On the relationship between Landsat-derived canopy cover and tree ring time-series across an urban-rural temperate forest region

Authors: Mitchell Bonney*, University of Toronto - Mississauga, Yuhong He, University of Toronto - Mississauga
Topics: Remote Sensing, Physical Geography, Environmental Science
Keywords: Landsat, time-series, tree rings, dendrochronology, canopy cover, temporal segmentation, forest change
Session Type: Virtual Paper
Day: 4/9/2021
Start / End Time: 6:15 PM / 7:30 PM
Room: Virtual 30
Presentation File: No File Uploaded

There is interest in better understanding the relationship between vegetation proxies derived from remote sensing and tree ring width (TRW) time-series, which would benefit dendrochronologists and remote sensing scientists. Previous research has focused on temperature or moisture limited coniferous forests, using lower spatial resolution (e.g., 8000 m) sensors and normalized difference vegetation index (NDVI) products to test relationships over 15-30 years. Advances in building standardized Landsat (30 m) time-series may allow for expansion into new areas (e.g., fragmented forests) and over longer periods (e.g., nearly 50 years). Here, we study the relationship between TRW and Landsat-derived canopy cover (%CC) from 1972-2018 across 16 deciduous, mixed, and coniferous stands in southern Ontario, Canada. Stand-level TRW chronologies are built with cores from canopy-dominant trees and compared with overlapping Landsat data. Time-series approaches, including detrending to a ring-width index (RWI), moving-window correlation and temporal segmentation are applied. We found significant positive correlations at 9 of 16 sites, with coniferous sites relating more strongly than deciduous and especially maple (Acer saccharum). Relationships are most strong in the 1980s, when RWI and %CC experience growth, and after 2000, where both markedly decline since 2010 at almost every site. We demonstrate the utility of tree rings in validating Landsat-observed change in urban-rural temperate forests, mainly in cedar (Thuja occidentalis), hemlock (Tsuga canadensis) and poplar (Populus balsamifera) dominated stands. We push Landsat to its limit, quantifying differences in %CC and monitoring gradual change over nearly 50 years in dense canopies.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login