Examining Extremes: A Comparison Between Machine Learning Based Southwestern CONUS Seasonal Drought and High Fire Risk Event synoptic Climatologies

Authors

• Ethan M. Greenberg, University of Nebraska-LincolnFollow

First Advisor

Mark D. Svoboda

Committee Members

Fransisco Munoz-Arriola, Tsegaye Tedere

Date of this Version

5-2026

Document Type

Thesis

Citation

A thesis presented to the faculty of the Graduate College at the University of Nebraska in partial fulfillment of requirements for the degree of Master of Science

Major: Natural Resource Science

Under the supervision of Professor Mark D. Svoboda

Lincoln, Nebraska, May 2026

Comments

Copyright 2026, Ethan M. Greenberg. Used by permission

Abstract

The Southwestern United States (SW CONUS), comprised of California, Arizona, Nevada, and Utah, is a vast region that tens of millions of people call home. Hosting ecosystems ranging from grasslands and shrublands to temperate forests and climate zones ranging from Mediterranean climates to arid deserts, the region is nearly unanimously prone to intense droughts and devastating wildfires. While fire weather conditions and drought are often studied separately or mentioned as background conditions for the other, there are relatively few studies that compare the typical meteorological conditions between the two. This study aims to more closely understand the meteorological relationship between drought and extreme fire weather by comparing synoptic climatologies of 200 (upper), 500 (middle), and 700 hPa (low-level) heights during days in months with seasonal droughts, defined with an SPEI-1 and SPEI-3 of up to -0.8, and days with a Fire Weather Index (FWI) of at least the 90th percentile in four subregions of the SW CONUS. Synoptic climatologies were generated using the machine learning based K-means clustering algorithm fit on the principal components of daily averaged geopotential fields on days with drought in at least one of four subregions of the SW CONUS. Droughts and extreme fire weather conditions often share hot, dry conditions and can have temporal overlaps, driven by the same global teleconnectionsand synoptic-scale weather patterns. Given the expectation of some temporal overlap and the general presence of similar driving weather conditions, there was an expectation of similar-looking geopotential clusters for SW CONUS extreme fire weather and drought events. This study found that both extreme fire weather and drought events have an overwhelming majority of their variance explained by clusters with dominant ridges over or in close proximity to the areas experiencing either drought or fire events. Despite similar-looking patterns at the macro-level, the exact shape and placement of the ridges were different between drought and fire weather events, with fire weather events showing clusters with slightly more amplified patterns. Understanding the synoptic climatologies of drought and fire weather can serve as a building block for improved long-term forecasts of drought and wildfires.

Advisor: Mark D. Svoboda