【目的】构建杉木(Cunninghamia lanceolata)人工林林分优势高预测模型,为其立地质量评价与生长预测提供支持。【方法】基于2021年与2022年两期全国林草生态综合监测中的杉木人工林样地数据,分别采用普通最小二乘法(OLS)、独立度量误差模型(IEIVM)和度量误差变量联立方程组模型(SEIVE)构建林分优势高估算模型,实现林分优势高、平均高与平均胸径间的相互预估,并利用决定系数(R²)、均方根误差与相对均方根误差评价模型性能。【结果】①IEIVM与SEIVE方法构建的模型预测精度显著优于OLS,R²为0.697~0.718,相对均方根误差为16.08%~16.59%;②SEIVE方法略优于IEIVM,在考虑变量间误差结构后,模型拟合效果更佳,且能实现3个变量间的相互精确预测。【结论】基于SEIVE方法并考虑误差结构所建立的杉木林分优势高估算模型具有较好的适用性与预测能力,可支持“平均胸径—平均高—优势高”的递推预测,为实现基于历史调查数据的杉木人工林立地质量精准评价提供基础模型。

【Objective】A stand dominant height estimation model for Cunninghamia lanceolata (Chinese fir) plantations was developed to provide a foundation for site quality assessment and growth prediction of Chinese fir plantations.【Method】Based on survey data from Chinese fir plantation plots from the two phases of the National Forest and Grassland Ecosystem Integrated Monitoring Program in 2021 and 2022, we constructed dominant height estimation models using the ordinary least squares regression (OLS), the independent measurement error-in-variable model (IEIVM), and the simultaneous measurement error-in-variable equations (SEIVE). These models were used to realize the conversion among stand dominant height, stand mean height, and stand mean diameter at breast height (DBH), and the models were evaluated using indicators such as the coefficient of determination (R²), root mean square error (RMSE), and relative root mean square error (RRMSE).【Result】(1) The stand dominant height estimation models established by the IEIVM and SEIVE methods performed significantly better than that by the OLS method, with R² ranging from 0.697 to 0.718 and RRMSE is between 16.08% and 16.59%. (2) After considering the error structure among variables, the SEIVE method was slightly superior to the IEIVM method, and the fitting effect of the model was better, and it could realize accurate mutual predictions in the three variables.【Conclusion】The stand dominant height estimation model of Chinese fir based on SEIVE and considering the error structures among variables has good applicability and predictive performance. It supports the sequential prediction of mean DBH to mean height and dominant height, which provides a fundamental model for the site quality evaluation of Chinese fir plantations.