Water productivity of rainfed maize and wheat: A local to global perspective

Water productivity (WP) is a robust benchmark for crop production in relation to available water supply across spatial scales. Quantifying water-limited potential (WPw) and actual on-farm (WPa) WP to estimate WP gaps is an essential first step to identify the most sensitive factors influencing production capacity with limited water supply. This study combines local weather, soil, and agronomic data, and crop modeling in a spatial framework to determine WPw and WPa at local and regional levels for rainfed cropping systems in 17 (maize) and 18 (wheat) major grain-producing countries representing a wide range of cropping systems, from intensive, high-yield maize in north America and wheat in west Europe to low-input, low-yield maize systems in sub-Saharan Africa and south Asia. WP was calculated as the quotient of either water-limited yield potential or actual yield, and simulated crop evapotranspiration. Estimated WPw upper limits compared well with maximum WP reported for field-grown crops. However, there was large WPw variation across regions with different climate and soil (CV = 29% for maize and 27% for wheat), which cautions against the use of generic WPw benchmarks and highlights the need for region-specific WPw. Differences in simulated evaporative demand, crop evapotranspiration after flowering, soil evaporation, and intensity of water stress around flowering collectively explained two thirds of the variation in WPw. Average WP gaps were 13 (maize) and 10 (wheat) kg ha^-1 mm^-1, equivalent to about half of their respective WPw. We found that non-water related factors (<i>i.e.</i>, management deficiencies, biotic and abiotic stresses, and their interactions) constrained yield more than water supply in <i>ca.</i> half of the regions. These findings highlight the opportunity to produce more food with same amount of water, provided limiting factors other than water supply can be identified and alleviated with improved management practices. Our study provides a consistent protocol for estimating WP at local to regional scale, which can be used to understand WP gaps and their mitigation.