College of Forestry,Northeast Forestry University,Harbin 150040,China
Show less
文章历史+
Received
Published
2024-11-29
2025-09-15
Issue Date
2025-10-30
PDF (2525K)
摘要
林分进界是林分生长动态变化过程中的一个重要环节,对于维持森林资源中的生物多样性和群落结构稳定至关重要。为此,在帽儿山实验林场设立的61块样地数据,从林分因子及生物多样性因子等方面进行考虑,通过肯德尔tau-b(Kendall-Tau-b)相关系数分析及考虑变量多重共线性选择最适变量,采用泊松(Poisson)模型、负二项模型(negative binomial model,NB)、零膨胀模型及障碍(Hurdle)模型构建进界模型。采用层次分割法分析变量贡献率,以找出影响进界模型的关键因素。研究结果表明,林分拥挤度(K)、林分算数平均胸径(d)、辛普森(Simpson)指数和林分平均高(mean height,MH)是影响每公顷进界株数(number of advance regeneration per hectare,Nn)的重要因子。通过赤池信息准则(akaike information criterion,AIC)、贝叶斯信息准则(bayesian information criterion,BIC)及对数似然值(logarithm likelihood,Loglike)指标对比,发现零膨胀负二项式(zero-inflated negative binomial,ZINB)模型与障碍负二项(hurdle negative binomial,HNB)模型显著优于其他模型。通过沃恩(Vuong)检验,发现负二项模型及其复合模型(ZINB、HNB)在拟合帽儿山天然林进界数量方面优于Poisson模型及其复合模型(零膨胀泊松模型(zero-inflated poisson model,ZIP)、障碍泊松模型(hurdle poisson model,HP)),且ZINB模型略优于HNB模型,故ZINB模型为拟合帽儿山天然林林分进界数量的最优模型,十折交叉检验也验证了这一结论。同时,通过层次分割法分析发现Simpson指数和林分平均高(MH)分别对最优进界模型(ZINB)的计数部分和零部分贡献度最高。所构建的帽儿山天然林进界模型具有一定的统计可靠性,可用于该地区进界生长预测,为当地天然林更新管理提供科学依据。
Abstract
Forest stand ingrowth is a critical component of the dynamic growth process of forest stands, essential for maintaining biodiversity and community structure stability in forest resources. Based on data from 61 plots established at the Maoer Mountain Experimental Forest Farm, this study considered factors such as stand characteristics and biodiversity. Through Kendall-Tau-b correlation coefficient analysis and selection of the most suitable variables considering multicollinearity among variables, models for ingrowth were constructed using Poisson, negative binomial (NB), zero-inflated, and Hurdle models. The contribution rate of variables was analyzed using hierarchical partitioning to identify key factors influencing the ingrowth model. The results showed that stand density (K), arithmetic mean diameter at breast height (d), Simpson's index, and mean stand height (MH) were significant factors affecting the number of ingrowth trees per hectare (Nn). Comparing models using AIC, BIC, and Loglike criteria, it was found that ZINB and HNB significantly outperformed other models. The Vuong test further revealed that the negative binomial models and their composite models (ZINB, HNB) performed better than Poisson models and their composites (ZIP, HP) in fitting the ingrowth quantity of natural forests in Maoer Mountain, with the ZINB model slightly outperforming the HNB model. Therefore, the ZINB model was the optimal model for fitting the ingrowth quantity of natural forest stands in Maoer Mountain, a conclusion also supported by ten-fold cross-validation. Additionally, hierarchical partitioning analysis indicated that the Simpson's index and mean stand height (MH) contributed most to the count and zero parts, respectively, of the optimal ingrowth model (ZINB). The natural forest ingrowth model constructed by this research has a certain statistical reliability and can be used for ingrowth prediction in the Maoer Mountain area, providing a scientific basis for local natural forest regeneration management.
