Award Date

5-1-2025

Degree Type

Thesis

Degree Name

Master of Science in Engineering (MSE)

Department

Civil and Environmental Engineering and Construction

First Committee Member

Haroon Stephen

Second Committee Member

Sajjad Ahmad

Third Committee Member

David James

Fourth Committee Member

Venkatesan Muthukumar

Fifth Committee Member

Matthew Petrie

Number of Pages

150

Abstract

Wildfires are increasingly common in sagebrush ecosystems across the western United States, leading to vegetation loss and ecosystem restructuring. This study investigated vegetation recovery within two large Nevada burn scars, the Snowstorm Fire of 2017 and the South Sugar Loaf Fire of 2018. Landsat 8 surface reflectance imagery supplied multi temporal spectral data, and vegetation burn severity was mapped with the difference Normalized Burn Ratio. The research aimed to quantify how vegetation health spectral indicators respond over time across severity gradients and to detect shifts in land cover composition from pre fire to post fire conditions.Four spectral indices—NDVI, MSI, MCARI2, and land surface temperature—were assessed with the Mann Kendall trend test, followed by a linear mixed effects model that linked time and severity class to spectral change. ISODATA clustering was conducted to determine pre-fire and post-fire classes, with a stability matrix and change matrix tracking the fate of each pre-fire class with post-fire classification. LANDFIRE EVT maps were assessed to test their sensitivity to short term ecosystem change. Severity mapping showed contrasting fire behavior. Snowstorm burned chiefly at low severity and covered sixty nine percent of the area, whereas South Sugar Loaf contained similar proportions of high and moderate high severity at thirty-two and thirty three percent respectively. NDVI rose at both sites, confirming recovery. In Snowstorm the increase was uniform across severity classes, but in South Sugar Loaf low severity pixels gained more greenness than high severity ones, indicating that high severity had an impact on vegetation regrowth. MSI declined over time at both sites, indicating moisture recovery, with the largest decline in moderate high severity zones at South Sugar Loaf, indicating the role of burn severity, MCARI2 a chlorophyll proxy, climbed steadily and was again most responsive in the lower severity classes of South Sugar Loaf as compared to high severity, whereas there was no effect of severity in Snowstorm fire. Land surface temperature declined at Snowstorm fire area and showed localized reductions in low severity areas at South Sugar Loaf. The Mann-Kendall trend test was non-significant for fire areas but showed localized positive trends. Linear Mixed Effects model and Mann-Kendall test revealed time since fire to be the primary driver of recovery, while severity modulated local trajectories. ISODATA distinguished five pre-fire and seven post-fire classes at South Sugar Loaf and five pre-fire and post-fire classes at snowstorm. South Sugar Loaf underwent pronounced compositional change: its dominant shrubland class retained just one third of its original area, shifting largely to herbaceous and sparse vegetation categories. Conifers in South Sugarloaf fire lost most of it area to sparse vegetation in post-fire. Snowstorm exhibited greater stability for shrubs species, with more than half of the Herbs Shrubs Mix-2 class persisting. Across both fires new post fire clusters displayed lower near infrared and higher short wave infrared reflectance, signatures typical of ash and bare substrate. LANDFIRE EVT maps failed to register these changes, remaining static in both South Sugar Loaf and Snowstorm. Combining time series spectral metrics with unsupervised classification captured both gradual recovery and abrupt compositional shifts, offering a comprehensive view of postfire dynamics. The persistence of outdated classes in LANDFIRE EVT underscores the need for more agile vegetation mapping frameworks that can accommodate rapid ecological change in fire prone landscapes.

Keywords

post-fire assessment; post-fire spectral signatures; post-fire vegetation monitoring; vegetation recovery; wildfire impact analysis

Disciplines

Remote Sensing

File Format

pdf

File Size

6100 KB

Degree Grantor

University of Nevada, Las Vegas

Language

English

Rights

IN COPYRIGHT. For more information about this rights statement, please visit http://rightsstatements.org/vocab/InC/1.0/

Download


Share

COinS