Deforestation
in the Brazilian Amazon continues at alarming rates, with large areas
of forest still being converted into pasture for cattle ranching.
Despite a wide range of policy interventions, including the Brazilian
Forest Code, intensification incentives, and supply-chain regulations,
effective monitoring of cattle production remains a critical gap. Recent
advances in artificial intelligence (AI) and very high-resolution
satellite imagery open new possibilities to directly observe cattle at
scale.
In this talk, Hodel
presents an AI-based approach to detect and count cattle from satellite
images and explore what these data can reveal about land-use dynamics
at the deforestation frontier. She will demonstrate how cattle-level
monitoring can be used to assess intensification patterns, detect
grazing incursions into protected areas and Indigenous lands, and
identify herds that fall outside existing supply-chain monitoring
systems.
Beyond
technical performance, she discusses how such data can be meaningfully
integrated into current policy frameworks to support enforcement and
transparency, while shedding light on the structural barriers to a just
transition toward more sustainable and inclusive land-use systems in the
Amazon.