Using coupled soil nutrients as dummy variables can improve the predictive performance of stand basal area growth model for subtropical Chinese fir plantations

Abstract Soil nutrient contents in ecosystems play a crucial role in forest growth and site productivity. We evaluated stand basal area growth in response to soil organic matters (OMs) and potassium in the plantations of Chinese fir [Cunninghamia lanceolata (Lamb.) Hook.]. Soil samples were collected from the top layer (0–20 cm) and dendrometric measurements were made in 150 subtropical Chinese fir plantations in Hunan Province. We carried out data analysis applying one-way variance analysis, dummy variables modeling, and K-means clustering methods. The results showed the following: (i) Available potassium (AK) was the most affecting factor, followed by OM. (ii) The predictive accuracy of the stand basal area growth model was substantially improved with coupled AK and OM used as dummy variables (AIC decreased from 820.6025 to 785.0356). (iii) In addition, the optimal nutrient type of group (SNECTG1, soil nutrient element combination type group) and soil factors affecting basal area growth and their range (OM 10 ~ 20 g kg−1 and AK 30 ~ 50 mg kg−1) were effectively explained. SNECTG1 had the greatest effect on stand basal area growth, resulting in the maximum potential productivity of 3.622 m2 ha−1 year−1 at a stand density index of 2826 and site index of 12 m. (iv) The stand basal area increments of middle-aged forest was the highest, followed by nearly mature forest, and mature forest. In summary, our stand basal area growth model can be used to predict stand increments under different site indices or soil nutrient scenarios, which can provide theoretical and practical guidance for the cultivation of large-diameter timber forests.