B8256 - ANALISI DI SEGNALE E METODI STATISTICI IN GEOFISICA

Learning outcomes

At the end of the course, the student knows and understands the theory of time-series and signal analysis in frequency space in relation to the internal, oceanographic, and atmospheric processes that generated them. The student will be able to create synthetic signals for the validation of physical theories, distinguish deterministic and stochastic signals with advanced computational methods of real data processing and give an interpretation useful to understand the issues of hazard and risk in the Earth System. He/she will also be able to understand the statistical methods that are used for the assessment of geophysical hazard and risk and in particular the maximum likelihood method applied both to the estimation of the parameters of space-time models and to error-in-variable regression techniques. Finally, he/she will know the information-based criteria for the comparison of different models and some techniques for the quantification of qualitative information.

Course contents

Time Series Analysis and Signal Processing are the most common ways of representing data and temporal models in Science. The study of physical processes in scientific/engineering fields cannot ignore these tools and the statistics that describe them, allowing studies at ime scales ranging from microseconds to millions of years.

Signal theory also represents an ideal introduction to statistical methods for hazard estimation that, starting from seismic examples, assume general value in all fields of Earth Physics.

Topics covered.

Time series analysis: time series as a mathematical/physical tool for the study of oscillatory phenomena.

Signal processing: construction of temporal filters and application to time series describing physical processes.

Computational processing: Advanced processing of time series through open access software, advanced correlation theories.

Visualization and interpretation: collaborative coding, visualization in 3D and 4D environment of geophysical data, interpretation of physical data.

Statistical seismology: statistical properties of occurrence, maximum likelihood methods.

Earthquake forecasting: deterministic forecasting, seismic hazard and risk.

It is assumed that the student has a good preliminary knowledge of the basic concepts of mathematical methods for physics.

The course provides the first elements of data processing and computational analysis of signals to students of the Degree Course. The topics covered provide the skills needed to work in a geophysics and applied physics environment, especially in research institutes and companies interested in the assessment of resources and seismic risk.

The course includes computational exercises that will provide the first training of students on the software and programming languages used by the scientific and engineering community to process signals.

Readings/Bibliography

Each lesson is accompanied by the projection of a PowerPoint file. The set of files, divided into chapters, contains a comprehensive exposition of the program and can act as a text for the study of the subject. The files in PowerPoint format are available during the course and can be accessed from the teaching web page.

If they wish to delve deeper into the topics covered in the course, students can consult the following texts available in the Library of the Department of Physics and Astronomy:

- K. Aki and P. G. Richards, Quantitative Seismology, 2nd edition, University Science Books, Sausalito CA, 2002.

- F. A. Dahlen and J. Tromp, Theoretical Global Seismology, Princeton University Press, Princeton NJ, 1998.

- E. Boschi and M. Dragoni, Seismologia, UTET, Turin, 2000.

- T. Lay and T. M. Wallace, Modern Global Seismology, Academic Press, S. Diego CA, 1995

- Stein S., Wysession M., An introduction to seismology, Earthquakes, and Earth structure, Blackwell Pub., 2003.

The following texts are available for consultation upon student request:

- D. Gubbins, Time Series Analysis and Inverse Theory, Cambride University Press, Cambridge UK, 2004.

- H. Igel, Computational Seismology, A Practical Introduction, Oxford University Press Books, Oxford UK, 2017.

Teaching methods

The course is presented in the form of a "frontal lesson" accompanied by visual documentation in PowerPoint.
The course includes computer exercises with programming in Julia, Matlab, and Python; with respect to these methods, active participation is expected from the attending students.

Assessment methods

The exam will be oral and will  last approximately 45 minutes.
The student will be asked in succession to illustrate three topics among those covered in the course. For each topic, they will first be asked to explain the general framework, then specific aspects will be explored in detail.
The depth and correctness of the answers will also be evaluated in light of practical and computational learning.
The mark will be an arithmetic average of the marks for each of the questions.

Teaching tools

The course uses presentations that will have connections to online resources, such as seismological databases and codes, which will contribute to the student's computational and data training.
The exercises include instructions given in advance for installing codes and downloading datasets on a personal computer, for in-class exercises and, optionally, outside of course time.

Office hours

See the website of Luca De Siena

See the website of Paolo Gasperini