基于多元统计分析的马尾松人工林健康评价研究——以广西热带林业实验中心为例
|
摘要
为构建一套科学实用的马尾松人工林健康评价体系,为马尾松健康经营提供理论依据。以广西热带林业实验中心的84块马尾松人工林样地为例,选取13个评价指标进行因子分析,计算森林健康指数(FHI)并对FHI值进行Ward聚类,以划分健康等级。此外,研究分析评价了马尾松样地的健康状况,并用判别分析建立Fisher判别函数对健康评价结果进行检验。结果表明:在调查的马尾松人工林中,健康林占32.1%,亚健康林占35.8%,不健康林占32.1%,且不同龄组马尾松人工林健康水平排序为:近熟林>中龄林>成熟林>幼龄林。判别分析与聚类结果具有较好的一致性,其中自身验证法正判率为97.8%,交互验证法正判率为94.3%。因此,构建的指标体系能够科学客观地反映广西热带林业实验中心马尾松人工林的健康状况。
In order to establish a scientific and practical health evaluation system for Pinus massoniana plantation and provide a theoretical basis for healthy management of Pinus massoniana, This study took 84 plots of Pinus massoniana plantations in Guangxi tropical forestry experimental center for example, then selected 13 indicators among these plots to calculate FHI index using the factor analysis method, then clustered FHI values to rank health level using the Ward cluster method. Besides, this study analyzed and assessed health state of these cluster groups, then established Fisher discriminant functional relations by discriminant analysis, finally, tested the health assessment results.The results showed that in these Pinus massoniana plantations surveyed, healthy forests accounted for 32.1%, sub-healthy forests accounted for 35.8% and unhealthy forests accounted for 32.1%, in addition, the health levels of Pinus massoniana plantations in different age groups were near-mature forest > middle age forest > mature forest > young forest.The result of discriminant analysis well agreed with clustering results, the positive verification rate of self-verification method and cross validation method were 97.8%, 94.3%.Therefore, the built indicator system can scientific and objective reflect the health status of the Pinus massoniana plantations in the Guangxi tropical forestry experimental center.
In order to establish a scientific and practical health evaluation system for Pinus massoniana plantation and provide a theoretical basis for healthy management of Pinus massoniana, This study took 84 plots of Pinus massoniana plantations in Guangxi tropical forestry experimental center for example, then selected 13 indicators among these plots to calculate FHI index using the factor analysis method, then clustered FHI values to rank health level using the Ward cluster method. Besides, this study analyzed and assessed health state of these cluster groups, then established Fisher discriminant functional relations by discriminant analysis, finally, tested the health assessment results.The results showed that in these Pinus massoniana plantations surveyed, healthy forests accounted for 32.1%, sub-healthy forests accounted for 35.8% and unhealthy forests accounted for 32.1%, in addition, the health levels of Pinus massoniana plantations in different age groups were near-mature forest > middle age forest > mature forest > young forest.The result of discriminant analysis well agreed with clustering results, the positive verification rate of self-verification method and cross validation method were 97.8%, 94.3%.Therefore, the built indicator system can scientific and objective reflect the health status of the Pinus massoniana plantations in the Guangxi tropical forestry experimental center.
引文
[1]CIESLA W M,DONAUBAUER E.Decline and dieback of trees and forests:a global overview[R].Fao Forestry Paper,1994.
[2]BURGER J A,ADAMS M B,CHANG S.Management effects on growth,production and sustainability of managed forest ecosystems:past trends and future directions[J].Forest Ecology&Management,2009,258(10):2335-46.
[3]甘敬,张振明,余新晓,等.森林健康监测与评价研究[J].水土保持研究,2006,13(3):177-80.
[4]MISTRETTA P A.Managing for forest health[J].Journal of Forestry Washington,2002,100(7):24-7.
[5]MENG J,LI S,WANG W,et al.Mapping Forest Health Using Spectral and Textural Information Extracted from SPOT-5Satellite Images[J].Remote Sensing,2016,8(9):719.
[6]COSTANZA R,NORTON B G,HASKELL B D.Ecosystem health:new goals for environmental management[J].Ecosystem Health New Goals for Environmental Management,1992,28:33-34.
[7]LUCIER A A.Criteria for success in managing forested landscapes[J].Journal of Forestry,1994,92(7):20-24.
[8]肖风劲,欧阳华,孙江华,等.森林生态系统健康评价指标与方法[J].林业资源管理,2004(1):27-30.
[9]甘敬,朱建刚,张国祯,等.基于BP神经网络确立森林健康快速评价指标[J].林业科学,2007,43(12):1-7.
[10]王秋燕,陈鹏飞,李学东,等.森林健康评价方法综述[J].南京林业大学学报(自然科学版),2018,42(2):177-83.
[11]郑景明,罗菊春,曾德慧.森林生态系统管理的研究进展[J].北京林业大学学报,2002,24(3):103-9.
[12]黄国胜,王雪军,李晓靖.三峡库区森林健康评价[J].江西农业大学学报,2017,39(2):334-41.
[13]国家林业局.国家森林资源连续清查技术规定[R].北京:国家林业局,2004.
[14]施明辉,赵翠薇,郭志华,等.基于SOM神经网络的白河林业局森林健康分等评价[J].生态学杂志,2011,30(6):1295-303.
[15]甘敬.北京山区森林健康评价研究[D].北京:北京林业大学,2008.
[16]汪晶,罗梅,郑小贤.杉木人工林健康评价研究--以福建将乐林场为例[J].西北林学院学报,2014,29(6):195-199.
