Abstract
Stalk lodging poses a significant global threat to agricultural productivity and food security in important cereal crops like maize and sorghum. This dissertation addresses the challenge by advancing multi-scale biomechanical understanding and developing robust phenotyping techniques to enhance stalk lodging resistance. At the cellular level, a semi-automated computational pipeline was developed for high-throughput digitization and finite element modeling of plant cell walls, which revealed the significant effect of cell wall thickness on tissue stiffness. Analyses across the tissue and whole-plant levels identified structural rather than material failure as the primary cause of stalk lodging, even under post-flowering water stress conditions. This highlights the importance of stalk geometry over material strength in breeding for lodging-resistant stalk varieties. A high-throughput biomechanical phenotyping pipeline was implemented on over 30,000 stalks to generate a large paired dataset of lodging-related traits across a diverse stalk population. This dataset enabled genomic prediction models that demonstrated the significance of multi-environment approaches in improving predictive accuracy for stalk lodging resistance. Finally, evaluation of X-ray diffraction methods for microfibril angle measurement in cereal stalks revealed limitations, including systemic biases and an inability to capture relevant biological variation, cautioning against their direct application without crop-specific calibration. Results reveal that effective lodging resistance strategies should incorporate the complex interplay involving multiscale stalk architecture, environmental conditions, and genetic factors. Collectively, this dissertation provides resourceful methodologies and identifies key stalk bending strength determinants that inform targeted breeding strategies and contribute to improving crop productivity.