广东省碳排放量测算、影响因素分析及预测模型选择  

Measurement of Carbon Emissions,Analysis of Influencing Factors and Selection of Forecasting Models in Guangdong Province

作　　者：任峰[1] 龙定洪 REN Feng;LONG Dinghong(School of Business and Administration,North China Electric Power University,Baoding Hebei 071003,China)

机构地区：[1]华北电力大学经济管理学院,河北保定071003

出　　处：《生态经济》2022年第5期21-27,32,共8页Ecological Economy

基　　金：河北省社会科学基金项目“供给侧结构性改革视角下京津冀城市群TFEE及其影响因素研究”(HB17GL068)。

摘　　要：广东省是中国经济最发达的省份之一,科学预测其碳排放量对“双碳”背景下探索省级控碳发展路径具有重要意义。论文首先对各类碳源和碳汇进行核算,得到了广东省1995—2019年的净碳排放量;运用扩展的STIRPAT模型分析广东省碳排放的影响因素,并将其作为碳排放量预测模型的输入变量;建立经鸡群算法优化的快速学习网CSO-FLN预测模型,对广东省碳排放量进行预测,并将预测结果与FLN、ELM模型进行比较分析以验证CSO-FLN模型的优越性和有效性。研究结果表明:广东省的净碳排放量在研究期间呈现上升趋势;人口规模、富裕度、产业结构与碳排放量之间均存在正相关关系,对外开放、技术水平、能源结构与碳排放量之间存在负相关关系;CSO-FLN模型的预测精度高于FLN、ELM模型,可作为后续碳排放量预测的有效建模方法。Guangdong is one of the most economically developed provinces in China,and scientific prediction of its carbon emissions is of great significance to explore low-carbon development at provincial level under the background of“double carbon”strategy.In this paper,the net carbon emissions of Guangdong from 1995 to 2019 are obtained by calculating various carbon sources and sinks,and the extended STIRPAT model is applied to analyze the influencing factors of carbon emissions,which are input variables of the carbon emission forecasting model.On this basis,a Fast Learning Network(FLN)prediction algorithm improved by Chicken Swarm Optimization is constructed to forecast the carbon emissions of Guangdong Province,the superiority and effectiveness of the CSO-FLN model is verified by comparing the prediction accuracy with FLN and ELM models.The results show that the net carbon emissions of Guangdong have generally shown an upward trend during the study period;the population size,wealth and industrial structure can promote the increase of carbon emissions,while openness,technological level,energy structure all play negative effects;the CSO-FLN model is superior to the FLN and ELM models in prediction accuracy,and can be used as an effective modeling method for follow-up carbon emission research.

关 键 词：碳排放 STIRPAT模型 鸡群算法 快速学习网 

分 类 号：F062.2[经济管理—政治经济学]