EARTH 362: Data Analysis for Earth and Planetary Sciences

Course Details

Class Lectures: Tues Thurs 12:30-1:50 PM, Tech F285

Instructor: Seth Stein

The Earth is a messy and complicated system that we study with data that are imprecise, inaccurate, inconsistent, and insufficient. As a result, our ideas about how the earth works and what it will do in the future have considerable uncertainties. This course introduces some approaches from statistics and probability and explores how are used to address issues of uncertainty and forecasting in the geosciences including natural hazards, climate change, and how the planet works. Topics include Fermi estimation, precision and accuracy, variance and covariances, propagation of errors, histograms, Gaussian distributions, lognormal distributions, central limit theorem, power law distributions, rejection of data, linear regression, basic probability, binomial distribution, Poisson distribution, chi-square tests. Grading is based on homework, in-class problems, a project, and write-ups of several department seminars.

Required Texts

An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements (Paperback) by John R. Taylor, University Science Books, 1997

Useless Arithmetic: Why Environmental Scientists Can’t Predict the Future by Orrin H. Pilkey & Linda Pilkey-Jarvis, Columbia U. Press, 2007

Grade:

30%: Weekly homework problems
30%: 2-page writeups of three department seminars, discussing the accuracy, precision, consistency, and adequacy of the data used, due within one week of the seminar
30%: Term project
10%: In-class problems

Portable electronic devices may not be used in class. Bring handouts to lecture. You may work with other students on the problems, if at the beginning of each assignment you list whom you worked with and on which parts. Problem sets are due one week after being assigned, unless prior arrangements have been made. Class question make-ups are only allowed by advance arrangements. Unexcused late work will be penalized. As per university policy, class attendance is expected and required.

 

Class handouts (downloadable pdf)

Excel file of data for sloss room table

Fortran function for Gaussian

Fortran subroutine for linear least squares 

Forsyth & Uyeda table for problem set 7 (xls)

Handout on using GPS data in teaching data analysis (pdf)

GPS data for problem set 9 (xls)