ABSTRACT
This research compares the capabilities of various Sentinel-2-derived spectral vegetation indices (SVIs) in particular red-edge SVIs to detect and classify spruce budworm (Choristoneura fumiferana) (SBW) defoliation using Support Vector Machine (SVM) and Random Forest (RF) models. The results showed the superiority of RF in model building for defoliation detection and classification into three classes (nil, light, and moderate) with overall errors of 17% and 32%, respectively. The most important variables for the best model were Enhanced Vegetation Index 7 (EVI7), Modified Chlorophyll Absorption in Reflectance Index (MCARI), Inverted Red-Edge Chlorophyll Index (IRECI), Normalized Difference Infrared Index 11 (NDII11) and Modified Simple Ratio (MSR). Red-edge SVIs were more effective variables for light defoliation detection compared to traditional SVIs such as Normalized Difference Vegetation Index (NDVI) and EVI8. These findings can help improve current remote sensing-based SBW defoliation detection and monitoring.
Acknowledgments
The authors would like to thank New Brunswick Department of Energy and Resource Development for providing defoliation field data for this project.
Declaration of interest statement
The authors declare no conflicts of interest.