Regionally Compatible Individual Tree Growth Model under the Combined Influence of Environment and Competition

文献类型： 外文期刊

第一作者： Zhang, Wenjie

作者： Zhang, Wenjie;Wu, Baoguo;Zhang, Wenjie;Yang, Guijun;Zhang, Wenjie;Yang, Guijun;Wu, Baoguo;Ren, Yi

作者机构：

关键词： regionally compatible; competition; environmental factors; individual tree growth

期刊名称：PLANTS-BASEL （ 影响因子：4.5； 五年影响因子：4.8 ）

ISSN：

年卷期： 2023 年 12 卷 14 期

页码：

收录情况： SCI

摘要： To explore the effects of competition, site, and climate on the growth of Chinese fir individual tree diameter at breast height (DBH) and tree height (H), a regionally compatible individual tree growth model under the combined influence of environment and competition was constructed. Using continuous forest inventory (CFI) sample plot data from Fujian Province between 1993 and 2018, we constructed an individual tree DBH model and an H model based on re-parameterization (RP), BP neural network (BP), and random forest (RF), which compared the accuracy of the different modeling methods. The results showed that the inclusion of competition and environmental factors could improve the prediction accuracy of the model. Among the site factors, slope position (PW) had the most significant effect, followed by elevation (HB) and slope aspect (PX). Among the climate factors, the highest contribution was made by degree-days above 18 & DEG;C (DD18), followed by mean annual precipitation (MAP) and Hargreaves reference evaporation (Eref). The comparison results of the three modeling methods show that the RF model has the best fitting effect. The R-2 of the individual DBH model based on RF is 0.849, RMSE is 1.691 cm, and MAE is 1.267 cm. The R-2 of the individual H model based on RF is 0.845, RMSE is 1.267 m, and MAE is 1.153 m. The model constructed in this study has the advantages of environmental sensitivity, statistical reliability, and prediction efficiency. The results can provide theoretical support for management decision-making and harvest prediction of mixed uneven-aged forest.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序