Date of Award

January 2018

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Atmospheric Sciences

First Advisor

Aaron D. Kennedy

Abstract

Areas that reside in the high-latitudes such as the northern United States can experience hazardous conditions during the winter months due to snowstorms. When strong winds exist with falling or freshly fallen snow, blizzard conditions are able create significant personal, societal, and economic impacts for the Northern Great Plains. While the climatology for these extreme snowstorms is known, the frequency and intensity of how these events may change in a warming climate is not certain. In order to determine how extreme snowstorms may change in the future climate, climate models can be used but the horizontal and vertical grid spacing makes identifying blizzard events difficult. Moreover, climate models do not include blowing snow, which means that blizzards that don’t have any falling snow are not considered. Therefore, another method must be used in order to identify these extreme snowstorm events.

The presented work will use a competitive neural network known as the Self-Organizing Map (SOM) to identify meteorological patterns associated with blizzard events over the Northern Great Plains from 1979-2015. Once these large-scale patterns are identified from observations, they will be identified in the Community Earth System Model (CESM) 4.0 20th Century forcing climate simulations run in support for the Coupled Model Intercomparison Project Phase 5 (CMIP-5). In specific, the methodology will rely on the ‘Mother of All Runs’ (MOAR) ensemble member, which allows for specific meteorological patterns to be identified. Blizzard events will be identified during historical time periods to determine biases, and then under future emissions scenarios.

Download

Share

COinS