基于清查数据的福建省马尾松生物量转换和扩展因子估算差异解析——3种集成学习决策树模型的比较

doi:10.13287/j.1001-9332.201806.019

应用生态学报 ›› 2018, Vol. 29 ›› Issue (6): 2007-2016.doi: 10.13287/j.1001-9332.201806.019

基于清查数据的福建省马尾松生物量转换和扩展因子估算差异解析——3种集成学习决策树模型的比较

欧强新¹,李海奎^1*,雷相东¹,杨英²

¹中国林业科学研究院资源信息研究所, 北京 100091;
²国家林业局调查规划设计院, 北京 100714

收稿日期:2017-10-21 修回日期:2018-03-19 出版日期:2018-06-18 发布日期:2018-06-18
通讯作者: E-mail: lihk@ifrit.ac.cn
作者简介:欧强新,男,1990年生,博士研究生.主要从事森林生长模型与模拟研究. E-mail: jonsinou@foxmail.com
基金资助:
本文由国家林业公益性行业科研专项(201504303)资助

Difference analysis in estimating biomass conversion and expansion factors of masson pine in Fujian Province, China based on national forest inventory data: A comparison of three decision tree models of ensemble learning.

OU Qiang-xin¹, LI Hai-kui^1*, LEI Xiang-dong¹, YANG Ying²

¹Research Institute of Forest Resource Information Techniques, Chinese Academy of Forestry, Beijing 100091, China;
²Academy of Forestry Inventory and Planning, State Forestry Administration, Beijing, 100714, China

Received:2017-10-21 Revised:2018-03-19 Online:2018-06-18 Published:2018-06-18
Supported by:
This work was supported by the National Forestry Industry Research Special Funds for Public Welfare Projects (201504303)

摘要/Abstract

摘要： 生物量转换和扩展因子(BCEFs)是估算森林生物量碳储量普遍使用的重要参数.厘清BCEFs估算差异的来源,可降低森林生物量碳储量评估的不确定性.利用基于集成学习的决策树模型能够很好地解决BCEFs估算差异来源问题.然而,不同此类模型的对比研究目前尚未见报道.本研究以第8次国家森林资源清查福建省331块马尾松的固定样地数据作为材料,分别利用增强回归树(BRT)、随机森林(RF)和立体派(Cubist)模型分析BCEFs(包括地上和地下部分)估算差异的来源.结果表明: 研究区马尾松BCEFs呈右偏分布,平均值为0.69 t·m^-3,最小值为0.67 t·m^-3,最大值为0.71 t·m^-3.BRT、RF和Cubist模型对BCEFs的拟合和预测能力均很好,均能够解释92.8%以上的BCEFs变异.3种模型均给出了相同的前2个相对贡献率最大的自变量,为平均胸径和蓄积量.BCEFs随着平均胸径、蓄积量的增加呈逐渐减小的趋势.平均胸径、蓄积量、平均年龄和平均树高等林分特征因子对BCEFs的影响极大,而土壤因子和地形因子对BCEFs的影响均很小.在建立BCEFs模型时,利用平均胸径、蓄积量、平均年龄和平均树高等少量包含较多BCEFs预测信息的变量便能获取很好的预估精度.当应用固定BCEFs时,应选择在平均年龄、平均胸径以及蓄积等方面具有广泛代表性的样本计算BCEFs.

Abstract: Biomass conversion and expansion factors (BCEFs) are important parameters for estimating carbon storage in forest biomass. Clarifying the source of differences in estimating BCEFs could reduce uncertainties in forest biomass carbon estimation. The decision tree models of ensemble learning can be used to properly figure out the source of differences in estimating BCEFs. However, the comparison of different decision tree models for analyzing differences in estimating BCEFs has never been reported. In this study, three models [the boosted regression trees (BRT), random forest(RF), and Cubist] and data of 331 masson pine plots from the 8^th Chinese National Forest Inventory for Fujian Province were used to analyze the differences in estimating BCEFs (including above- and below-ground). The results showed that BCEFs were following right-skewed distribution, with the mean, minimum and maximum value being 0.69 t·m^-3, 0.67 t·m^-3and 0.71 t·m^-3, respectively. All three models performed well in BCEFs prediction and fitting, and could explain more than 92.8% variations of BCEFs. All three models showed that average DBH and volume were the top two highest relative importance predictors. BCEFs decreased with the increases of average DBH and volume. Stand characteristics factors, such as average DBH, volume, average age and average height, had great influence on BCEFs. Both soil factors and topographic factors had little influence on BCEFs. Using a few variables (such as average DBH, volume, average age and avera-ge height) which contained more BCEFs prediction information could have preferable forecasting precision when building BCEFs models. Moreover, widely representative samples with different average tree ages, average DBH and volume should be chosen to calculate BCEFs when applying constant BCEFs.

欧强新,李海奎,雷相东,杨英. 基于清查数据的福建省马尾松生物量转换和扩展因子估算差异解析——3种集成学习决策树模型的比较[J]. 应用生态学报, 2018, 29(6): 2007-2016.

