ARCTIC GREENING: CHARACTERIZING TUNDRA VEGETATION FROM IN-SITU AND REMOTELY SENSED OBSERVATIONS

• Main Author: Ellenson, Shira Ann
• Format: Thesis
• Language: unknown
• Published: University of Montana 2022
• Subjects: photogrammetry; tundra; vegetation; Canopy Height Model; remote sensing; Geographic Information Sciences; Physical and Environmental Geography; Arctic; Arctic Greening; Tundra
• Online Access: https://scholarworks.umt.edu/etd/12031; https://scholarworks.umt.edu/context/etd/article/13150/viewcontent/Ellenson_Shira_Thesis.pdf

As the Arctic has warmed at twice the rate of the global average, vegetation productivity has also been increasing. While satellite remote sensing is useful for summarizing Arctic-wide trends, changes in tundra species heights, densities, composition, and distribution can be missed at coarse resolution. Smaller, plot-scale studies are necessary to better understand vegetation dynamics at fine scales occurring on the ground. In 1995, high-resolution traditional aerial photographs and in-situ measurements of vegetation characteristics were taken at a series of plots established on the Alaskan North Slope. Repeat field surveys in 2021 revealed increases in plant cover for deciduous shrubs and graminoids and decreases for bryophytes that were substantial at some sites. The overall mean canopy height doubled from 6.1 to 13.7 cm (individual sites increased 63 to 513%) and overall mean maximum shrub height doubled from 31.7 to 60.0 cm (individual sites increased 26 to 170%) between 1995 and 2021. Tundra vegetation within these plots has changed substantially over this 26-year period. I used traditional air photos to measure tundra vegetation by applying a modern photogrammetry workflow and LiDAR-based classification to generate canopy height models (CHMs) at varying resolutions. When compared to field measurements, mean point-cloud estimated canopy heights calculated at 10 cm pixel resolution showed differences of 0 to 2.8 cm at sites with mean canopy heights of 2.3 to 5.8 cm. At sites with taller vegetation (mean canopy heights of 13.8 and 18.5 cm) differences were 0.7 and 1.6 cm, respectively, from point-cloud estimates made at 25 cm pixel resolution. Finer resolution CHMs performed better with shorter canopy heights, and coarser resolution CHMs worked better with taller canopy heights. These results suggest that using high-resolution remote sensing paired with in situ measurements can estimate canopy heights for tundra vegetation, allowing small changes to be detected. This study contributes to the record of ...