基于IPSO-BP神经网络模型的山东省碳排放预测及生态经济分析  

Carbon emissions prediction based on IPSO-BP neural network model and eco-economic analysis of Shandong province

作　　者：张迪[1,4] 王彤彤 支金虎[1,4] ZHANG Di;WANG Tongtong;ZHI Jinhu(College of Plant Sciences,Tarim University,Alar 843300,China;College of Natural Resources and Environment,Northwest A&F University,Yangling 712100,China;Chongqing branch,Changjiang River Scientific research institute of Changjiang Water Resources Commission,Chongqing 400026,China;Research Center of Oasis Agricultural Resources and Environment in Sourthern Xinjang,Tarim University,Alar Xinjiang 843300,China)

机构地区：[1]塔里木大学植物科学学院,阿拉尔843300 [2]西北农林科技大学资源环境学院,杨凌712100 [3]长江水利委员会长江科学院重庆分院,重庆400026 [4]塔里木大学南疆绿洲农业资源与环境研究中心,阿拉尔843300

出　　处：《生态科学》2022年第1期149-158,共10页Ecological Science

基　　金：国家重点研发计划(2017YFC0504300,2017YFD0201900);环境材料与修复技术重庆市重点实验室开放基金(CEK1805)。

摘　　要：在低碳经济发展背景下,针对山东省碳排放数据更新迟缓、以往预测模型难以满足现实需求的问题,统计相关数据,根据政府间气候变化专门委员会(IPCC)推荐方法测算山东省2000—2017年碳排放量和排放强度,并借助脱钩分析、碳承载力和碳赤字探究碳排放的动态变化趋势;基于5项最重要的碳排放影响因素,建立改进的粒子群算法(IPSO)优化BP神经网络模型,对山东省的碳排放量和排放强度进行仿真预测。结果表明:山东省工业耗能占总量的78.5%左右。2000—2017年间山东省碳排放量呈增长趋势,年平均为52328.10万吨;碳排放强度却呈下降趋势,年平均为1.73万吨/亿元。总体而言,2000—2017年间山东省碳排放量与GDP之间呈弱脱钩的态势,碳承载力呈先增长后减小的趋势,18年间碳承载力减少了8%,全省从2005年开始出现碳赤字,并呈现增大趋势。IPSO算法明显优化了BP神经网络,误差更小、精度更高,更适于碳排放量及相关指标的预测。预测结果显示山东省未来碳排放量呈缓慢增长趋势,而碳排放强度有所降低,以期为政府决策提供科学依据。In a low-carbon economic context,the prior prediction models cannot meet the requirement due to the slow update of date regarding the carbon emission in Shandong province.According to the statistical yearbook and the IPCC guideline,this paper estimates the carbon emissions and emission intensity of Shandong Province from 2000 to 2017.The dynamic behavior for carbon emissions is analyzed with carbon carrying capacity, carbon deficit and decoupling. Based on the five most important carbon emission factors, this paper proposes an improved particle swarm optimization algorithm (IPSO) optimized BP neural network model to simulate the carbon emissions and emission intensity for Shandong province. The result showed that the industrial energy consumption in Shandong Province accounts for about 78.5 % of the total. From 2000 to 2017, the carbon emissions showed an increase, and with an average carbon emission of 523.281 million tons. However, the intensity of carbon emissions described a downward trend, with an average of 17,300 tons per 100 million yuan. In general, the carbon emissions and GDP of Shandong province showed a weak decoupling trend between 2000 and 2017. The carbon carrying capacity showed an increasing trend first and then decreasing. The carbon carrying capacity decreased by 8% in 18 years. Since 2005, the carbon deficit has appeared and shown an increasing trend. Obviously, the IPSO algorithm optimized BP neural network with smaller error and higher accuracy, which is more suitable for the prediction regarding carbon emissions and related indicators. The prediction results provide a scientific basis for government decision-making because the carbon emissions of Shandong province will show a slow growth trend in the future and the carbon emission intensity will be reduced simultaneously.

关 键 词：BP神经网络 IPSO优化算法 碳排放 预测 山东省 

分 类 号：X196[环境科学与工程—环境科学]