OU Qiang-xin, LI Hai-kui, LEI Xiang-dong, YANG Ying. Difference analysis in estimating biomass conversion and expansion factors of masson pine in Fujian Province, China based on national forest inventory data: A comparison of three decision tree models of ensemble learning.[J]. Chinese Journal of Applied Ecology, 2018, 29(6): 2007-2016.

参考文献

[1] Luo YJ, Zhang XQ, Wang XK, et al. Dissecting variation in biomass conversion factors across China’s fore-sts: Implications for biomass and carbon accounting. PLoS One, 2014, 9(4): e94777
[2] Houghton RA, Hall F, Goetz SJ. Importance of biomass in the global carbon cycle. Journal of Geophysical Research Biogeosciences, 2009, 114: G00E03, doi: 10.1029/2009JG000935
[3] Luo Y-J (罗云建), Zhang X-Q (张小全), Wang X-K (王效科), et al. Forest biomass estimation methods and their prospects. Scientia Silvae Sinicae (林业科学), 2009, 45(8): 129-134 (in Chinese)
[4] Mcroberts RE, Nsset E, Gobakken T, et al. Indirect and direct estimation of forest biomass change using forest inventory and airborne laser scanning data. Remote Sensing of Environment, 2015, 164: 36-42
[5] Brown SL, Schroeder P, Kern JS. Spatial distribution of biomass in forests of the eastern USA. Forest Ecology and Management, 1999, 123: 81-90
[6] Zuo S-D (左舒翟), Ren Y (任引), Wang X-K (王效科), et al. Biomass estimation factors and their determinants of Cunninghamia lanceolata forests in China. Scientia Silvae Sinicae (林业科学), 2014, 50(11): 1-12 (in Chinese)
[7] Petersson H, Holm S, St hl G, et al. Individual tree bio-mass equations or biomass expansion factors for assessment of carbon stock changes in living biomass: A comparative study. Forest Ecology and Management, 2012, 270: 78-84
[8] Johnson WC, Sharpe DM. The ratio of total to merchan-table forest biomass and its application. Canadian Journal of Forest Research, 1983, 13: 372-383
[9] Lim H, Lee KH, Lee KH, et al. Biomass expansion factors and allometric equations in an age sequence for Japanese cedar (Cryptomeria japonica) in southern Korea. Journal of Forest Research, 2013, 18: 316-322
[10] Lehtonen A, Makipaa R, Heikkinen J, et al. Biomass expansion factors (BEFs) for Scots pine, Norway spruce and birch according to stand age for boreal forests. Forest Ecology and Management, 2004, 188: 211-224
[11] Ashraf MI, Zhao ZY, Bourque CPA, et al. Integrating biophysical controls in forest growth and yield predictions with artificial intelligence technology. Canadian Journal of Forest Research, 2013, 43: 1162-1171
[12] De’Ath G. Boosted trees for ecological modeling and prediction. Ecology, 2007, 88: 243-251
[13] Kuhn M, Johnson K. Trans, Lin H (林荟). Applied Predictive Modeling. Beijing: China Machine Press, 2016 (in Chinese)
[14] Yoo S, Im J, Wagner JE. Variable selection for hedonic model using machine learning approaches: A case study in Onondaga County, NY. Landscape and Urban Planning, 2012, 107: 293-306
[15] Ou Q-X (欧强新), Li H-K (李海奎), Yang Y (杨英). Factors affecting the biomass conversion and expansion factor of masson pine in Fujian Province. Acta Ecologica Sinica (生态学报), 2017, 37(17): 5756-5764 (in Chinese)
[16] Yang RM, Zhang GL, Liu F, et al. Comparison of boosted regression tree and random forest models for mapping topsoil organic carbon concentration in an alpine ecosystem. Ecological Indicators, 2016, 60: 870-878
[17] Nguyen DB, Rose MT, Rose TJ, et al. Impact of glyphosate on soil microbial biomass and respiration: A meta-analysis. Soil Biology and Biochemistry, 2016, 92: 50-57
[18] Chen K (陈凯), Yu H-Y (于海燕), Zhang J-W (张汲伟), et al. Predictive model based multimetric index of macroinvertebrates for river health assessment. Chinese Journal of Applied Ecology (应用生态学报), 2017, 28(6): 1993-2002 (in Chinese)
[19] Yin C (尹才), Liu M (刘淼), Sun F-Y (孙凤云), et al. Influencing factors of non-point source pollution of watershed based on boosted regression tree algorithm. Chinese Journal of Applied Ecology (应用生态学报), 2016, 27(3): 911-919 (in Chinese)
[20] Hao J-F (侯建峰), Lyu X-T (吕晓涛), Wang C (王超), et al. Variation of soil respiration and its underlying mechanism in grasslands of northern China. Chinese Journal of Applied Ecology (应用生态学报), 2014, 25(10): 2840-2846 (in Chinese)
[21] Forest Resources Management Department of the State Forestry Administration (国家林业局森林资源管理司). Handbook on Forest Resource Management. Beijing: China Metrology Press, 2007 (in Chinese)
