Estimating crop yield potential at regional to national scales

World population will increase 35% by 2050, which may require doubling crop yields on existing farm land to minimize expansion of agriculture into remaining rainforests, wetlands, and grasslands. Whether this is possible depends on closing the gap between yield potential (Yp, yield without pest, disease, nutrient or water stresses, or Yw under water-limited rainfed conditions) and current average farm yields in both developed and developing countries. Quantifying the yield gap is therefore essential to inform policies and prioritize research to achieve food security without environmental degradation. Previous attempts to estimate Yp and Yw at a global level have been too coarse, general, and opaque. Our purpose was to develop a protocol to overcome these limitations based on examples for irrigated rice in China, irrigated and rainfed maize in the USA, and rainfed wheat in Germany. Sensitivity analysis of simulated Yp or Yw found that robust estimates required specific information on crop management, +15 years of observed daily climate data from weather stations in major crop production zones, and coverage of 40–50% of total national production area. National Yp estimates were weighted by potential production within 100-km of reference weather stations. This protocol is appropriate for countries in which crops are mostly grown in landscapes with relatively homogenous topography, such as prairies, plains, large valleys, deltas and lowlands, which account for a majority of global food crop production. Results are consistent with the hypothesis that average farm yields plateau when they reach 75–85% of estimated national Yp, which appears to occur for rice in China and wheat in Germany. Prediction of when average crop yields will plateau in other countries is now possible based on the estimated Yp or Yw ceiling using this protocol.