基于生长模型的速生阔叶树种林分碳汇潜力研究

Carbon storage potentials for fast growing broad-leaf species based on forest growth models

【目的】比较不同立地条件和林分密度下速生阔叶树种的碳汇潜力,旨在为碳汇造林提供决策依据。【方法】采用昆明市第4次森林资源二类调查数据,选取桤木(Alnus cremastogyne)、银荆(Acacia dealbata)、蓝桉(Eucalyptus globulus)和直干桉(Eucalyptus maideni)林分作为研究对象,运用优化算法求解林分断面积生长备选模型Richards、Schumacher和Korf参数;将林分断面积生长模型和单木生物量模型与各器官含碳率链接,模拟不同立地条件和林分密度下林分碳储量生长过程;采用林分碳储量平均增长量和连年增长量分析各速生阔叶树种林分碳汇潜力。【结果】桤木、银荆、蓝桉和直干桉林分断面积生长模型拟合决定系数分别为0.94,0.89,0.92和0.84;优化算法求解参数迭代次数从小到大顺序为麦夸特算法、粒子群算法、差分进化算法、模拟退火算法和遗传算法,Schumacher模型和Korf模型拟合参数欧式距离小于Richards模型,但Korf模型和Schumacher模型的拟合参数稳定性优于Richards模型。地位级为Ⅲ级、林分密度指数为800时15年生桤木、银荆、蓝桉和直干桉林分碳储量分别为13.81,20.33,38.89和45.27t/hm^2。单因素方差分析表明,林龄为1~5年时各树种林分间碳储量生长不存在差异(α=0.10),林龄为6~10和11~15年时各树种林分间碳储量生长存在极显著差异(α=0.01);桤木、银荆、蓝桉和直干桉林分碳储量平均增长量最大值分别为1.14,1.76,3.10和3.11t/(hm^2·年),连年增长量最大值分别为1.34,2.18,3.66和3.41t/(hm^2·年)。【结论】建立的速生阔叶树种生长模型可用于林分碳储量生长分析,优化算法能对生长模型更好地进行参数估计,各树种的固碳潜力顺序为直干桉>蓝桉>银荆>桤木。

[Objective] This study compared the carbon sequestration potentials of different fast grow- ing broad-leaf species with different sites and densities to provide basis for afforestation decision-making. [Method] Forest inventory data of ALnus cremastogyne ,Acacia deaLbata ,Eucalyptus globulus ,E. maideni from the 4th forest resources survey in Kunming were used for establishment of basal area growth models. Optimal algorithms were applied in solving parameters of alternative models including Richards,Schumach- er,and Korf. Tree-level biomass models were linked with growth models to simulate the growth of carbon storage with site situations and stand densities. Mean increment and annual increment of carbon storagewere used in comparison of carbon sequestration potentials of different species. [Result] Decision coeffi- cients of fitting were 0. 94,0. 89,0. 92 and 0.84 for Alnus cremastogyne ,Acacia dealbata ,Eucalyptus glob- ulus ,and Eucalyptus maiden. Iterations of different algorithms for parameter solving were in the order of levenberg-marquardt〈particle swarm optimization〈 differential evolution〈 simulated annealing〈genetic algorithm. Euclidean distances of solved parameters for Sehumacher and Korf were bigger than that of Richards. Carbon storages of Alnus cremastogyne, Acacia dealbata, Eucalyptus globulus, and Eucalyptus maiden were 13.81,20.33,38.89 and 45.27 t/hm2 in 15 years old stands（ SC. Ill ,SDI.8OO）,respectively. There was no difference in carbon storage between species in 1--5 years old stands （c~=0.10）. The differ- ence of carbon storage was extremely significant between species in 6-- 10 years and 11 -- 15 years old stands （a=0. 01）. The maximum mean increments of carbon storage were 1. 14,1. 76,3. 10,and 3. 11 t/（hm2 ~ a） and the maximum annual increments of carbon storage were 1. 34, 2. 18, 3. 66, and 3. 41 t/（hm2 ~ a） for Alnus cremastogyne ,Acacia dealbata, Eucalyptus globulus, and Eucalyptus maideni , re- spectively. [Conclusion] The established models could be applied in carbon storage analysis and the optimal algorithms were good at parameters solving. Carbon storage potentials for fast growing broad-leaf species were in the order of Eucalyptus maideni〉 Eucal yptus globulus〉 Acacia dealbata 〉 Alnus cremastog yne.