Date of Award

May 2026

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Earth System Science & Policy

First Advisor

Jeffrey VanLooy

Abstract

Glaciers in high mountain regions are the source of critical water supply for a large portion of the global population, providing freshwater for several uses such as agriculture, irrigation, drinking water, energy production, and the economy. Climate change is causing the increased melting of mountain glaciers, and in turn putting constraints on several of the aforementioned uses of freshwater that is sourced from high mountain regions as well as increasing the risk of natural disasters such as mudslides and glacial lake outburst floods. Understanding the factors that influence glacial melt is important when assessing the mass balance of glaciers and the outcomes that arise from increased melting. While many melt and mass balance models applied at regional scales are based on temperature and precipitation, there is an increased interest in the role of albedo on influencing glacial melt. Using remote sensing, albedo data can be collected over large areas with little cost and relative ease. However, limitations in doing so include temporal resolution of the satellite being used coupled with the variability in clear sky observations which are necessary for accurate data collection. Oversaturation of imagery is also an obstacle faced in imagery that has fresh snowfall, resulting in inaccurate data. The objective of this study was to create an albedo melt season model based on melt season data (2000, 2007, 2015, 2017, and 2018) that allows the estimation of albedo at any particular point in the Wind River Range, Wyoming melt season, which creates a bypass to the temporal resolution, oversaturation, and cloud cover limitations of remote sensing. To create this model, albedo is correlated with several other variables, namely elevation, temperature, precipitation, previous winter snowfall, and radiation, using data that can be obtained from the USGS Earth Explorer website, PRISM Climate Data website, and a SRTM DEM. All of the Landsat remote sensing data are 30 m spatial resolution and 16-day temporal resolution. The PRISM data can be obtained at individual coordinate points, enabling the PRISM data to coincide with 30 m spatial resolution remote sensing data. A generalized additive model is used to analyze correlation, which allows the analysis of a predictive relationship between albedo and the other variables at any single pixel in the imagery, applied across the entire study area which will result in an estimation of albedo at random points based on the PRISM and DEM variables. A stepwise analysis of correlation coefficients will determine which variables are best to include in the model, and if any, which are redundant. The same model is tested with different combinations of interactions and without any interactions to compare and determine the best result. Validation of the model is completed by comparing the albedo model results with actual observed albedo data that is not used in the creation of the model. This approach resulted in a generalized additive model (GAM) with main effect and interaction variables producing the best model fit with the independent variables explaining about 43% of the variance in albedo using five years of melt season data during the 2000-2018 period.

Download

Share

COinS