关键词: 长白落叶松, 分位数回归, 冠幅模型, 林分经营密度数表

Abstract: Based on the survey data of 55 permanent plots of Larix olgensis plantation in Maoershan Experimental Forest Farm of Northeast Forestry University, we divided the stand density index (SDI) of the plots into three grades by mean±standard deviation method, namely SDI Ⅰ∈(0, 695], SDI Ⅱ∈(695, 1027], and SDI Ⅲ∈(1027, ∞] trees·hm^-2. Based on the Logistic equation, we constructed the maximum crown width prediction model of individual tree at three quantiles (0.90, 0.95, and 0.99) under different SDI grades by coupling dummy variables and quantile regression. We further used the crown projection area method to compile the stand management density number table under different SDI grades and quantified the influence of stand density on stand volume and carbon storage. The results showed that the maximum crown width models of individual tree with different SDI had significant differences, and that the maximum crown width of individual tree could be better simulated at 0.90 quantile. The independent sample test showed that compared with the Logistic basic model, the R_adj² of the single-tree dummy variable quantile maximum crown model could be significantly increased by 0.20, while the root mean square error was significantly reduced by 0.15 m. Based on the established single tree dummy variable quantile maximum crown width model and crown projection area method, we compiled the stand management density number table under different SDI grades. When the target diameter of cultivation was 30 cm, compared with SDI Ⅰ, stand volume and carbon storage were overestimated by 26.16 m³·hm^-2 and 10.10 t C·hm^-2, respectively, when SDI grades were not distinguished. Similarly, compared with SDIⅡ, these values were overestimated by 15.99 m³·hm^-2 and 6.12 t C·hm^-2. However, compared with SDIⅢ, they were significantly underestimated by 85.13 m³·hm^-2 and 33.04 t C·hm^-2. Our results indicated that ignoring the differences of SDI grades could overestimate the carbon sequestration capacity of low-density stands but underestimate the actual contribution of high-density stands. Therefore, implementing stand density regulation based on SDI grades is conducive to achieving precise quality improvement of L. olgensis plantations.

Key words: Larix olgensis, quantile regression, crown width model, stand management density number table