基于STIRPAT模型的上海市碳达峰预测研究  

Research on Shanghai Carbon Peak Forecast Based on STIRPAT Model

作　　者：姚明秀 王淼薇[2] 雷一东[1] YAO Mingxiu;WANG Miaowei;LEI Yidong(Department of Environmental Science and Engineering,Fudan University,Shanghai 200438,China;CNAO􀆳s Shanghai Resident Office,Shanghai 200051,China)

机构地区：[1]复旦大学环境科学与工程系,上海200438 [2]中华人民共和国审计署驻上海特派员办事处,上海200051

出　　处：《复旦学报（自然科学版）》2023年第2期226-237,共12页Journal of Fudan University：Natural Science

基　　金：上海市生态环境局科研项目(沪环科[2021]第39号)。

摘　　要：本文根据1997年—2019年上海市碳排放相关数据,筛选后确定人口、人均GDP、能源强度、能源结构以及交通周转量5个特征因子建立STIRPAT模型,利用Python的sklearn机器学习库中的岭回归模型对方程进行拟合,采用划分数据集的方式选取岭回归合适的超参数α,最终得到上海市碳排放预测模型。采用每5年设置一次变化率、5年内变化率匀速下降的方式设定未来特征变化率,且各特征值变化率均采用等差方式设定高中低3档,组合形成5种情景,运用情景分析法对上海市2040年前的碳排放量做出预测:在低碳发展情景与节能情景下,上海市可以完成2025年前碳达峰的目标;在中间基准情景、理想情景与自由发展情景下,无法完成2025年前碳达峰的目标。根据碳排放预测模型系数,能源强度、人均GDP增速和人口数量是降低上海市碳排放总量的关键影响因素,建议上海市优先选择节能情景模式,严格控制能源消耗总量,优化能源结构,降低能源强度,同时适当降低经济发展速度,有利于实现双碳目标。Based on Shanghai's carbon emissions related data from 1997 to 2019,this paper determines the five characteristic factors of population,per capita GDP,energy intensity,energy structure,and traffic turnover after screening to establish the STIRPAT model,using the ridge regression model in the sklearn machine learning library of Python fitting the equation,selecting the appropriate hyperparameterαfor ridge regression by dividing the data set,and finally obtaining the Shanghai's carbon emission prediction model.The change rate is set every 5 years and the rate of change decreases uniformly within 5 years to set the future feature change rate,and the change rate of each feature value is set by the arithmetic method to set high,middle and low grades,combining to form 5 scenarios,using scenarios analytical method predicts Shanghai's carbon emissions before 2040:under lowcarbon development scenarios and energy-saving scenarios,Shanghai can achieve its goal of peaking carbon before 2025;in the middle baseline scenario,ideal scenario and free development scenario,the goal of reaching the peak by 2025 cannot be achieved.According to the carbon emission prediction model coefficients,energy intensity,per capita GDP growth and population are the key factors in reducing Shanghai's total carbon emissions.It is suggested that Shanghai should give priority to the energy-saving scenario mode,strictly control the total energy consumption,optimize the energy structure,reduce energy intensity,and appropriately reduce the economy.The speed of development is conducive to achieving the dual-carbon goal.

关 键 词：STIRPAT模型 碳达峰 低碳发展 上海市 

分 类 号：X24[环境科学与工程—环境科学]