本研究数据来源于帽儿山61块天然林复测数据,共有25种树种、22 848株样木,包括进界木495株。其中28块样地设立于2007年,之后每隔5 a进行复测,直至2022年。另外33块样地设立于2017年,于2022年进行复测。本研究所用样地面积为0.06~0.210 hm2,设立好样地后对样地内的每个样木进行每木检尺,记录胸径、树高、冠幅和枝下高等。然后对林分起源、林分年龄、郁闭度、海拔因子进行调查,并在5 a后对样地进行复测,记录样地的进界情况。整个调查区间内的每公顷进界株数(number of advance regeneration per hectare,Nn)的频数分布直方图如图1所示。
对于Poisson模型、NB模型、ZINB模型、ZIP模型、HNB模型和HP模型的拟合情况采用赤池信息准则(akaike information criterion,AIC)、贝叶斯信息准则(bayesian information criterion,BIC)以及对数似然值(logarithm likelihood,Loglike)这3个指标来进行比较。AIC、BIC值越小,Loglike值越大代表拟合效果越好。沃恩(Vuong)检验在比较一般计数模型与复合模型的分析中表现出较高的效能[28]。当AIC值相近时,采用Vuong检验来比较拟合模型。模型采用十字交叉检验来评估各模型的预测精度,使用均方根误差(RMSE)和平均绝对误差(MAE)2个指标对模型进行评价,RMSE和MAE越小,模型预测精度越高。其中,RMSE(式中记为RMSE)和MAE(式中记为MAE)的公式为
GRISCOMB W, BUSCHJ, COOK-PATTONS C,et al.National mitigation potential from natural climate solutions in the tropics[J].Philosophical Transactions of the Royal Society B,2020,375(1794):20190126.
[2]
李武赫.大兴安岭东部落叶松-白桦混交林进界生长模型[D].哈尔滨:东北林业大学,2021.
[3]
LIW H.Ingrowth model of Larix gmelinii-Betula platyphylla mixed forest in eastern Daxing'an Mountains[D].Harbin:Northeast Forestry University,2021.
[4]
韩有志,王政权.森林更新与空间异质性[J].应用生态学报,2002(5):615-619.
[5]
HANY Z, WANGZ Q.Spatial heterogeneity and forest regeneration[J].Chinese Journal of Applied Ecology,2002(5):615-619.
YANX Y, ZHANGG Q, ZANGL P,et al.Based on the 2012—2021 English literature metrology analysis of the forest seedling research progress update[J].Journal of Mountain Agriculture and Biology,2024,43(4):23-30.
LEI Y C,ZHANG X Q,Tree recruitment model of Larix olgensis [J].Forest Research,2013,26(5):554-561.
[10]
ANDREASSENK.Development and yield in selection forest[J].Meddelelser fra Skogforsk,1994,47(5):1-37.
[11]
BUONGIORNOJ, MICHIEB R.A matrix model of uneven-aged forest management[J].Forest Science,1980,26(4):609-625.
[12]
GRIMESR F, PEGGR E.Growth data for a spotted gum-ironbark forest in south-east Queensland[J].Queensland,Department of Foestry,Technical Paper,1979.
[13]
FORTINM, DEBLOISJ.Modeling tree recruitment with zero-inflated models:The example of hardwood stands in southern Quebec,Canada[J].Forest Science,2007,53(4):529-539.
ZHENGR, XIAOS P, YANGX C,et al.Estimate model of ingrowth diameter for the middle and young artifical forest of Pinus elliottii in Xiajiang County[J].Jiangxi Forestry Science and Technology,2004(2):15-16.
CHENK Q, ZHANGH R, ZHANGB,et al.Spatial distribution simulation of recruitment trees of natural secondary forest based on geographically weighted regression[J].Journal of Beijing Forestry University,2021,43(2):1-9.
[23]
RUSSELLM B, WESTFALLJ A, WOODALLC W.Modeling browse impacts on sapling and tree recruitment across forests in the northern United States[J].Canadian Journal of Forest Research,2017,47(11):1474-1481.
