Changes of agroclimate conditions and induced yield variability in the breadbaskets of the world

Abstract:

Global warming is changing both daily weather and long-term climate. In a connected world, climate shifts in any breadbasket of the world, particularly at the daily timescale, can threaten global food security. While consensus results have been reached regarding the impacts of temperature changes on crop yields over these regions, it is not known how agriculture-significant climate conditions, as derived from daily weather, are changing and how are these changes contributing to the changes of crop yields. In this study, we analyze the changes in 11 agroclimate indices, as indicators of agriculture-significant climate conditions, for four globally important crops: maize, rice, soybeans, and wheat, as well as contributions of these indices to the changes in crop yields over each breadbasket region during the past decades. We derive the agroclimate indices for each crop growing season at each grid point over the global cropland, including: last spring frost, growing season length, the start of field operations, field conditions suitable for plating, growing degree days, growing season stress, field conditions suitable for harvest, first fall frost, accumulated frost days, growing season total precipitation, and growing season dry days. Daily gridded temperature, precipitation, and soil moisture from various sources are used for calculating these agroclimate indices. We will present trends and variations of these agroclimate indices at different locations over the global cropland, their relative contributions to the crop yields over different breadbaskets of the world as revealed by machine learning models, and discuss potential strategies to intervene or harness these changes. Our observational analysis based on multiple daily agroclimate datasets includes sources of uncertainty in the impact assessments and informs tailored climate adaptation strategies to ensure food security in a changing climate.