Dendrocalamus brandisii is a dual-purpose bamboo species valued for shoots and timber. Its culms possess superior physical and chemical properties for industrial applications, and its shoots command market prices more than five times higher than comparable products, providing substantial economic benefits. To estimate the growth potential of D. brandisii and guide plantation management, this study compiled environmental datasets—including soil, climate, and topography—from major production regions across Yunnan, Guizhou, Guangdong, Guangxi, Fujian, and Hainan. Among machine learning approaches, the random forest model was selected to develop a biomass prediction model. Model results showed that soil variables—soil organic carbon, cation exchange capacity, bulk density, soil texture, pH, and nitrogen content—were the dominant drivers of biomass variation, followed by topographic and climatic factors including upslope position, aspect, and bio5. The relationships between these variables and biomass accumulation were examined using partial dependence analysis, which revealed response curves across different environmental gradients. Based on biomass, a site quality evaluation system was established for D. brandisii , and site quality classes were mapped accordingly. The results indicated a clear association between site quality and elevation, with high-quality sites concentrated in areas of moderate altitude. Therefore, valley regions are recommended for planting in plateau–mountain areas, whereas relatively higher hilly and mountainous zones are suitable in coastal regions. By quantifying environmental impacts on D. brandisii productivity and classifying site quality levels, this study provides practical guidance for regional plantation planning and offers support for improving the economic benefits of the D. brandisii industry. • Covers main Dendrocalamus brandisii plantations in China, largest study area so far. • Soil has the greatest impact; climate and terrain are less important. • Model predicts growth potential and supports site quality classification. • The correlation and response pattern of cation exchange capacity are different.
Wang et al. (2026) studied this question.