A productive friction: Leveraging misalignments between local ecological knowledge and remotely sensed imagery for forest conservation planning

Earth's forests are continually monitored by both the satellite record and the lived experiences of nearly 2 billion forest‐proximate peoples. 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 in what contexts they are expected to correlate or deviate, or how they might be 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. These findings indicate that the alignment between local knowledge and satellite observations of mangrove cover change systematically increases with the scale of change in this system. Finally, we demonstrate a practical workflow for quantitatively leveraging these misalignments by optimizing across both data sources to identify restoration priority areas. This study assesses the social–ecological factors that drive alignment between local ecological knowledge and satellite‐observed measures of mangrove cover change in Pemba, Tanzania. Results show that alignment between these two data sources, or lack thereof, is driven by differences in selective harvesting, mangrove recolonization, and the prevalence of clear cutting. These findings are used to suggest an optimization approach for identifying restoration priority areas using both data sources in combination.