Fusion of Sentinel-1B and Sentinel-2B Data for Forest Disturbance Mapping: Detection of Bark Beetle Mortality in the Southern Sierra Nevada

Abstract

Since 2010 approximately 129 million trees in California have been killed by bark beetle attacks. The detection and mapping of bark beetle infestations is an important component for monitoring overall forest health and sustainability. Using remote sensing techniques and European Space Agency (ESA) Sentinel-1 and Sentinel-2 imagery, bark beetle infestation in the southern Sierra Nevada range were mapped. Sentinel-1 Synthetic Aperture Radar imagery was co-registered with Sentinel-2 multispectral imagery and the machine learning algorithms Random Forest and Support Vector Machines were utilized in a supervised classification scheme on two user generated training and validation data samples. The observed accuracy for this analysis ranged from 79.1% to 88.7%. As climate change intensifies, it can be expected that the extent and severity of epidemic bark beetle mortality will continue to spread in western North America. The investigation of bark beetle outbreaks and subsequent impacts is an important component to monitoring overall forest health and sustainability.