Doctoral Dissertations
Date of Award
8-2024
Degree Type
Dissertation
Degree Name
Doctor of Philosophy
Major
Geology
Major Professor
Molly C. McCanta
Committee Members
Nick Dygert, Tim Glotch, Linda Kah
Abstract
Earth and Venus are dominated by volcanic landforms, making accurate characterization of the diversity of igneous compositions on the planets’ surfaces key to understanding their geologic histories. Spectral observations in the visible-near-infrared (VNIR) and mid-infrared (MIR) provide rapid characterization of surface composition and mineralogy but are not without limitations. Spectra are sensitive to a number of factors, including observation geometry, collection method (emissivity or reflectance), grain size, phase assemblage, and composition. It is impossible to constrain the various effects of all factors when interpreting orbital data, and simplifications are made, guided by laboratory investigations.
A common simplification is to ignore changes due to observational geometry. Particulate surfaces are not perfect isotropic scatterers, though variations in bi-directional reflectance geometry yield only minor changes to spectra. Venusian surface spectra, however, are collected in the uncommon VNIR configurations of hemispherical reflectance or directional emissivity. Chapter 1 investigates the differences that collection in bi-directional or hemispherical reflectance geometries can have on VNIR spectra. Changes to spectral slope and spectral feature shape are observed between the two geometries, particularly for mafic materials. Chapter 1 also investigates how applicable radiative transfer modeling is to convert between the two geometries.
Chapter 2 investigates the detectability of sulfate alteration on Venus surface compositions using VNIR hemispherical reflectance analyses of Venus-analog basalt-hematite-sulfate mixtures. Surfaces composed of multiple phases are more difficult to interpret spectrally than those of a single phase as spectral signals from the phases interact. Hematite and two sulfate phases, anhydrite and thenardite, are used to simulate alteration mineralogy, resulting in different trends in spectral variability relative to sulfate content. In the spectra of samples with greater than 20% alteration phases, primary basalt spectral features are completely obscured.
Chapter 3 investigates the utility of partial least squares (PLS) regression models in characterizing the oxide composition and crystallinity of volcanic samples from VNIR and MIR data. PLS models provide quantitative predictions that are insensitive to compounding factors, minimizing the assumptions necessary to obtain surface composition. PLS efficacy is also not strongly affected by variations to spectral resolution, making it a powerful tool for maximizing data analysis.
Recommended Citation
Leight, C.J., "Spectral Characterization of Volcanic Materials on Earth and Venus. " PhD diss., University of Tennessee, 2024.
https://trace.tennessee.edu/utk_graddiss/10476