关键词: 树高-胸径模型, 树高分级, 双重迭代算法, 参数灵敏度分析

Abstract:

Site plays an important role in fitting height-diameter model in such a large scale as nation, whereas it is not known in most cases. Using the classified height method with double iteration algorithm, height-diameter model of six major tree species in china, including Quercus, Cunninghamia lanceolata, Pinus massoniana, Populus, Larix gmelinii and Pinus tabulaeformis,were constructed. The data came from height-measurement file in 7^th national consecutive forest inventory,total observations were 118 441,of which 79 095 observations used for model development and 39 346 observations used for model validation. Contrasting to ordinary method, the coefficients of determination reached 0.942 8-0.966 5 from 0.520 3-0.753 2. Parameter sensitivity analysis and model validation showed that models were robust in general and fitted parameters were statistically reliable. The classified height method is useful to construct height-diameter model for all tree species in china. Considering stand factors,the method is helpful to evaluate national forest site.

Key words: height-diameter model, classified height, double iteration algorithm, parameter sensitivity analysis