[17]陈绍玲.马尾松人工林多形地位指数曲线模型的建模方法[J].中南林业科技大学学报,2008,28(2):125-128.
[18]何晓群.多元统计分析(第2版)[M].北京:中国人民大学出版社,2008.
[19]BARTLETT M S.Tests of significance in factor analysis[J].British Journal of Mathematical&Statistical Psychology,1950,3(2):77-85.
[20]KAISER H F.An index of factorial simplicity[J].Psychometrika,1974,39(1):31-6.
[21]刘恩田,赵忠,宋西德,等.渭北黄土高原刺槐林健康评价指标体系的构建[J].西北农林科技大学学报(自然科学版),2010,38(10):67-75.
[22]杨枝林,刘君昂,周国英,等.衡山马尾松人工林林分活力与健康状况评价[J].中国农学通报,2012,28(19):42-7.
[23]楚春晖,佘济云,陈冬洋.基于SOM神经网络的五指山市森林健康评价[J].中南林业科技大学学报,2015,35(10):69-73.
[24]刘素芝,何小东,李建军.基于知识粒度的森林生态系统健康评价指标赋权方法[J].生态学杂志,2014,33(4):1082-1088.
[25]田耀武,黄志霖,肖文发,等.基于物元模型和小班尺度的兰陵溪流域森林生态系统健康评价[J].生态学杂志,2017,36(5):1458-1464.
[26]李静锐.基于人工神经网络模型的森林生态系统健康评价[D].北京:北京林业大学,2007.
[27]谭三清,王湘衡,肖维,等.基于复杂网络的森林健康评价研究[J].中南林业科技大学学报,2015,35(8):13-16.
[2]BURGER J A,ADAMS M B,CHANG S.Management effects on growth,production and sustainability of managed forest ecosystems:past trends and future directions[J].Forest Ecology&Management,2009,258(10):2335-46.
[3]甘敬,张振明,余新晓,等.森林健康监测与评价研究[J].水土保持研究,2006,13(3):177-80.
[4]MISTRETTA P A.Managing for forest health[J].Journal of Forestry Washington,2002,100(7):24-7.
[5]MENG J,LI S,WANG W,et al.Mapping Forest Health Using Spectral and Textural Information Extracted from SPOT-5Satellite Images[J].Remote Sensing,2016,8(9):719.
[6]COSTANZA R,NORTON B G,HASKELL B D.Ecosystem health:new goals for environmental management[J].Ecosystem Health New Goals for Environmental Management,1992,28:33-34.
[7]LUCIER A A.Criteria for success in managing forested landscapes[J].Journal of Forestry,1994,92(7):20-24.
[8]肖风劲,欧阳华,孙江华,等.森林生态系统健康评价指标与方法[J].林业资源管理,2004(1):27-30.
[9]甘敬,朱建刚,张国祯,等.基于BP神经网络确立森林健康快速评价指标[J].林业科学,2007,43(12):1-7.
[10]王秋燕,陈鹏飞,李学东,等.森林健康评价方法综述[J].南京林业大学学报(自然科学版),2018,42(2):177-83.
[11]郑景明,罗菊春,曾德慧.森林生态系统管理的研究进展[J].北京林业大学学报,2002,24(3):103-9.
[12]黄国胜,王雪军,李晓靖.三峡库区森林健康评价[J].江西农业大学学报,2017,39(2):334-41.
[13]国家林业局.国家森林资源连续清查技术规定[R].北京:国家林业局,2004.
[14]施明辉,赵翠薇,郭志华,等.基于SOM神经网络的白河林业局森林健康分等评价[J].生态学杂志,2011,30(6):1295-303.
[15]甘敬.北京山区森林健康评价研究[D].北京:北京林业大学,2008.
[16]汪晶,罗梅,郑小贤.杉木人工林健康评价研究--以福建将乐林场为例[J].西北林学院学报,2014,29(6):195-199.
[17]陈绍玲.马尾松人工林多形地位指数曲线模型的建模方法[J].中南林业科技大学学报,2008,28(2):125-128.
[18]何晓群.多元统计分析(第2版)[M].北京:中国人民大学出版社,2008.
[19]BARTLETT M S.Tests of significance in factor analysis[J].British Journal of Mathematical&Statistical Psychology,1950,3(2):77-85.
[20]KAISER H F.An index of factorial simplicity[J].Psychometrika,1974,39(1):31-6.
[21]刘恩田,赵忠,宋西德,等.渭北黄土高原刺槐林健康评价指标体系的构建[J].西北农林科技大学学报(自然科学版),2010,38(10):67-75.
[22]杨枝林,刘君昂,周国英,等.衡山马尾松人工林林分活力与健康状况评价[J].中国农学通报,2012,28(19):42-7.
[23]楚春晖,佘济云,陈冬洋.基于SOM神经网络的五指山市森林健康评价[J].中南林业科技大学学报,2015,35(10):69-73.
[24]刘素芝,何小东,李建军.基于知识粒度的森林生态系统健康评价指标赋权方法[J].生态学杂志,2014,33(4):1082-1088.
[25]田耀武,黄志霖,肖文发,等.基于物元模型和小班尺度的兰陵溪流域森林生态系统健康评价[J].生态学杂志,2017,36(5):1458-1464.
[26]李静锐.基于人工神经网络模型的森林生态系统健康评价[D].北京:北京林业大学,2007.
[27]谭三清,王湘衡,肖维,等.基于复杂网络的森林健康评价研究[J].中南林业科技大学学报,2015,35(8):13-16.