[24]
SHIH, ZHOUQ, HER,et al.Temperature sensitivity of tree recruitment at alpine treelines increases along latitudinal gradients[J].Ecography,2023,2023(9):e06700.
[25]
KÄBERY, MEYERP, STILLHARDJ,et al.Tree recruitment is determined by stand structure and shade tolerance with uncertain role of climate and water relations[J].Ecology and Evolution,2021,11(17):12182-12203.
LIC M, LIL X.Simulating study on tree recruitment of Quercus mongolica based on zero-inflated model and mixed effect model methods[J].Journal of Beijing Forestry University,2020,42(6):59-67.
[28]
GUOY, LIZ, LIUP,et al.Modeling correlation and heterogeneity in crash rates by collision types using full Bayesian random parameters multivariate Tobit model[J].Accident Analysis and Prevention,2019,128:164-174.
[29]
QUEROM G, GUICCIARDIS, PENNAP,et al.Increasing trends in faecal pollution revealed over a decade in the central Adriatic Sea (Italy)[J].Water Research,2024,262:122083.
HUIG Y, ZHANGL J, HUY B,et al.Stand crowding degree and its application[J].Journal of Beijing Forestry University,2016,38(10):1-6.
[32]
KENDALLM G.A new measure of rank correlation[J].Biometrika,1938,30(1/2):81-93.
[33]
PERDECKA C.Poisson regression as a flexible alternative in the analysis of ring-recovery data[J].European Union for Bird Ringing Newsletter,1998,2:37-39.
LAMBERTD.Zero-Inflated poisson regression,with an application to defects in manufacturing[J].Technometrics,2012,34(1):1-14.
[36]
ESKELSONB N I, TEMESGENH, BARRETTT M.Estimating cavity tree and snag abundance using negative binomial regression models and nearest neighbor imputation methods[J].Canadian Journal of Forest Research,2009,39(9):1749-1765.
[37]
MULLAHYJ.Specification and testing of some modified count data models[J].Journal of Econometrics,1986,33(3):341-365.
GUOF T, HUH Q, JINS,et al.Relationship between forest lighting fire occurrence and weather factors in Daxing'an Mountains based on negative binomial model and zero-inflated negative binomial models[J].Chinese Journal of Plant Ecology,2010,34(5):571-577.
[40]
LAIJ, ZOUY, ZHANGS,et al.glmm.hp:an R package for computing individual effect of predictors in generalized linear mixed models[J].Journal of Plant Ecology,2022,15(6):1302-1307.
[41]
LAIJ, ZHUW, CUID,et al.Extension of the glmm.hp package to zero-inflated generalized linear mixed models and multiple regression[J].Journal of Plant Ecology,2023,16(6):rtad038.
[42]
BARRYC S, WELSHA.Generalized additive modelling and zero inflated count data[J].Ecological Modelling,2002,157(2/3):179-188.
[43]
YAACOBW, FAIROSW Y W, LAZIMM A,et al.A practical approach in modelling count data[C].Proceedings of the Regional Conference on Statistical Sciences.Malaysia,2010:176-183.
[44]
CUNNINGHAMB R, LINDENMAYERB D.Modeling count data of rare species:Some statistical issues[J].Ecology,2005,86(5):1135-1142.
[45]
FENGC X.A comparison of zero-inflated and hurdle models for modeling zero-inflated count data[J].Journal of Statistical Distributions and Applications,2021, 8(1):8.
XIAOC, TIAND Y, MAR,et al.Compatibility predictive model for regeneration quantities of Larix gmelinii natural forest in Daxing’anling Mountains,China[J].Chinese Journal of Applied Ecology,2023,34(9):2345-2354.
[48]
田栋元.大兴安岭地区主要林分类型更新空间格局及数量模型[D].哈尔滨:东北林业大学,2022.
[49]
TIAND Y.Spatial pattern and quantitative model for the regeneration of major forest types in the Daxing'anling region[D].Harbin:Northeast Forestry University,2022.