Optimizing rice ( Oryza sativa L.) production in Brazil through climate‐driven nonlinear modeling

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Understanding how climate variability influences crop yield is essential for improving rice production, particularly in a country like Brazil with diverse agroclimatic zones. This study aimed to identify the key climatic drivers affecting lowland rice ( Oryza sativa L.) yields and to delineate optimal production areas using nonlinear modeling. We integrated 21,749 records from multiple‐environment trials conducted between 1982 and 2018 with large‐scale climate data spanning five Brazilian regions. To capture the nonlinear relationships between yield and environmental variables, generalized additive models (GAMs) were fitted separately for each region. As a novel contribution, we identified region‐specific climatic drivers, such as growing degree days, precipitation, temperature, humidity, and altitude, and applied the potential of mean increment (PMI) metric to pinpoint zones where yields are most likely to benefit from favorable climatic conditions. Results showed that temperature consistently influenced yield across all regions, though specific thresholds varied. For instances, in the South, high grain‐filling temperatures reduced yield, while in the North, higher minimum temperatures at flowering improved outcomes. Optimal areas were identified where both high PMI and observed yields coincided, offering spatially explicit targets for breeding and management. These findings highlight the utility of GAMs in capturing complex genotype–environment interactions and guiding climate‐smart agricultural practices. The approach informs regionalized strategies for crop calendar adjustment and ideotype development, enhancing climate resilience in rice production systems. By uncovering “environmental rules” for yield optimization, this framework contributes valuable insights for breeders, agronomists, and policymakers seeking sustainable intensification under climate uncertainty.
Funding Information
  • Fundação de Amparo à Pesquisa do Estado de Goiás (202310267000216)
  • Conselho Nacional de Desenvolvimento Científico e Tecnológico (4/2021 ‐ 310209/2021‐8)

This publication has 24 references indexed in Scilit: