基于哑变量和分位数回归的兴安落叶松更新幼树的树高-胸径模型

doi:10.13287/j.1001-9332.202309.005

摘要/Abstract

摘要： 本研究以大兴安岭地区翠岗林场2018—2019年55块固定样地2054株兴安落叶松幼树为对象,采用四分位数法将林分密度指数(SDI)划分为4个等级,即等级Ⅰ(SDI₁<1863株·hm^-2)、等级Ⅱ(1863≤SDI₂<2155株·hm^-2)、等级Ⅲ(2155≤SDI₃<2459株·hm^-2)和等级Ⅳ(SDI₄≥2459株·hm^-2),并采用哑变量方法引入SDI构建兴安落叶松幼树树高-胸径的哑变量模型和分位数回归模型。结果表明: 选取的5个代表性非线性树高曲线模型中,Richards模型的拟合效果最好,其R_a²、RMSE、MAE分别为0.7637、0.8250 m、0.5696 m;基于Richards模型构建的包含SDI的哑变量模型,其R_a²较基础模型提高了1.3%,而RMSE、MAE、AIC分别降低了2.1%、1.5%和11.2%;当分位点τ=0.5时,分位数回归模型的R_a²最大,RMSE、MAE、AIC最小,分别为0.7612、0.8294 m、0.5657 m、-767.19。相较于SDI₁,SDI₂～SDI₄林分内幼树的树高分别增加5.6%、5.6%和11.3%。因此,合理调控兴安落叶松林的林分密度有利于增加更新幼树的高生长。

关键词: 哑变量, 分位数回归, 树高, 胸径, 林分密度指数

Abstract: Based on data collected from 2054 saplings of Larix gmelinii forest in 55 fixed plots in 2018-2019 in Cuigang Forestry Station, Daxing’anling area, we classified the stand density index (SDI) into four classes, i.e., Class Ⅰ (SDI₁<1863 plants·hm^-2), Class Ⅱ (1863 plants·hm^-2≤SDI₂<2155 plants·hm^-2), Class Ⅲ (2155 plants·hm^-2≤SDI₃<2459 plants·hm^-2) and Class Ⅳ (SDI₄≥2459 plants·hm^-2) by using the quartile method. We constructed a dummy variable model and quantile regression model for the height-breast diameter of saplings of L. gmelinii with dummy variable method introduced SDI. The results showed that among the five selected representative non-linear tree height curve models, the Richards model fitted the best, with R_a², RMSE and MAE of 0.7637, 0.8250 m and 0.5696 m. The dummy variable model including the SDI constructed based on the Richards model showed a 1.3% increase in R_a² compared with the base model, while RMSE, MAE, and AIC decreased by 2.1%, 1.5%, and 11.2%, respectively. When the quantile τ was 0.5, R_a² of quantile regression model was the maximum, and RMSE, MAE, AIC was the minimum, being 0.7612, 0.8294 m, 0.5657 m, and -767.19, respectively. Compared with SDI₁, sapling height in SDI₂-SDI₄ was increased by 5.6%, 5.6%, and 11.3%, suggesting reasonable that regulation of stand density was conducive to increase the height growth of saplings in regeneration.

Key words: dummy variable, quantile regression, tree height, diameter at breast height, stand density index

吕乐乐, 王文彬, 董灵波. 基于哑变量和分位数回归的兴安落叶松更新幼树的树高-胸径模型[J]. 应用生态学报, 2023, 34(9): 2355-2362.

LYU Lele, WANG Wenbin, DONG Lingbo. Height-diameter models of regenerated saplings of Larix gmelinii based on dummy variable and quantile regression[J]. Chinese Journal of Applied Ecology, 2023, 34(9): 2355-2362.

