Multi-Scale Detection and Characterization of Physical and Ecological Change in the Arctic Using Satellite Remote Sensing

Abstract

The Arctic environment is in a state of transition as the result of climate warming. There are tremendous implications for ecology, human well-being, national security, and energy and economic development among many other critical issues. The effects of Arctic warming are multifarious. Both physical and ecological processes have been affected in a set of complex feedback mechanisms.

The objective of this dissertation research is to develop new technology though advanced computing and improved data availability to identify environmental baselines and trends. This is accomplished through the use of different satellite remote sensing platforms at different spatial and temporal scales. This research provides needed methods and baseline data and identifies temporal trends for environmental monitoring programs in the Arctic. Among other uses, the data presented here will allow future evaluation of continued change and disturbance.

Using electro-optical satellite data spanning the pan-Arctic, and standardizing in terms of data and satellite platform, an unbiased comparison across numerous disparate variables and across terrestrial and marine environments provides a method to detect and evaluate environmental change. Using MODIS standard products of land surface temperature, percent snow covered area, NDVI, EVI, phenology, burned area, marine chlorophyll, CDOM, sea surface temperature, and marine primary productivity, time series data 2000-2017 shows significant temporal trends in almost all variables. Analysis of seasonal data reveals significant breakpoints in temporal trends over the 18-year period. Within the terrestrial environment, data shows significant increasing trends in land surface temperature and NDVI. In the marine environment, significant increasing trends are detected in primary productivity. Significantly earlier onset of greenup date and longer end of growing season are observed in certain bioclimate subzones. Terrestrial and marine observations show similar rates of change with unidirectional change in terrestrial and significant directional and magnitude shifts in marine.

There is great potential to use a specific type of active remote sensing, Synthetic Aperture Radar (SAR), for Arctic change detection and disturbance monitoring due to its all-weather imaging capabilities and the ability for SAR to provide information on vegetation structure, biomass, and moisture conditions. In contrast to electro-optical studies from the same region, measures of landscape recovery to pre-burn conditions, as detected by SAR, are on the order of four to five years instead of one. A 3 dB difference exists between burned and unburned tundra, with the best time for burned area detection being late in the growing season before frozen ground conditions develop. Models developed using in-situ field data in burned and unburned areas of the tundra show SAR backscatter is a function of moisture condition, shrubs, tussocks, and fire. Contributions of different biophysical variables to the backscatter signal vary as a result of year since fire, but models show contributions from all four categories in almost all fire histories.

Advanced methods to look at synoptic change at a variety of scales are needed to increase our understanding of this remote environment and provide a baseline status upon which to evaluate change. This dissertation works toward the development of methods and datasets for change evaluation.