基于VAR模型的森林植被碳储量影响因素分析——以陕西省为例

森林作为陆地生态系统最大的碳库,对现在及未来的气候变化、碳平衡都具有重要影响。而对影响森林植被碳库的自然和非自然因素进行研究更是对增强森林的碳汇作用,继而改善生态环境状况意义重大。现有的森林动态模型虽然可以很好的模拟碳储量各影响因子之间的联系,但研究往往集中于小尺度从单一影响因素着手,且由于确定模型输入变量和参数的复杂性,使得这些模型在区域甚至更大尺度上的应用存在着一些困难。因此,运用VAR模型,以陕西省为例,构建森林植被碳储量与病虫害发生面积、木材产量、森林火灾面积、森林抚育面积、人工更新造林面积、降水和温度之间的动态关系,来验证该模型在省级尺度条件下的区域森林植被碳储量影响因素分析中的可行性。结果表明:各变量在5%的显著性水平下呈一阶单整序列并具有长期稳定的均衡关系,VAR模型也通过了平稳性检验满足运行的前提条件。通过脉冲响应和方差分解分析可知,森林病虫害、木材产量对陕西省森林植被碳储量呈现出很明显的负作用,并且贡献度很高,分别为5.61%和4.52%;森林抚育、人工更新造林对碳储量的影响存在一定的滞后期;火灾、温度和降水的冲击给碳储量带来的影响均不明显。模型较好的模拟了各影响因素对陕西省碳储量的影响,且具有一定的现实意义,因此,该模型可应用于省级尺度条件下的区域森林植被碳储量影响因素分析。

Analysis of factors that influence forest vegetation carbon storage by using the VAR model: a case study in Shanxi Province

Forest vegetation carbon dynamics are widely considered as basic indicators of the capacity of a forest ecosystem for carbon sequestration and carbon exchange with the atmosphere and the ability of the forest ecosystem to function as a carbon sink or source. Forest management practices such as harvesting, afforestation, and reforestation are the most important factors that influence forest carbon dynamics. Therefore, elucidation of the effects of human silvicultural activities on forest ecosystem dynamics will provide a valuable insight into ways of expanding the size of the carbon pool and improving the ecological environment of China. Previous studies have attempted to identify the factors that affect forest carbon sequestration; however, most of these investigations focused on a single factor and did not consider the interactions occurring between multiple factors. In addition, most previous studies were conducted on a small scale, and no investigations conducted at a regional or larger scale examined the influence of interactions occurring between multiple factors on forest carbon sequestration. Forest dynamic models represent a valuable tool for resolving this problem; however, existing models do not adequately reflect the interactive effects of multiple factors such as forest management practices, climate change, and other human-induced disturbances on forest carbon storage. In the present study, we used a vector autoregression (VAR) model that included seven parameters (factors) (timber output, forest-tending area, area affected by forest diseases and/or pests, fire-affected area, reforestation area, annual total precipitation, and annual average temperature) to simulate forest carbon storage dynamics in Shanxi Province. The results of variable stationary and cointegration analyses revealed that the investigated variables exhibited a single whole-sequence order at the 5% significance level; in addition, they showed long-term stable equilibrium. Hence, we applied each variable to the VAR model and used the model stationary test to show that all the characteristic roots of the VAR model were <1, i.e., these roots fell within the unit circle. Our results indicate that the VAR model system is stable and meets the prerequisites for identifying the effects of multiple factors on forest vegetation carbon storage dynamics. Finally, we used impulse response and variance decomposition analyses to show that forest pests and timber output had significant negative impacts on carbon storage and contributed to a high degree of 5.61% and 4.52%, respectively. Forest tending and artificial reforestation had weakly negative effects on carbon storage, whereas forest fires, precipitation, and temperature had no significant effects on carbon storage. Taken together, our findings verify the suitability of the VAR model for identifying the effects of multiple factors on forest ecosystem carbon storage dynamics. Therefore, in future studies, this model can be used to analyze factors that influence forest vegetation carbon storage at a provincial scale.