1 Socio-Environmental Drivers of Forest Change in Rural Uganda Maia Call 1,2 , Tony Mayer 1,3 , Samuel Sellers 1,3 , Diamond Ebanks 1,3 , Margit Bertalan 1,3 , Elisabeth Nebie 1,4 , and Clark Gray 2 Paper for presentation at the 2016 Annual Meeting of the Population Association of America 1 Authors contributed equally to this research. 2 Department of Geography, University of North Carolina at Chapel Hill 3 Curriculum for the Environment and Ecology, University of North Carolina at Chapel Hill 4 Department of Anthropology, University of North Carolina at Chapel Hill Abstract Tropical deforestation and forest degradation remain among the top global threats to biodiversity, carbon storage and rural livelihoods, but the social processes underlying these changes remain difficult to observe across large spatial scales and in data-poor contexts such as tropical Africa. We link population-representative longitudinal survey data from agricultural households in rural Uganda to high-resolution satellite data on forest cover change, and use this linked dataset to investigate processes at two scales: tree planting and harvesting at the parcel scale, and deforestation and reforestation at the community scale. This multi-scale analysis reveals that tree planting is more common on parcels with secure tenure, by educated heads and in isolated communities. Deforestation is highest in wealthier, more accessible communities with low population density and high baseline forest cover. These results provide explicit evidence that the social drivers of forest change in Uganda vary across scales, indicating a need for additional multi-scale studies in land use change science. Introduction The social drivers of forest change have been a major focus of study in the human- environment sciences for at least three decades, and will likely remain so given ongoing processes of tropical deforestation and forest degradation (Hansen, Stanford). Over that time, a clear consensus has emerged that forest change must be understood as a social processes with drivers at multiple spatial scales (Lambin), and a growing number of studies have explicitly examined the multi-scale drivers of tropical deforestation (Gray, Vance, Pan). However, despite these advances, most studies remain bound to a particular spatial scale of forest change (the pixel), ignoring sub-pixel processes such as tree planting and harvesting that can have important implications for rural livelihoods and carbon storage (###). Meanwhile, a distinct literature on smallholder tree planting has evolved in isolation from these debates, focusing on technical and legal approaches to promote tree planting (###). These issues are of particular importance in East Africa and in our study country, Uganda. A majority of the highly biodiverse forests of East Africa have already been lost to