Narrations of lecture material as well as research presentations can be found on Jef Caers YouTube Channel
GS 240: Data science for geoscience
Overview of some of the most important data science methods (statistics, machine learning & computer vision) relevant for geological sciences, as well as other fields in the Earth Sciences. Areas covered are: extreme value statistics for predicting rare events; compositional data analysis for geochemistry; multivariate analysis for designing data & computer experiments; probabilistic aggregation of evidence for spatial mapping; functional data analysis for multivariate environmental datasets, spatial regression and modeling spatial uncertainty with covariate information (geostatistics). Identification & learning of geo-objects with computer vision. Focus on practicality rather than theory. Matlab exercises on realistic data problems.
Terms: Win | Units: 3 | Grading: Letter (ABCD/NP)
GS 260: Quantifying Uncertainty in Subsurface Systems
Broad conceptual overview of the various components required to uncertainty quantification (UQ) for decision making in subsurface engineering problems such as oil/gas production, groundwater management, contaminant remediation, geothermal energy and mineral deposits. The emphasis lies on learning how to synthesize rather than the details of each individual discipline. The class will cover the basic data science for UQ: dimension reduction methods, Monte Carlo & global sensitivity analysis. Introduction to Bayesianism and how it applies to subsurface prediction problems, in particular, the formulation of geological prior models and the role of geostatistics. Strategies for integrating geological science, geophysics, data science and decision science into decision making under uncertainty. Team work on real field applications.
Terms: Spr | Units: 3 | Grading: Letter or Credit/No Credit
ENERGY 241: Seismic Reservoir Characterization (ENERGY 141)
Practical methods for quantitative characterization and uncertainty assessment of subsurface reservoir models integrating well-log and seismic data. Multidisciplinary combination of rock-physics, seismic attributes, sedimentological information and spatial statistical modeling techniques. Student teams build reservoir models using limited well data and seismic attributes typically available in practice, comparing alternative approaches. Software provided (SGEMS, Petrel, Matlab). Offered every other year.
Terms: Spr | Units: 3-4 | Grading: Letter (ABCD/NP)
ENERGY 246: Reservoir Characterization and Flow Modeling with Outcrop Data (ENERGY 146)
Course gives an overview of concepts from geology and geophysics relevant for building subsurface reservoir models. Includes a required 1-day field trip and hands-on lab exercises. Target audience: MS and 1st year PhD students in PE/ERE/GS with little or no background in geology or geophysics. Topics include: basin and petroleum systems, depositional settings, deformation and diagenesis, introduction to reflection seismic data, rock and fluid property measurements, geostatistics, and flow in porous media.
Terms: Aut | Units: 3 | Grading: Letter (ABCD/NP)
GEOPHYS 112: Exploring Geosciences with MATLAB
How to use MATLAB as a tool for research and technical computing, including 2-D and 3-D visualization features, numerical capabilities, and toolboxes. Practical skills in areas such as data analysis, regressions, optimization, spectral analysis, differential equations, image analysis, computational statistics, and Monte Carlo simulations. Emphasis is on scientific and engineering applications. Offered every year, autumn quarter.
Terms: Aut | Units: 1-3 | Grading: Letter or Credit/No Credit