Krishna Pasupuleti
Institution: National Education Services
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Climate action
Open Access status: gold
OpenAlex ID: W7214533366
Cites:
Cited by:
Related to: