| 邱子健,李天玲,申卫收.江苏省农田生态系统固碳时空分布特征与趋势预测[J].农业环境科学学报,2024,43(1):226-236. |
| 江苏省农田生态系统固碳时空分布特征与趋势预测 |
| Spatiotemporal distribution characteristics and trend prediction of carbon sequestration in farmland ecosystems in Jiangsu Province，China |
| 投稿时间：2023-02-16 |
| DOI：10.11654/jaes.2023-0110 |
| 中文关键词: 江苏省 农业碳中和 土壤固碳 农田生态碳汇 机器学习预测 |
| 英文关键词: Jiangsu Province agricultural carbon neutrality soil carbon fixation farmland ecological carbon sink machine learning prediction |
| 基金项目:江苏省发展和改革委员会碳达峰策略和路径前期研究；南京信息工程大学气候与环境治理研究院开放课题（ZKKT2022A09） |
|
| 摘要点击次数: 503 |
| 全文下载次数: 333 |
| 中文摘要: |
| 为探讨江苏省农田生态系统固碳时空分布特征及未来固碳趋势，利用固碳速率法对江苏省2005—2020年农田固碳进行估算，重点分析2005、2010、2015年和2020年时空分布特征，并运用机器学习的方法对2021—2060年全省农田生态系统固碳进行预测。结果表明：在时间序列上，江苏省近年农田生态系统固碳量整体呈现升高的趋势，2020年估算量为282.55万t·a-1 （以C计，下同），在全省陆地生态系统固碳总量中占比达20.17%；在空间分布上，固碳贡献最大的是苏北地区，无论是施用肥料还是秸秆还田贡献的固碳量，苏北地区均呈现高于苏中、苏南地区的态势；根据机器学习的重要性分析，秸秆还田量是最为重要的影响因素；两种模型中，BP神经网络相较于随机森林具有更高的预测精度，该模型预测2021—2060年农田生态系统固碳量仍会在短期内持续升高，但随后将进入较稳定的平台期，其中2021—2026年间固碳量将持续升高并达峰值，为365.26万t·a-1，而到2060年固碳量则为348.12万t·a-1。研究表明，江苏省农田生态系统固碳量已逐步提升，但未来增长速率将趋于减缓，有必要进一步强化固碳措施，重点是提升秸秆还田率及其固碳效率，同时现有研究方法也有待于进一步优化，未来应将有机肥施用、绿肥还田、轮作等因素考虑在内，从而实现对农田生态系统固碳更为精准、全面的估算。 |
| 英文摘要: |
| To explore the spatiotemporal distribution characteristics and future carbon sequestration trends of farmland ecosystems in Jiangsu Province, China, the carbon sequestration rate method was used to estimate the carbon sequestration of farmland in Jiangsu Province from 2005—2020, with a focus on analyzing the spatiotemporal distribution characteristics of 2005, 2010, 2015, and 2020. Machine learning methods were used to predict the carbon sequestration of farmland ecosystems in the province from 2021—2060. The results revealed that in terms of time series, the overall carbon sequestration of farmland ecosystems in Jiangsu Province had demonstrated an increasing trend in recent years, with an estimated 2 825 500 t·a-1 （calculated by C, the same below）in 2020, accounting for 20.17% of the total carbon sequestration of terrestrial ecosystems in the province. In terms of spatial distribution, the largest contribution of carbon sequestration was in the northern region of Jiangsu compared to the central and southern regions, owing to high contributions from both fertilizer application and straw returning. According to the importance analysis of machine learning, the amount of straw returned to the field was the most important factor. In the two models, BP neural network had a higher prediction accuracy than that of random forest. The BP neural network predicted that from 2021—2060, the carbon sequestration of farmland ecosystem would continue to increase in the short term but would then entered a more stable platform period. The carbon sequestration would continue to rise and reach a peak of 3 652 600 t · a-1 from 2021—2026, and 3 481 200 t · a-1 by 2060. Studies had revealed that the carbon sequestration capacity of agricultural ecosystems in Jiangsu Province had gradually increased but the growth rate would slow down in the future. It is necessary to further strengthen carbon sequestration measures, with a focus on improving the straw return rate and carbon sequestration efficiency. Meanwhile, existing research methods need to be further optimized. In the future, factors such as organic fertilizer application, green manure return, and rotation should be considered to achieve more accurate and comprehensive carbon sequestration of agricultural ecosystems estimation. |
| HTML
查看全文
查看/发表评论 下载PDF阅读器 |
|
|
|
