Paula Crock

Date of Award

January 2017

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Space Studies

First Advisor

Santhosh Seelan

Abstract

Determining the ages of glacier drifts in Antarctica can help paleoclimatologists determine

the changes Earth’s climate has gone through and thereby inform models for future climate

change prediction. However, many of these drifts are difficult to reach for sample collection

necessary to determine their ages. This research attempts to use multispectral remote sensing

data to expand the mapping of drift ages from known point measurements regionally. This

research is based on existing drift ages from Ong Valley, Transantarctic Mountains. Two

methods were used to determine a combination of the image band data that would sufficiently

distinguish the three age-distinct drift regions: Principal Component Analysis (PCA) and an

empirical analysis based on observed trends in the data. The PCA results showed that virtually

all bands contribute equally to the differences in the image data from the three drift regions,

precluding the use of a small number of bands in an index to classify the regions. An index was

developed from the empirical analysis but this index was unable to sufficiently overcome the

count variations in the data sets to successfully classify the regions. Although neither method

provided a conclusive means to distinguish the drift regions from the remote sensing data used in

this analysis other remote sensing data, e.g. – data at different or more extensive bands ranges, or

other analysis techniques, e.g. – more preprocessing of the data or machine learning algorithms

applied to the image data, may yet yield successful results.

Download

Share

COinS