A new machine-learning system developed at the University of Alaska Fairbanks aims to automatically produce detailed maps from satellite data to show locations of likely beetle-killed spruce trees in Alaska. This automated process can assist forestry and wildfire managers in their decision-making, especially as the spruce beetle infestation spreads. The identification system, designed by assistant professor Simon Zwieback, fills a knowledge gap by mapping likely spruce beetle infestations in areas of low to moderate severity. Existing methods rely on expert observations from airplanes, which are expensive and limited in space and time, hindering stakeholders’ ability to respond effectively to the outbreak.
Alaska foresters currently rely on survey flights, manual interpretation of high-resolution imagery, and automated analysis of coarser satellite imagery to identify dead spruce trees in mixed forests. Zwieback’s innovative method combines the efficiency of automation with the detail of high-resolution satellite images, utilizing machine learning techniques to recognize dead spruce based on their characteristic shape, color, and contextual clues. The system has been successfully tested in a study area heavily affected by a beetle infestation, demonstrating its ability to identify dead spruce even in stands containing only a few affected trees, where traditional methods may fall short.
The widespread infestation in Alaska has affected approximately 2 million acres, predominantly in Southcentral Alaska, since 2016, with the beetle spread northwards to Cantwell and the Alaska Range mountains by 2020. The death of large numbers of spruce trees leads to various ecosystem changes, such as alterations in understory vegetation, an increase in ground-level fuel due to dead branches, and heightened wildfire danger. Zwieback’s method can be valuable in decisions related to fire prevention and suppression, as well as in understanding the outbreak dynamics and informing responses as it migrates into the Interior region.
Looking ahead, Zwieback aims to implement his system for the entire state of Alaska to provide ongoing monitoring and analysis of new satellite images to track the outbreak dynamics. The research was funded by NASA EPSCoR and the National Science Foundation EPSCoR, highlighting the importance of remote sensing technology in understanding and addressing forest health issues. Field sites have been established on Ahtna-owned land in Southcentral Alaska to better understand the progress and consequences of the outbreak as it moves into the Interior. By combining machine learning with high-resolution imagery, Zwieback’s method offers a valuable tool for forest and wildfire management in Alaska, helping to mitigate the impacts of the spruce beetle infestation and protect the state’s valuable resources and landscapes.