参考文献

[1] 贾炜玮, 解希涛, 姜生伟, 等. 大兴安岭新林林业局3种林分类型天然更新幼苗幼树的空间分布格局. 应用生态学报, 2017, 28(9): 2813-2822
[2] EerikäInen K, Valkonen S, Saksa T. Ingrowth, survival and height growth of small trees in uneven-aged Picea abies stands in southern Finland. Forest Ecosystems, 2014, 1: 1-10
[3] Ducey MJ, Knapp RA. A stand density index for complex mixed species forests in the northeastern United States. Forest Ecology and Management, 2010, 260: 1613-1622
[4] 周晏平, 雷泽勇, 赵国军, 等. 沙地樟子松不同树高-胸径模型比较分析. 华南农业大学学报, 2019, 40(3): 75-81
[5] 樊伟, 许崇华, 崔珺, 等. 基于混合效应的大别山地区杉木树高-胸径模型比较. 应用生态学报, 2017, 28(9): 2831-2839
[6] 徐庆华, 杨进良, 黄练忠, 等. 次生常绿阔叶林群落林冠结构对林下植被的影响. 浙江农林大学学报, 2019, 36(6): 1151-1157
[7] Pya N, Schmidt M. Incorporating shape constraints in generalized additive modelling of the height-diameter relationship for Norway spruce. Forest Ecosystems, 2016, 3: 112-125
[8] 陈甲瑞, 王小兰. 藏东南尼洋河流域高山松树高曲线的研究. 中南林业科技大学学报, 2019, 39(8): 16-20
[9] 刘济明, 程晶, 李丽霞, 等. 米槁种群树高-胸径关系研究. 甘肃农业大学学报, 2019, 54(5): 151-162
[10] 冯国红, 朱玉杰, 王景峰. 小兴安岭天然林不同树种的树高曲线模型. 东北林业大学学报, 2016, 44(12): 6-9
[11] 李春明, 李利学. 基于非线性混合模型的栓皮栎树高与胸径关系研究. 北京林业大学学报, 2009, 31(4): 7-12
[12] 李婉婷, 姜立春, 万道印. 基于混合效应的兴安落叶松树高与胸径关系模拟. 植物研究, 2014, 34(3): 343-348
[13] 王小明, 李凤日, 贾炜玮, 等. 帽儿山林场天然次生林阔叶树种树高-胸径模型. 东北林业大学学报, 2013, 41(12): 116-120
[14] Wang J. Bayesian quantile regression for parametric nonlinear mixed effects models. Statistical Methods & Applications, 2012, 21: 279-295
[15] 卢军, 张会儒, 雷相东, 等. 长白山云冷杉针阔混交林幼树树高-胸径模型. 北京林业大学学报, 2015, 37(11): 10-25
[16] 姚丹丹, 徐奇刚, 闫晓旺, 等. 基于贝叶斯方法的蒙古栎林单木树高-胸径模型. 南京林业大学学报: 自然科学版, 2020, 44(1): 131-137
[17] 王冬至, 张冬燕, 李永宁, 等. 基于贝叶斯法的针阔混交林树高与胸径混合效应模型. 林业科学, 2019, 55(11): 85-94
[18] Koenker R, Bassett G. Regression quantiles. Econome-trica, 1978, 46: 33-50
[19] 何晓申. 分位回归理论及应用研究. 经济研究导刊, 2016(16): 6-7
[20] Bohora SB, Cao QV. Prediction of tree diameter growth using quantile regression and mixed-effects models. Forest Ecology and Management, 2014, 319: 62-66
[21] Ducey MJ. The ratio of additive and traditional stand density indices. Western Journal of Applied Forestry, 2009, 24: 5-10
[22] Zang H, Lei X, Zeng W. Height-diameter equations for larch plantation in northern and northeastern China: A comparison of the mixed-effects, quantile regression and generalized additive models. Forestry, 2016, 89: 434-445
[23] Vans AM, Gregoire TG. A geographically variable model of hemlock woolly adelgid spread. Biological Invasions, 2007, 9: 369-382
[24] Mehtatalo L, Gregoire TG, Burkhart HE. Comparing strategies for modeling tree diameter percentiles from remeasured plots. Environmetrics, 2010, 19: 529-548
[25] 高慧淋, 董利虎, 李凤日. 基于分位数回归的长白落叶松人工林最大密度线. 应用生态学报, 2016, 27(11): 3420-3426
[26] 马岩岩, 姜立春. 基于非线性分位数回归的落叶松树干削度方程. 林业科学, 2019, 55(10): 68-75
