Date of Award

Spring 5-2018

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mathematics

First Advisor

Frederick A. Adkins, Ph.D.

Second Advisor

Nicholas D. Deardorff, Ph.D.

Third Advisor

Yongtao Cao, Ph.D.

Abstract

In recent decades, lidar has revolutionized topographic mapping of the Earth and planets through the use of digital elevation models (DEMs). However, the return amplitudes of the reflected laser pulses, typically collected as part of a lidar dataset, have seldom beenused as a means of identifying and characterizing volcanic surface features such as lava flows, rafted tephra and agglutinate, and pyroclastic deposits consisting of tephra and ashfall. Here, we find an effective process for remotely characterizing volcanic terrains using a simple but rigorous cluster analysis of lidar return amplitudes and DEM data to define the parameters for a self-organizing mapping routine. The data used for this study, collected from the Northwest Rift Zone on Newberry Volcano in central Oregon, has been accurately geo- referenced, providing 3 foot horizontal and 4.5 centimeter vertical resolution. In addition, the return amplitude values were recorded with a horizontal resolution of 1.5 feet. An appropriate number of terrain categories is found by applying an incremental within-cluster sum of squares algorithm to generate clusters from random subsets of lidar DEM and reflectivity data chosen from the study area. From these results, a silhouette analysis determines the optimum number of clusters to be used as a necessary input parameter for the categorization of each data cell by means of a self-organizing mapping function. These results, confirmed by comparison with field work conducted at Mokst Butte and its associated lava flows, is then applied to other, less accessible volcanic terrains. The resulting false-color imagery allows precise identification of volcanic morphologies that are otherwise unrecognizable in remote sensing data such as lidar, InSAR, and orthorectified color photography, and in regions where traditional field work is difficult or unfeasible.

Share

COinS