Jennifer Burney (UCSD), Craig McIntosh (UCSD), Bruno Lopez-Videla (UCSD), Krislert Samphantharak (UCSD), Alexandre Gori Maia (Unicamp)

Proceedings of the National Academy of Sciences (PNAS)


Climate change can affect agriculture across levels–from plants to farms to institutions–but these impacts are difficult to measure and project consistently. We propose a statistical approach for estimating the sensitivity of agricultural systems to different dimensions of climate change and modeling future shifts that incorporate human adaptation. Applying this in the Brazilian context reveals that climate has a powerful effect on yields and agricultural revenues and drives default for a large public sector bank. Future projections suggest increased yield and revenue volatility at midcentury, along with higher rates of climate-driven default that create correlated risks for financial institutions. This approach is thus able to capture often hard-to-model emergent climate risks and inform more tailored approaches to building resilience.