As climate change warms winters in North America, the range of an invasive pest, the Hemlock Woolly Adelgid (Adelges tsugae: HWA), is expanding into the eastern hemlock (Tsuga canadensis) dominated forests in New England and southern Canada. Infestation leads to the death of individual hemlock trees within 5 to 15 years, removing this foundation species from the landscape, and resulting in a transition of nearly all eastern hemlock dominated forests in North America into mixed deciduous forests.
The scale of the HWA disturbance lends itself to regional-scale monitoring efforts with spaceborne remote sensing. The Global Ecosystem Dynamics Investigation (GEDI) mission, planned for launch in 2019, will characterize forest structure across a landscape and record data from the stratified layers of the forest canopy. It may be able to capture structural changes brought about by disturbances such as the HWA and provide snapshots of the spread of such infestations over continental areas This project explicitly explores the ability of large-footprint lidar scanners to discern stages of HWA disturbance. This study uses simulated GEDI waveforms from NSF NEON and NASA G-LiHT airborne lidar data, acquired over infested forest stands, and relates structural metrics from the waveforms to hemlock mortality.
Results/Conclusions
Hemlock tree mortality data were collected over HWA infested areas within the Smithsonian Institute ForestGEO plot at Harvard Forest in central Massachusetts during the summer of 2016. A trend surface of hemlock tree mortality between 2010 and 2016 within the area revealed a strong gradient, with southern plots showing a maximum of 31% hemlock mortality and northern plots showing 0% mortality.
Large-footprint waveform lidar data were simulated over the area using the GEDI Simulator, developed by Steven Hancock at the University of Maryland. These lidar waveforms were simulated from discrete lidar data collected by NASA G-LiHT in 2012 and NSF NEON four years later in summer 2016. The change in structural metrics between 2012 and 2016 was explored with relative heights of energy percentiles (RH10, 25, 50, 75, and 90) calculated for each waveform. These metrics were compared to the mortality of hemlock trees within the lidar footprint.
Initial results showed an increase in the intensity of reflectance of the ground and near-ground portion of lidar waveforms from 2012 to 2016. This could indicate that the HWA infestation damaged hemlock canopies from 2012 to 2016, which allowed more lidar energy to reach the forest floor in 2016.