Using nonparametric modeling approaches and remote sensing imagery to estimate ecological welfare forest biomass

作     者：Chaofan Wu Hongxiang Tao Manyu Zhai Yi Lin Ke Wang Jinsong Deng Aihua Shen Muye Gan Jun Li Hong Yang 

作者机构：College of Environmental and Resource Sciences Zhejiang University Hangzhou 310058 China Zhejiang Forestry Academy Hangzhou 310023 China School of Civil Engineering and Environmental Sciences and School of Meteorology University of Oklahoma Norman OK 73019 USA State Key Laboratory of Hydro Science and Engineering Department of Hydraulic Engineering Tsinghua University Beijing 100084 China 

出 版 物：《Journal of Forestry Research》 (林业研究（英文版）)

年 卷 期：2018年第29卷第1期

页      面：151-161页

核心收录：

学科分类：09[农学] 0903[农学-农业资源与环境] 

基　　金：support of Chinese Ministry of Environmental Protection(No.STSN-05-11) Ministry of Science and Technology of the People’s Republic of China(No.2015BAC02B00) Science Technology Department of Zhejiang Province(No.2015F50056) 

主　　题：Above ground biomass Ecological forest Forest management Landsat Random forest 

摘      要：The spatial distribution of forest biomass is closely related with carbon cycle, climate change, forest productivity, and biodiversity. Efficient quantification of biomass provides important information about forest quality and health. With the rising awareness of sustainable development, the ecological benefits of forest biomass attract more attention compared to traditional wood supply function. In this study, two nonparametric modeling approaches, random forest（RF） and support vector machine were adopted to estimate above ground biomass（AGB） using widely used Landsat imagery in the region,especially within the ecological forest of Fuyang District in Zhejiang Province, China. Correlation analysis was accomplished and model parameters were optimized during the modeling process. As a result, the best performance modeling method RF was implemented to produce an AGB estimation map. The predicted map of AGB in the study area showed obvious spatial variability and demonstrated that within the current ecological forest zone, as well as the protected areas, the average of AGB were higher than the ordinary forest. The quantification of AGB was proven to have a close relationship with the local forest policy and management pattern, which indicated that combining remote-sensing imagery and forest biophysical property would provide considerable guidance for making beneficial decisions.