Impact of green water anomalies on global rainfed crop yields
Edoardo Borgomeo, Hassaan Furqan Khan, Matias Heino, Esha Zaveri, Matti Kummu, Casey Brown, Anders Jägerskog
Abstract
Abstract The importance of green water (moisture from rain stored in soils) for global food and water security is widely recognized, with process-based simulation models and field-level studies demonstrating its role in supporting rainfed agriculture. Despite this evidence, the relationship between green water anomalies and rainfed agriculture has not yet been investigated using statistical models that identify a causal relationship between the variables. Here, we address this gap and use disaggregated statistical regression (panel data analysis) at the 30 arc-min grid level to study the response of observed yields (1982–2010) of four main crops (maize, rice, soybean and wheat) to green water anomalies globally over rainfed areas. Dry green water anomalies (1 or 2 standard deviations below long-term average) decrease rainfed crop yields worldwide. This effect is more pronounced for wheat and maize, whose yields decline by 12%–18% and 7%–12% respectively. Globally, agricultural production benefits from wet green water anomalies. This effect is intensified in arid climates and weakened in humid climates where, for wheat, soybean and rice, periods of green water availability 2 standard deviations above long-term averages lead to declines in crop yield. This confirms existing evidence that excess soil moisture is detrimental to crop yield. These findings (1) advance our understanding of the impact of green water on rainfed food production and (2) provide empirical evidence supporting arguments for better management of local green water resources to reduce the impact of agricultural drought and waterlogging on rainfed crop production and capture the yield increasing effects of positive green water anomalies.