| 基于组合模型及多情景模拟的中国西部地区建筑业碳排放峰值预测 |
| 摘要点击 677 全文点击 14 投稿时间：2025-06-29 修订日期：2025-09-03 |
| 查看HTML全文
查看全文 查看/发表评论 下载PDF阅读器 |
| 中文关键词 中国西部地区 建筑业碳排放 峰值预测 多情景模拟 机器学习方法 |
| 英文关键词 Western China carbon emissions of the construction industry peak prediction multi-scenario simulation machine learning methods |
| DOI 10.13227/j.hjkx.202506347 |
|
| 中文摘要 |
| 旨在科学评估中国西部地区建筑业碳排放的演变规律，识别关键驱动因子，并通过模型优化构建碳排放预测体系，为平衡经济增长与“双碳”目标提供决策依据. 基于IPCC碳排放系数法测算西部地区2003~2022年建筑业碳排放量. 通过融合LASSO-XGBoost-RFE特征筛选和SPBO-BP优化模型构建了预测框架，并模拟5种情景探索达峰路径. 结果表明：①西部地区建筑业碳排放呈现“上升-波动-平稳”的阶段性特征，间接碳排放占比高达93.9%；②城镇化率、建筑业能源强度、固定资产投资、建筑业房屋竣工面积及第二产业增加值是关键影响因子；③相较于其他基准模型，目标模型的预测性能显著提升，R2达到0.959 7；④极低碳、低碳和基准情景分别于2028年（6.12亿t）、2029年（6.50亿t）和2030年（7.04亿t）达峰，而高碳和极高碳情景则分别在2039年（8.67亿t）和2040年（12.65亿t）达峰. 研究成果可为我国西部地区建筑业的低碳发展提供重要理论支撑与技术参考. |
| 英文摘要 |
| The objective of this study is to scientifically assess the evolution pattern of carbon emissions in the construction industry in China's western region， identify key driving factors， and construct a carbon emission prediction system through model optimization， providing a decision-making basis for balancing economic growth and the “dual carbon” goals. Based on the IPCC carbon emission coefficient method， the carbon emissions of the construction industry in the western region from 2003 to 2022 were calculated. A prediction framework was constructed by integrating LASSO-XGBoost-RFE feature selection and SPBO-BP optimization models， and five scenarios were simulated to explore the peak path. The research shows that： ① The carbon emissions of the construction industry in the western region exhibited a phased characteristic of “rise-fluctuation-stability，” with indirect carbon emissions accounting for 93.9%. ② Urbanization rate， energy intensity of the construction industry， fixed asset investment， completed floor area of construction industry buildings， and added value of the secondary industry were key influencing factors. ③ Compared with other benchmark models， the prediction performance of the target model was significantly improved， with R2 reaching 0.959 7. ④ The very low-carbon， low-carbon， and benchmark scenarios were predicted to peak in 2028 （612 million tons）， 2029 （650 million tons）， and 2030 （704 million tons）， respectively， while the high-carbon and very high-carbon scenarios would peak in 2039 （867 million tons） and 2040 （1265 million tons）. The research results can provide important theoretical support and technical references for the low-carbon development of the construction industry in China's western region. |