[22] State Forestry Administration (国家林业局). Tree Biomass Models and Related Parameters to Carbon Accounting for Pinus massoniana (LY/T 2263-2014 ). Beijing: State Forestry Administration, 2014 (in Chinese)
[23] Zefferman EP, Harris DJ. Predicting drivers of nuisance macrophyte cover in a regulated California stream using boosted regression tree models. Journal of Aquatic Plant Management, 2016, 54: 78-86
[24] Lewin WC, Mehner T, Ritterbusch D, et al. The influence of anthropogenic shoreline changes on the littoral abundance of fish species in German lowland lakes vary-ing in depth as determined by boosted regression trees. Hydrobiologia, 2014, 724: 293-306
[25] Cheong YL, Leit o PJ, Lakes T. Assessment of land use factors associated with dengue cases in Malaysia using boosted regression trees. Spatial and Spatio-temporal Epidemiology, 2014, 10: 75-84
[26] Lyu X-L (吕晓玲), Song J (宋捷). Big Data Mining and Statistical Machine Learning. Beijing: Renmin University of China Press, 2016 (in Chinese)
[27] Chiu Y-W. Trans, Pan Y (潘怡). Machine Learning with R Cookbook. Beijing: China Machine Press, 2016 (in Chinese)
[28] Sun L (孙亮), Huang Q (黄倩). Practical Machine Learning. Beijing: Posts and Telecom Press, 2017 (in Chinese)
[29] Friedman JH. Greedy function approximation: A gra-dient boosting machine. Annals of Statistics, 2001, 29: 1189-1232
[30] Zhang L (张雷), Liu S-R (刘世荣), Sun P-S (孙鹏森), et al. Comparative evaluation of multiple models of the effects of climate change on the potential distribution of Pinus massoniana. Chinese Journal of Plant Ecology (植物生态学报), 2011, 35(11): 1091-1105 (in Chinese)
[31] Li X-H (李欣海). Using “random forest” for classification and regression. Chinese Bulletin of Entomology (应用昆虫学报), 2013, 50(4): 1190-1197 (in Chinese)
[32] Breiman L. Random forests. Machine Learning, 2001, 45: 5-32
[33] Zhang R-Y (张瑞英), Pang Y (庞勇), Li Z-Y (李增元), et al. Canopy closure estimation in a temperate forest using airborne LiDAR and LANDSAT ETM+ data. Chinese Journal of Plant Ecology (植物生态学报), 2016, 40(2): 102-115 (in Chinese)
[34] Kuhn M, Weston S, Keefer C, et al. Cubist Models for Regression [EB/OL]. (2016-12-10) [2017-10-19]. https://cran.r-project.org/web/packages/Cubist/vignettes/cubist.pdf
[35] Zhou Y (周银), Liu L-Y (刘丽雅), Lu Y-L (卢艳丽), et al. Regional scale mapping of soil organic matter using remote sensing and visible-near infrared spectroscopy. Journal of Remote Sensing (遥感学报), 2015, 19(6): 998-1006 (in Chinese)
[36] Dai S (戴舒), Fu Y-C (付迎春), Zhao Y-L (赵耀龙). The remote sensing model for estimating urban impervious surface percentage based on the Cubist model tree. Journal of Geo-Information Science (地球信息科学学报), 2016, 18(10): 1399-1409 (in Chinese)
[37] Zang H (臧颢). Regional-scale Climate-sensitive Stand Growth Models for Larch Plantations. PhD Thesis. Beijing: Chinese Academy of Forestry, 2016 (in Chinese)
[38] Liski J, Korotkov AV, Prins CF, et al. Increased carbon sink in temperate and boreal forests. Climatic Change, 2003, 61: 89-99
[39] Soares P, Tome M. Biomass expansion factors for Eucalyptus globulus stands in Portugal. Forest Systems, 2012, 21: 141-152
[40] Lehtonen A, Cienciala E, Tatarinov F, et al. Uncertainty estimation of biomass expansion factors for Norway spruce in the Czech Republic. Annals of Forest Science, 2007, 64: 133-140
[41] Peichl M, Arain MA. Allometry and partitioning of above- and belowground tree biomass in an age-sequence of white pine forests. Forest Ecology and Management, 2007, 253: 68-80
[42] Jalkanen A, M kip R, St hl G, et al. Estimation of the biomass stock of trees in Sweden: Comparison of biomass equations and age-dependent biomass expansion factors. Annals of Forest Science, 2005, 62: 845-851
[43] Tobin B, Nieuwenhuis M. Biomass expansion factors for Sitka spruce (Picea sitchensis (Bong.) Carr.) in Ireland. European Journal of Forest Research, 2007, 126: 189-196
[44] Pajtík J, Konpka B, Lukac M. Biomass functions and expansion factors in young Norway spruce (Picea abies[L.] Karst) trees. Forest Ecology and Management, 2008, 256: 1096-1103

基于清查数据的福建省马尾松生物量转换和扩展因子估算差异解析——3种集成学习决策树模型的比较

Difference analysis in estimating biomass conversion and expansion factors of masson pine in Fujian Province, China based on national forest inventory data: A comparison of three decision tree models of ensemble learning.

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 0

编辑推荐

Metrics

本文评价