A Productive Friction: Leveraging Misalignments Between Local Ecological Knowledge and Remotely Sensed Imagery for Forest Conservation Planning

Abstract Both satellites and nearly 2 billion forest-proximate peoples continually monitor Earth’s forests. Generally, the satellite record summarizes production estimates, such as percent tree cover, at regular, relatively coarse scales. Conversely, local perceptions tend to capture changes at irregular and very fine scales. While the utility of both of these sources of information has been widely demonstrated in isolation, little work has explored how they might be systematically and quantitatively integrated. Here, we collect gridded information on community perceived and remotely sensed mangrove cover change across 719 0.5-km grids in Pemba Island, Tanzania. We reveal variation in the association between these two data sources across different wards ( shehia ) and explore the reasons for this variation using interviews and direct observation. We find that shehia with the greatest alignment between perceived and remotely sensed mangrove change tended to have little planting or natural regeneration of mangrove propagules and large areas of complete cover loss. Alternatively, in shehia with the lowest alignment, we find high levels of natural and/or human-assisted mangrove recolonization and selective harvesting of individual trees and branches. Finally, we demonstrate a practical workflow for quantitatively leveraging these misalignments by optimizing across both data sources to identify restoration priority areas.