Defense Presentation Title: Betting the Farm: Does Homogenization Threaten Smallholders’ Climate Resilience?

Advisor: Dr. Josh Gray, faculty fellow and associate professor in the Department of Forestry and Environmental Resources

Abstract: Smallholder farms (<2 ha) contribute 80% of the food supply in Asia and Sub-Saharan Africa. Their ability to produce high yields will largely determine the success of the United Nations’ Sustainable Development Goals (SDGs) to eliminate poverty and hunger. Fortunately, recent advancements in remote sensing and machine learning enable us to assess progress toward the SDGs at broad spatial scales, even when ground truth data are sparse. The first part of this dissertation covers two reproducible and scalable strategies to monitor changes in smallholder communities: 1) automatically delineating fields using publicly available satellite images and AI; and, 2) augmenting public satellite data with minimum amounts of commercial satellite data, identified a priori, to monitor crop growth at the smallholder field level. Next, the delineated fields and their crop growth metrics are paired with household survey data to assess whether homogenized smallholder communities – in which farmers have adopted the same management practices – are more vulnerable to climate threats than their homogeneous counterparts, posing a risk to their long-term stability in food security.