论文部分内容阅读
本文利用闽北四县市收集的杉木９９４株、马尾松１０８７株、阔叶树１０３２株样木资料，按树种随机抽取１００组胸径与去皮地径的成对值，用５种胸径与去皮地径的回归方程对各树种进行拟合与检验．以剩余平方和最小者为优，比较得出：杉木、阔叶树以一元线性而马尾松以幂函数方程为佳，其相关系数分别为０．９９８５、０．９９５２、０．９８６５。用５０株模外样木资料对估测模型进行Ｆ检验和Ｕ检验，估测模型的系统误差和约方误均小于５％，表明模型预测精度较高，可供林业生产经营管理使用．
In this paper, 994 Chinese fir, 1,087 Masson massoniana and 1032 broadleaved trees collected from four counties of northern Fujian were used to select 100 paired DBH and DBH paired trees according to tree species. Five DBH and peeled The path regression equation fitted and tested the species. The best of the least square residuals was obtained. The results showed that the fir and broadleaf trees were linear and the mass function of the masson pine was better than the power function equation. The correlation coefficients were 0.9985, 0.9952 and 0.9865 respectively. F-test and U-test were used to evaluate the estimation model with 50 out-of-wood samples. The systematic errors and approximate square errors of the estimated models were both less than 5%, indicating that the prediction accuracy of the model was high and could be used by forestry production and management.