TIME SERIES ANALYSIS IN THE EARTH SCIENCES
Steve P. Lund
Overview
This class is designed for first or second year graduate students to familiarize them with the adaptation of numerical and graphical methods in the Earth Sciences to a computer environment. The course will stress the use or development of computer alogrithms to solve typical numerical or graphical problems. A secondary goal will be to learn basic elements of computer hardware, operating system languages, and practical computing skills that are necessary to effectively use the Department of Earth Sciences computer facilities.
There are no prerequisites for this course, but a senior-level undergraduate mathematics background (calculus and differential equations) is preferred. No previous knowledge of computer languages is required, but students without a working knowledge of Fortran can expect to participate in two weeks of remedial classes (see below).
Course Structure
The class will entail two one-hour lectures and one two-hour lab each week. The lectures will cover computer algorithm development and usage of 'canned' algorithms for the solution of numerical problems. The lab will cover computer hardware, computer system languages, and practical skills needed to operate the departmental computers. A remedial class will be offered by the instructor one extra hour per day for the first two weeks to teach Fortran to anyone without previous computer experience.
Grading will be based completely on the solution of problem sets (75%) associated with the lecture material and a final project (25%) on a computer topic of your choice. The book 'NUMERICAL RECIPES - The Art of Scientific Computing' by Press, Flannery, Teukolsky, and Vetterling is the recommended (but not required) text for the class. The text also has a fortran example book associated with it. Both books are available in the bookstore.
Lecture Topics
TOPIC 1: Modeling Studies (1): Linear least-squares estimation (text chapter 14).
TOPIC 2: Modeling Studies (2): Interpolation/extrapolation (text chapter 3)
TOPIC 3: Random Processes (1): Elementary statistics and tests (text chapter 13)
TOPIC 4: Random Processes (2): Vector statistics and tests
TOPIC 5: Random Processes (3): Random numbers and simple time series (text chapter 7)
TOPIC 6: Modeling Studies (3): Jackknifing and bootstrapping techniques (text chapter 14)
Example 1: Smoothing paleomagnetic time series
Example 2: Jackknife test in magnetostratigraphy
Example 3: Earthquake focal depth estimation
TOPIC 7: Modeling Studies (4): Generalized model building
Example 1: Parametric Earth models (PREMs)
Example 2: Sedimentary basin evolution
TOPIC 8: Modeling Studies (5): Using STELLA for modeling physical systems in the Earth Sciences.
TOPIC 9: Contouring spatial data using PGPLOT and STEREONET
TOPIC 10: Random Processes (4): Spectral Analysis (text chapters 7, 12)
Example 1: Milankovich frequencies in paleoclimate records.
TOPIC 11: Random Processes (5): Multivariate statistics - factor analysis using BMDP.
Example 1: Fourier grain-shape analysis in sedimentology
TOPIC 12: Finite Difference Techniques
Example 1: Dynamical systems - mass on a spring
Example 2: Groundwater flow through porous media
TOPIC 13: Numerical Integration (text chapters 4, 15)
Lab Topics
LAB 1: An introduction to the Earth Sciences computer facilities and your default VAX user account.
a) Computer room layout and machines
b) User room layout
c) Remote access
d) Computer facility rules
e) Login to your VAX user account
LAB 2: Introduction to the VAX operating environment.
a) VAX/VMS
b) LOGIN.COM
c) subdirectories
d) editting
LAB 3: Hardcopy and graphics (PGPLOT) on the VAX computers.
LAB 4: Macintosh operating system (Macintosh computers)
LAB 5: MS/DOS operating system (IBM PCs)
LAB 6: Unix operating system (Sun computers)
LAB 7: Hardcopy and graphics on the Sun computers.
LAB 8: Computer networks/communications (TELNET, FTP, KERMIT)
LAB 9: Special software packages (MATLAB, STELLA, BMDP)
LAB 10: The INTERNET - Communication with databases around the world
LAB 11: Hardware structure of computers
a) microVax
b) IBM PC
c) Macintosh
Comments on Grading
Problem Sets (75%): I will grade the problem sets on the basis of scientific method (show all source code developed), the computer results, and a clearly written (less than one page) summary of the problem and what you did.
Term Paper (25%): I will grade the term papers on the basis of scientific content, grammar, and editorial style. Treat this paper as an introduction, background, or methods section for an MS thesis or as a minor proposal for the Phd qualifying exam. Elements of style which I will consider are listed below.
A) Outline of term paper:
1) introduction: background, what scientific problem do you want to consider?
2) body: Describe your algorithm (user developed or canned), its purpose, and any results.
3) conclusions: did you summarize all data, are conclusions reasonable, did you forget any applicable information?
B) Points I will consider:
do you: 1) describe the purpose of the paper?
2) provide a background for the paper?
3) describe algorithm/scientific problem?
4) properly interpret any data?
5) summarize the project?
6) draw reasonable conclusions?
7) indicate where more work is needed?
8) have a logical and well-organized paper?
9) use good grammar?
10) provide adequate references?
Possible Topics
Any scientific problem in the Earth Sciences that requires computing to generate, analyse, or display data to clarify a problem may be appropriate. YOU MUST CLEAR YOUR RESEARCH TOPIC WITH ME BEFORE YOU START!