COS 199-3 - Moving towards forecasting forest canopy recovery rates and trajectories following invasive insect pest defoliation events

Abstract:

Background: Insect pests are some of the most destructive disturbance agents in forests worldwide, yet there remains a lack of predictive modeling frameworks for understanding their impacts in space and time, particularly at the landscape scale. In recent years there has been increased interest in quantifying the consequences of forest pest outbreaks, particularly in the face of rapidly changing climate regimes and at scales relevant to managers. Here we demonstrate the use of remote sensing to characterize the disparate recovery trajectories that arise between sites with and without tree mortality. To do this, we leverage observations from a Landsat-based forest condition monitoring product that provides near-real-time estimates of canopy greenness to model forest canopy recovery across southern New England following an outbreak of Lymantria dispar, a defoliation moth species. We hypothesized that we would observe two distinct recovery trajectories following a widespread L. dispar outbreak dependent on whether or not a site experienced canopy tree mortality. Furthermore, we hypothesized that at the landscape scale, recovery rates at our study sites will differ according to demographic, environmental, and edaphic factors. To test our hypotheses, we use a process-based approach to model forest recovery following the 2016 L. dispar defoliaiton event across Massachusetts, Connecticut, and Rhode Island.



Results: We found a positive relationship between the magnitude of L. dispar defoliation and rate of recovery following outbreaks. In addition to the magnitude of disturbance, we tested a suite of environmental variables to identify key drivers of recovery rates following defoliation and determine contributing factors to recovery success. In the summer of 2022, we conducted surveys of 205 field plots in central Massachusetts originally established in 2017 to assess tree mortality and understory demographics following a 2016-2017 L. dispar outbreak across a mixed hardwood forest. We used these field data to validate our model outputs generated from 5000 sample sites across the three states encompassing our study area. Our results suggest that tree mortality creates divergent recovery trajectories that can be accounted for in models. We further suggest these results could be useful for providing management recommendations for assisted recovery of oak trees following L. dispar outbreaks.