2026-07-27 ミシガン大学
＜関連情報＞
- https://news.umich.edu/from-up-on-high-a-view-from-space-for-a-future-for-small-farms/
- https://www.science.org/doi/10.1126/science.aee1344
衛星データは食料システムの変革に役立つ可能性がある
衛星データは、小規模農家の農業システムにおける収穫量の増加と持続可能性の向上に活用できる
Satellite data can help transform food systems
Satellite data can be used to increase yields and improve sustainability in smallholder agricultural systems
Meha Jain
Science Published:5 Feb 2026
DOI:https://doi.org/10.1126/science.aee1344
Climate change is disrupting biological systems across the globe, from the bleaching of coral reefs to shifting habitat ranges. Agricultural ecosystems essential to humanity are projected to be one of the systems most negatively affected by climate change. Simultaneously, climate change is exacerbating the environmental impacts of agriculture (1). Actionable solutions that reduce the effects of climate change, increase food production, and enhance environmental sustainability are urgently needed.
My research program examines how we can produce enough food to sustainably feed humanity in the face of climate change. I focus my research on smallholder agroecosystems across the Global South, which are facing some of the greatest food production pressures and are the backbone of millions of rural livelihoods. Historically, it has been difficult to develop and deploy effective solutions at scale in agricultural systems across the Global South because these systems are extremely heterogeneous, with solutions that are the most effective in one location being suboptimal or maladaptive in other regions. At the same time, these systems are typically data scarce, making it challenging to identify contextually successful strategies and to develop data-informed interventions and policies.
To overcome these challenges, my research program develops methods that use state-of-the-art satellite sensors and artificial intelligence to create field-level datasets on farm management and outcomes in data-scarce smallholder systems across the Global South. These datasets allow me to understand real-world farm management and its impacts on both crop production and the environment, identifying the most effective ways to sustainably increase production at the landscape scale. Although researchers have used Earth observation satellites to map agricultural characteristics for decades, it has historically been challenging to produce such datasets in smallholder systems given the small size of fields, the high heterogeneity in farm management and outcomes, and the lack of ground data for training and validating models. By leveraging new high-resolution satellite sensors and methods that I developed that use limited to no ground data for calibration, I overcame these challenges to produce critical datasets in locations and time periods that have never been measured.
