EOSC 454 · Applied Geophysics

Combining physics driven and data driven methods for using geophysical and remote sensing data to solve problems in the geosciences. Case histories related to societal challenges due to climate change, environmental problems, and natural resources motivate the methods covered. This course emphasizes practical application of quantitative methods in numerical modelling, inverse theory, and machine learning; a course project provides the opportunity for students to select a topic of interest to explore in more depth. 

Writing and sharing scientific software is an essential part of any quantitative workflow. We will review fundamentals in the use of Python for scientific programming, and using version control including Git and GitHub for tracking changes and sharing work. Practices that facilitate the reproducibility of computational results are a theme throughout the course.

key concepts 

• numerical modelling: considerations for running numerical simulations including: how to design a mesh for a simulation, understanding boundary conditions and how to test the setup of a numerical simulation 
• inversion: fitting a physics-based model to observed data. Non-uniqueness of the inverse problem, the use of regularization, and avenues for bringing additional data and information into an inversion
• machine learning and advanced topics: use of machine learning for integrating data sets, areas of research in combining physics-driven and data-driven approaches 
• reproducibility and scientific software: writing and sharing code in a manner that enables others to reproduce your work, and best practices for sharing scientific software