Spatio-temporal dynamics of water stress for common bean production in Goiás, Brazil

Abstract

Water stress is among the most critical abiotic factors limiting agricultural productivity. Adjusting sowing dates to align with periods of higher water availability is a proven strategy for mitigating drought impacts on common bean production. This study aimed to evaluate the dynamics of water stress in common bean production regions across Goiás, Brazil, during the rainfed wet and dry seasons, utilizing functional data analysis (FDA). The CSM-CROPGRO-Dry Bean model was employed to simulate common bean yields and environmental factors (EF), including water stress. Simulations incorporated a dataset encompassing soil characteristics, daily climate data (1980–2016), management practices, and genetic coefficients. Water stress data were analyzed by municipality, sowing date, and year, focusing on the critical period from 20 days before flowering to 40 days after flowering. Accumulated water stress curves were generated and grouped using functional K-means clustering for each sowing date. Based on these clusters, regions were characterized by water stress levels, and a spatio-temporal sowing calendar was developed. The results indicated that during the wet season, the optimal sowing period is between October 20 and November 10, when water stress is minimal. Conversely, the dry season exhibited the highest levels of water stress, typically beginning before flowering and intensifying through grain filling. To mitigate these effects, sowing between January 1 and February 10 is recommended. By assessing the spatio-temporal dynamics of water stress, this study offers actionable insights to optimize sowing schedules, enhance yield potential, efficiently manage water resources, and promote sustainable common bean production in Goiás.