Quantifying four decades of disturbance dynamics in North American Arctic and Boreal ecosystems using Landsat time series
OA Version
Citation
Abstract
Climate change in the Arctic and Boreal Regions (ABR) is unfolding at twice the rate of the global average over recent decades. In the past 40 years, satellite observations have revealed extensive changes in land cover within this area, attributed to various disturbances coupled with a warming climate. However, with relatively sparse satellite data density in high latitudes, it has been challenging to quantify the extent of disturbances at moderate resolution and assess how climate change influences the nature and frequency of these disturbance events. The first step of mapping disturbances in the Arctic and Boreal Region is to determine the spatial and temporal frequency of usable satellite observations. Remote sensing satellites effectively characterize ecosystem changes and monitor disturbances, but finer details like logging patches and pest damage can be challenging to detect with sensors like AVHRR and MODIS. Our analysis indicates that the current acquisition status and strategies of the Landsat Program enable sufficient observations for time series analysis in high latitudes, expect for Alaska before 2000 (Zhang et al., 2022a).
A new method using pre- and post-disturbance Tasseled Cap (TC) values at peak summer derived from the Landsat time series can distinguish multiple causal agents of disturbance over the core domain of NASA’s Arctic-Boreal Vulnerability Experiment, and was used to map disturbances caused by fire, insect, and logging activities on an annual basis from 1987 to 2012. In addition, an iterative post-processing procedure improved the accuracy of our map (Zhang et al., 2022b).
Application of this new methodology and several significant improvements enabled the extension of the spatial and temporal scope for the quantification of disturbance agents of fire, insects and logging of Arctic and Boreal Canada and Alaska from 1987 to 2021. The findings indicate: 1) insect damage has increased significantly over the past two decades; 2) the area burned is highly variable from year to year but does not exhibit temporal trends; 3) interannual variability in logging is low, with macro trends in logging activities closely bound to economic factors and market dynamics; and 4) post-disturbance boreal forest recovery is too complex to be characterized exclusively by the widely used Normalized Difference Vegetation Index (NDVI). The Landsat-derived Tasseled Cap Wetness (TCW) index better characterizes the vegetation recovery, particularly in conifer forests. Results for field sites reveal that conifer forests often require 35 years for natural recovery, and insect-induced forest mortality even longer.
Description
2024
License
Attribution 4.0 International