上海管理科学2025，Vol.47Issue(6)：51-56,6.
基于BP神经网络的林业生物量预测
Forest Biomass Prediction Based on BP Neural Network
摘要
Abstract
Climate warming has become a hot issue in today's world,and forestry carbon sinks have at-tracted much attention because of their potential in maintaining global climate stability.Forestry car-bon sequestration refers to the preservation of CO2 in the air in plants and soil,and the absorption of forest to reduce the concentration of CO2 in the air.The key of forest carbon sink research is whether the estimation of forest biomass and forest carbon sink is accurate and fast.In this paper,BP neural network is used to estimate forest biomass.Seven parameters,such as DBH,tree height,altitude,soil type,slope direction,slope and geographical location,were taken as input layer parameters.Meanwhile,there were 448 nodes in the hidden layer of the model,and the variables in the output layer were above-ground biomass and subsurface biomass.A program is written to train the model on the Python software platform,and the BP neural network model is optimized internally.The R2 of the final model is 0.9992 when estimating above-ground biomass and 0.9829 when estimating subsur-face biomass,which is more accurate than the traditional binary regression model.关键词
BP神经网络/模型/林业碳汇/地上生物量Key words
BP neural network/the model/forestry carbon sink/aboveground biomass分类
农业科技引用本文复制引用
曹泽康,丁宁..基于BP神经网络的林业生物量预测[J].上海管理科学,2025,47(6):51-56,6.基金项目
浙江省自然科学基金白马湖实验室区域创新发展联合基金资助项目(LBMHY24E060014) （LBMHY24E060014）
