Authors: Mitchell Bonney*, University of Toronto - Mississauga, Yuhong He, University of Toronto - Mississauga
Topics: Physical Geography, Remote Sensing
Keywords: dendroecology, remote sensing, Landsat, time-series, forest dynamics
Session Type: Paper
Presentation File: No File Uploaded
Dendroecologists are interested in increasing the spatial scope of their results, while those who analyze remote sensing time-series to monitor vegetation change are concerned with ground validating past changes through time (i.e., change validation). Both tree ring widths (TRW) and Landsat remote sensing provide information related to forest productivity at a yearly temporal resolution. Researchers in both fields also make use of remarkably similar time-series methods to extract useful trends from often-noisy temporal datasets. Despite these connections, research relating Landsat (30 m) time-series with TRW is almost non-existent. Most studies that relate dendrochronology with remote sensing have done so with low resolution (e.g., 8 km) satellite data and have focused on boreal locations. In this study, we test the relationship between TRW and a canopy cover (%CC) time-series derived from Landsat (1972 - 2018) across 16 temperate deciduous, mixed and coniferous stands in a southern Ontario watershed. Stand-level TRW chronologies are built with cores collected from 20-30 canopy-dominant trees at each stand, visually and statistically crossdated at the tree and stand-level and compared with the overlapping Landsat data. Time-series approaches, including different standardization and detrending methods, are applied to both the TRW chronologies and Landsat time-series for comparison. Analysis is ongoing, but preliminary results indicate that TRW and Landsat %CC follow similar temporal patterns of growth and decline, at either a 0- or 1-year lag for TRW. Furthermore, the most notable narrow marker year (1988) from TRW aligns with an exceptionally low productivity year in the Landsat record.