Canadian Forest Service Publications

Boreal Shield forest disturbance and recovery trends using Landsat time series. 2015. Frazier, R.J.; Coops, N.C.; Wulder, M.A. Remote Sensing of Environment 170. pp 317–327

Year: 2015

Issued by: Pacific Forestry Centre

Catalog ID: 36526

Language: English

Availability: PDF (download)

Available from the Journal's Web site.
DOI: 10.1016/j.rse.2015.09.015

† This site may require a fee

Mark record


Monitoring forest recovery following disturbance is important for both forest management as well as assessing possible climate change impacts on forest dynamics. To do so, an improved understanding of forest recovery processes and their relationship to remotely sensed spectral measures of recovery is required. Our objective in this research is to develop and apply a methodology for using Landsat time series to characterize forest recovery using spectral recovery trajectories. We focus our efforts in the Canadian Boreal Shield ecozone where a known geographic east to west distinction in disturbance regimes remains to be quantified. Results show that forest recovery following a stand replacing disturbance is detectable and quantifiable using a dense Landsat time series of spectral reflectance. All Tasseled Cap indices were found to capture an element of forest recovery following disturbance, with Wetness offering additional information on increasing vegetation structure and complexity. Tasseled Cap component trajectories of recovery show clear differences in both disturbance detection and forest recovery across the east and west Boreal Shield sections. The Cohen's d similarity metric indicated large differences (d N .08) in Wetness and Greenness-based spectral recovery trajectories when comparing the two Boreal Shield sections with the East Boreal Shield having markedly more above average recovery (+2 std. dev. from the mean) than theWest. Based on our spectral recovery results,we also observe that forest recovery varies over the entire ecozone and is different between the east and west Boreal Shield forests.

Plain Language Summary

We develop and apply a Landsat time series methodology for spectral recovery characterization. We construct and characterize spectral forest recovery trajectories in four spectral indices. East and West Boreal Shield sections disturbance/recovery trends are compared. Different disturbance and recovery trends exists in the East and West Boreal Shield sections. Selection of spectral index affects the characterization of recovery through time. Given sufficient time trees generally return following disturbance, and as such, recovery is better understood as a process than a state. Better understanding of how remotely sensed spectral measures can provide information on the forest recovery processes is required. All indices are found to capture the return of vegetation post-disturbance, with additional information on increasing vegetation content and complexity offered via indices using longer optical wavelengths.