Date of Award

8-2015

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School

College of Science and Mathematics

Department/Program

Earth and Environmental Studies

Thesis Sponsor/Dissertation Chair/Project Chair

Mark J. Chopping

Committee Member

Sandra Passchier

Committee Member

Dirk Vanderklein

Committee Member

Marc Imhoff

Abstract

The observed greening of Arctic vegetation and the expansion of shrubs in the last few decades has likely had profound implications for the tundra ecosystem, including feedbacks to climate. Uncertainty surrounding the magnitude, direction, and implications of this vegetation shift calls for monitoring of vegetation structural parameters, such as fractional cover of shrubs. Due to the extent of the North Slope of Alaska and its extreme environments, remote sensing may be the most suitable tool to produce wall-to-wall fractional shrub cover maps for the entire region, however, most regional maps have relied on vegetation indices or needed many years worth of data to cover the whole region. Here, a new mapping approach is presented that uses satellite imagery from the Multi-angle Imaging SpectroRadiometer (MISR) sensor and some landscape variables to predict tall shrub (> 0.5 m) cover with the ultimate goal of evaluating temporal changes in tall shrub fractional cover during the period of 2010-2000. Specifically, we: 1) undertook two field surveys in the North Slope of Alaska to obtain estimates of tall shrub cover, canopy height, crown radius, and total number of shrubs at 26 sites (250 m × 250 m each); 2) evaluated the ability of the semi-automated image interpretation algorithm CANAPI - CANopy Analysis from Panchromatic Imagery, to derive structural data for tall (> 0.5 m) shrubs in the Arctic; 3) constructed a robust reference database with estimates of shrub structural parameters; 4) trained and validated the boosted regression tree model to predict tall shrub fractional cover from moderate resolution imagery; 5) created the 2000 and the 2010 tall shrub fractional cover map for the North Slope of Alaska; and 6) evaluated the changes in shrub abundance during the period 2010-2000 in the North Slope of Alaska. Results from the field surveys suggested that tall shrub fractional cover was less than 5% at 250 m scales. The evaluation of the CANAPI algorithm showed that CANAPI could successfully retrieve fractional cover (R2 = 0.83, P < 0.001), mean crown radius (R2 = 0.81, P < 0.001), and total number of shrubs (R2 = 0.54, P < 0.001) from very-high resolution imagery. As a result, a robust reference database was constructed with estimates of tall shrub fractional cover, canopy radius, and total number of shrubs for 1,039 sites across the domain of the North Slope. After the training and validation of the Boosted Regression Tree (BRT), the best model used 14 predictor variables and explained 52% of the variation in the response variable, fractional cover. The red reflectance, slope, nadir Bidirectional Reflectance Distribution Function (BRDF) adjusted weight of determination, and isotropic scattering kernel were the variables more often used to generate the regression trees, and therefore they contributed the most to the model. The trained BRT model was used to construct the tall shrub fractional cover map for the year 2000 and 2010 using moderate resolution imagery. The maps revealed that cover ranged from 0.00 to 0.21 and about 75% of the sites had a fractional cover less than 0.013. High cover values were predicted along floodplains, creeks, and sloped terrain. The 2000 MISR-derived fractional cover map presented here outperformed the 2000 Landsat-derived tall shrub fractional cover map when compared to the robust validation data set (R2= 0.38, Root Mean Square Error (RMSE) = 0.08). Temporal comparisons of tall shrub abundance in the MISR-derived maps suggested that shrubs expanded during the period 2000-2010. The extent of the area that unequivocally experienced a robust change in tall shrub cover was less than 1 % (1,487 km2) of the total area of the North Slope of Alaska (213,090 km2). It is possible that tall shrubs may have expanded throughout a larger area but there is insufficient precision in the MISR-based estimates to make an unequivocal determination. Nevertheless, it seems that there was a positive trend toward an increase in shrub cover considering that 95% of the locations that had a robust change saw an increase. The tall shrub cover expansion rate varied between 0.006 yr-1 and 0.017 yr-1, being higher along the forest-tundra ecotone, north of the Brooks Range. More research is necessary to determine if the increase in cover corresponded to the advance of the tree line, or to the expansion of the tall shrubs, or both.

Share

COinS