Outbreaks of insect pests periodically cause large losses of volume in Canada’s forests. Compounded with climate change, outbreaks create significant challenges for managing the sustainable delivery of ecosystem services. Current methods to monitor damage by these pests involve both field and aerial surveys. While relatively cost effective and timely, aerial survey consistency and spatial coverage may be insufficient for detailed monitoring across Canada’s vast forest-land base. Remote sensing can augment these methods and extend monitoring capabilities in time and space by incorporating knowledge of pest-host interactions and of how damage translates into a remote sensing signal for detection and mapping. This review provides a brief introduction to major forest insect pests in Canada (two bark beetles (Coleoptera: Curculionidae) and six defoliators) and the damage they cause, a synthesis of the literature involving aerial survey and remote sensing, and a discussion of how these two approaches could be integrated into future pest monitoring from a Canadian perspective. We offer some lessons learned, outline roles that remote sensing could serve in a management context, and discuss what ongoing and upcoming technological advances may offer to future forest health monitoring.

The oil and gas industry has grown significantly throughout the boreal and arctic ecosystems of North America. A major feature of the ecological footprint of oil and gas exploration is seismic lines—narrow corridors used to transport and deploy geophysical survey equipment. These lines, which traverse forests, tundra, uplands, and peatlands, were historically up to 10 m wide. Over the past decade, seismic lines have decreased in width (in some cases down to 1.75–3 m); however, their density has increased drastically and their construction is expected to continue in regions of Canada and the United States that are rich in oil and gas resources. We examine the literature related to the environmental impacts of, and restoration and reclamation efforts associated with, seismic lines in the boreal and arctic ecosystems of North America. With respect to conventional seismic lines, numerous studies report significant and persistent environmental changes along these lines and slow recovery of vegetation that translates into a lasting fragmentation of the landscape. This fragmentation has many ramifications for biodiversity and ecosystem processes, including significant implications for threatened woodland caribou herds. While modern, low-impact seismic lines have comparatively lower ecological effects at the site-level, their high density and associated potential edge effects suggest that their actual environmental impact may be underestimated. Seismic line restoration is a critical aspect of future integrated landscape management in hydrocarbon-rich regions of the boreal-arctic, and if widely applied, has the potential to benefit a wide range of species and maintain or re-establish key ecosystem services such as carbon sequestration and biodiversity.

Monitoring tree regeneration in forest areas disturbed by resource extraction is a requirement for sustainably managing the boreal forest of Alberta, Canada. Small remotely piloted aircraft systems (sRPAS, a.k.a. drones) have the potential to decrease the cost of field surveys drastically, but produce large quantities of data that will require specialized processing techniques. In this study, we explored the possibility of using convolutional neural networks (CNNs) on this data for automatically detecting conifer seedlings along recovering seismic lines: a common legacy footprint from oil and gas exploration. We assessed three different CNN architectures, of which faster region-CNN (R-CNN) performed best (mean average precision 81%). Furthermore, we evaluated the effects of training-set size, season, seedling size, and spatial resolution on the detection performance. Our results indicate that drone imagery analyzed by artificial intelligence can be used to detect conifer seedling in regenerating sites with high accuracy, which increases with the size in pixels of the seedlings. By using a pre-trained network, the size of the training dataset can be reduced to a couple hundred seedlings without any significant loss of accuracy. Furthermore, we show that combining data from different seasons yields the best results. The proposed method is a first step towards automated monitoring of forest restoration/regeneration.