Trait-based screening of heterogeneous culms to support selection and allocation in bamboo-based restoration

This is the authors' abstract. We don't add key points for this paper.

In bamboo-based restoration systems, matching harvested culms to appropriate end uses is complicated by strong culm-to-culm variability and by grading practices that remain uneven or context-specific. This study examined whether measured culm descriptors could predict bending strength (BS), modulus of elasticity (MOE), and embodied carbon (EC) as a basis for culm-level assessment particularly for structural applications and lower-carbon allocation decisions. Using 102 culm specimens of Bambusa blumeana and Bambusa vulgaris tested under ISO 22157-type bending procedures, culm descriptors were related to BS, MOE, and EC through an interpretable, data-driven screening framework. Predictive performance was high across all three indicators (R² = 0.96–0.98), enabling model-based translation of culm-level heterogeneity. Key results show that favorable structural and carbon outcomes do not arise from a single universal trait but instead reflect different trait combinations and trade-offs across indicators. Within the study’s dataset, culms approximately 80–90 mm in diameter and 1.6–2.1 kg/m in linear mass tended to combine stronger predicted structural performance with lower embodied carbon. While denser culms (about 725 kg/m³) with relatively thin walls (<8 mm) were associated with favorable stiffness in specific prediction regimes. By translating trait-level variability into candidate allocation cues, the framework provides a transparent basis for assessing heterogeneous bamboo stocks and guiding decisions that direct culms toward structural and lower-carbon uses, thereby supporting restoration objectives in managed bamboo systems.