| 基于Lasso-Transformer神经网络模型的海南省碳排放预测 |
| 摘要点击 523 全文点击 24 投稿时间：2025-01-25 修订日期：2025-04-01 |
| 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 碳排放 Lasso-Transformer神经网络模型 海南省 LMDI模型 预测 |
| 英文关键词 carbon emission Lasso-Transformer neural network model Hainan Province LMDI model forecast |
| DOI 10.13227/j.hjkx.20260211 |
| 作者 | 单位 | E-mail | | 金雨洁 | 南京大学地理与海洋科学学院, 南京 210023
自然资源部碳中和与国土空间优化重点实验室, 南京 210023 | jinyujie@smail.nju.edu.cn | | 金晓斌 | 南京大学地理与海洋科学学院, 南京 210023
自然资源部碳中和与国土空间优化重点实验室, 南京 210023 | jinxb@nju.edu.cn | | 洪星明 | 南京大学集成电路学院, 苏州 215163 | | | 张舟遥 | 南京大学地理与海洋科学学院, 南京 210023
自然资源部碳中和与国土空间优化重点实验室, 南京 210023 | | | 韩博 | 南京大学地理与海洋科学学院, 南京 210023
自然资源部碳中和与国土空间优化重点实验室, 南京 210023 | | | 周寅康 | 南京大学地理与海洋科学学院, 南京 210023
自然资源部碳中和与国土空间优化重点实验室, 南京 210023 | |
|
| 中文摘要 |
| 海南省作为我国生态文明试验区和自由贸易港，在当前“碳达峰、碳中和”战略实施背景下，承担着减排降碳与经济协同发展的重要任务. 在对海南省2004~2023年的碳源、碳汇和净碳排放量核算基础上，运用LMDI模型和Lasso分析对海南省碳排放影响因素进行分解和筛选，并纳入4种Lasso-Transformer神经网络模型对2024~2030年海南省碳排放量进行预测. 结果表明：①海南省2004~2023年碳汇总量变化趋势较为平稳，净碳排放量变化趋势与碳源总量基本保持一致. ②海南省碳排放的主要影响因素由强到弱分别为：能源强度、土地碳排放强度、经济效率、土地利用结构、人口规模和土地利用效率. ③通过模型优选，利用Lasso-PatchTST模型对海南省2024~2030年碳排放量及各影响因素进行预测，得出2030年的碳排放量为4 345.53万t，土地利用效率因素增长速度最快，人口规模因素增长速度最慢. 通过优化产业结构、提升资源利用效率并加强生态系统保护，可以促进海南省减排降碳与经济协调发展. 研究结果可为海南省低碳经济发展提供决策参考. |
| 英文摘要 |
| As an important ecological civilization pilot zone and free trade port in China， Hainan Province undertakes the important task of coordinated development of carbon reduction and economic development under the background of the implementation of the strategy of "carbon peak and carbon neutrality." Based on the calculation of carbon source， carbon sink， and net carbon emissions in Hainan Province from 2004 to 2023， the LMDI model and Lasso analysis were used to decompose and screen the influencing factors of carbon emissions in Hainan Province， and four Lasso-Transformer neural network models were included to predict carbon emissions in Hainan Province from 2024 to 2030. The results showed that： ① The trend of total carbon sink in Hainan Province from 2004 to 2023 was relatively stable， and the change trend of net carbon emission was basically consistent with the total carbon source. ② The main influencing factors of carbon emissions in Hainan Province were energy intensity， land carbon emission intensity， economic efficiency， land use structure， population size， and land use efficiency. ③ Through model optimization， the Lasso-PatchTST model was used to predict the carbon emission of Hainan Province from 2024 to 2030 and its influencing factors， and the carbon emission in 2030 was predicted to be 43，455，300 tons. The growth rate of land use efficiency factor was the fastest， and the growth rate of population size was the slowest. By optimizing industrial structure， improving resource utilization efficiency and strengthening ecosystem protection， it can promote the coordinated development of carbon reduction and economy in Hainan Province. The results of this study can provide a reference for decision-making of low-carbon economic development in Hainan Province. |
