| Arctic regions have proven to be particularly vulnerable and sensitive to global climate change. In fact, Alaskan tundra ecosystems have experienced accelerated warming over the past three decades, which has caused changes in the rates and magnitudes of carbon exchange between the vegetation, soils and atmosphere. Remotely sensed vegetation indices, such as the normalized difference vegetation index (NDVI), provide a near continuous, global-scale means of monitoring ecosystem form and functions related to carbon uptake and storage. This unique tool could be especially useful for monitoring the response of remote Arctic tundra ecosystems to the observed warming. However, there are confounding factors that inhibit our ability to accurately interpret these unique, large-scale datasets. In this thesis, I relate plot-level measurements of NDVI, derived from radiance data collected with a field portable spectrometer, to ecosystem form and function at multiple spatial scales in an Arctic tundra ecosystem. Specifically, I establish relationships among NDVI, aboveground plant biomass and ecosystem carbon fluxes in two of the three dominant tundra vegetation communities near the Toolik Lake Long-Term Ecological Research (LTER) site, located on the north slope of Alaska. I found that across all treatments in wet sedge tundra communities, NDVI is correlated with aboveground biomass (r2 = 0.84), gross ecosystem production (r2 = 0.75) and ecosystem respiration (r2 = 0.71). Similarly, NDVI was found to be correlated with aboveground biomass (r2 = 0.56) in tussock tundra communities. I proceed to show how these relationships are community specific as a result of the contrasting degree of heterogeneity in species and functional types that make up wet sedge and tussock tundra communities. Subsequently, I explore, compare and contrast inter-annual variability in NDVI in each of these two vegetation communities. Finally, I explore our ability to scale-up NDVI from the plot-level to the landscape-level at Imnavait Creek, a small Arctic tundra watershed that has been intensively and consistently studied since the mid-1980's. I found that an abrupt and persistent increase in the average NDVI value between our plot-level sampling average and our landscape-level sampling average, caused by the combination of our systematic, plot-level sampling scheme, and the coincidental systematic 10 to 30 m spacing between adjacent watertracks, landscape features shown to be associated with higher NDVI values than the surrounding vegetation. |