[27] 辛士冬, 姜立春. 利用分位数回归模拟人工樟子松树干干形. 北京林业大学学报, 2020, 42(2): 1-8
[28] Özçelik R, Cao QV, Trincado G, et al. Predicting tree height from tree diameter and dominant height using mixed-effects and quantile regression models for two species in Turkey. Forest Ecology and Management, 2018, 419-420: 240-248
[29] Crecente-Campo F, Marshall P, LeMay V, et al. A crown profile model for Pinus radiata D. Don in northwestern Spain. Forest Ecology and Management, 2009, 257: 2370-2379
[30] 王君杰, 夏宛琦, 姜立春. 基于分位数回归和哑变量模型的大兴安岭兴安落叶松树高-胸径模型. 中南林业科技大学学报, 2020, 40(9): 24-32
[31] 高东启, 邓华锋, 王海宾, 等. 基于哑变量的蒙古栎林分生长模型. 东北林业大学学报, 2014, 21(1): 61-64
[32] 杨英, 冉啟香, 陈新云, 等. 哑变量在云杉地上生物量模型中的应用研究. 林业资源管理, 2015, 12(6): 71-76
[33] 郑冬梅, 曾伟生. 用哑变量方法构建东北落叶松和栎类分段地上生物量模型. 北京林业大学学报, 2013, 35(6): 32-35
[34] Chrimes D, Nilson K. Overstorey density influence on the height of Picea abies regeneration in northern Sweden. Forestry, 2005, 78: 433-442
[35] Saksa T, Valkonen S. Dynamics of seedling establishment and survival in uneven-aged boreal forests. Forest Ecology and Management, 2011, 261: 1409-1414
[36] Huang S, Price D, Titus SJ. Development of ecoregion-based height-diameter models for white spruce in boreal forests. Forest Ecology and Management, 2000, 129: 139-141
[37] Peng C, Zang L, Liu J. Developing and validating nonlinear height-diameter models for major tree species of Ontario’s boreal forests. Northern Journal of Applied Forestry, 2001, 18: 87-94
[38] 臧颢, 雷相东, 张会儒. 红松树高-胸径的非线性混合效应模型研究. 北京林业大学学报, 2016, 38(6): 8-16
[39] Reineke LH. Perfecting a stand-density index for even-aged forests. Journal of Agricultural Research, 1993, 46: 627-638
[40] 张会儒, 唐守正, 胥辉. 关于生物量模型中的异方差问题. 林业资源管理, 1999(1): 46-49
[41] Arias-Rodil M, Diéguez-Aranda U, Rodríguez Puerta F, et al. Modelling and localizing a stem taper function for Pinus radiata in Spain. Canadian Journal of Forest Research, 2015, 45: 647-658
[42] Sharma RP, Vacek Z, Vacek S, et al. Modelling individual tree height-diameter relationships for multi-layered and multi-species forests in central Europe. Trees, 2019, 33: 103-119
[43] 李海奎, 法蕾. 基于分级的全国主要树种树高-胸径曲线模型. 林业科学, 2011, 47(10): 83-90
[44] 张鹏, 何怀江, 范春雨, 等. 吉林蛟河针阔混交林主要树种树高-胸径模型. 林业科学研究, 2018, 31(2): 11-18
[45] 田瑞松. 立地条件对华北落叶松生长的影响. 现代园艺, 2018(19): 76-77
[46] Canham CD, Papaik MJ, Uriarte M, et al. Neighborhood analyses of canopy tree competition along environmental gradients in New England forests. Ecological Applications, 2006, 16: 540-554
[47] 历胤男, 刘盛, 李国伟, 等. 落叶松人工林冠下光环境对更新幼树生长的影响. 东北林业大学学报, 2016, 44(10): 8-11
[48] 刘春延. 河北赛罕坝落叶松人工林生长及生境因子关系研究. 博士论文. 北京: 北京林业大学, 2010
[49] 赵梦草, 顾海燕. 用分位数回归建立的混交林树高生长模型. 东北林业大学学报, 2020, 48(2): 24-28
[50] 田德超, 李凤日, 董利虎. 依据分位数回归建立德长白落叶松潜在最大冠幅预测模型. 东北林业大学学报, 2019, 47(8): 41-46