Predicted Tree Canopy Cover (1972 - 2020) from: Using Landsat time-series to investigate nearly 50 years of tree canopy cover change across an urban-rural landscape

Paper Abstract: Canadian urban and adjacent landscapes have undergone dynamic changes over the last 50 years due to land management practices, land cover alternations, climate change, and various disturbances. Remote sensing, particularly the Landsat archive, provides the only means to spatially quantify these long-term dynamics at a local scale. Here, we explore how Landsat, including the often-forgotten MSS sensor, can be used to investigate tree canopy cover (CC) change over nearly 50 years (1972-2020) in a Canadian urban-rural landscape. We built a CC time-series by training random forest models using visually interpreted CC from reference high-resolution imagery. Predictors included topographic and yearly LandTrendr-fitted tasseled cap indices. Harmonized tasseled cap indices were available throughout the full Landsat archive by applying LandsatLinkr. Yearly binary tree canopy maps were also built to mask consistent non-canopy areas and limit noise. To increase confidence in observed CC change in the absence of historical reference imagery, we consider multiple temporal validation options. We explore changes across broad landscape types and build connections with different influential drivers (e.g., agricultural reforestation, housing development, emerald ash borer, ice storms). Results demonstrate how long-term Landsat time-series can be used to better understand historical tree canopy change, providing important information for local organizations. Dataset details: See paper.