Global vulnerability of agricultural commodities to climate risk: Evidence from satellite data

Abstract

Using a new dataset of high-resolution satellite readings of soil moisture in conjunction with climate-economy econometric models and machine learning methods, we quantify the risk to global agricultural commodities from climate change. Our climate risk measure based on soil moisture fluctuations has a significantly negative impact on the yield of major cereals, and the effect is greater than temperature and precipitation. This impact concentrates on countries with lower income levels, harsher climate conditions, or less irrigated farmland. Moreover, we show agricultural commodities are less vulnerable to current drought risks in countries with higher frequencies of past droughts, suggesting that countries adapt to climate change through learning from prior experience.