数据与计算发展前沿 ›› 2026, Vol. 8 ›› Issue (1): 168-182.

CSTR: 32002.14.jfdc.CN10-1649/TP.2026.01.014

doi: 10.11871/jfdc.issn.2096-742X.2026.01.014

• 技术与应用 • 上一篇    下一篇

基于工作流的陆地生态系统碳循环实时同化预测系统

万萌1(),何洪林2,3,4,任小丽2,3,聂宁明1,*(),曹荣强1,王宗国1,李凯1,王晓光1,王彦棡1,王珏1,高超4   

  1. 1.中国科学院计算机网络信息中心，北京 100083
    2.中国科学院地理科学与资源研究所生态系统网络观测与模拟重点实验室，北京 100101
    3.国家生态科学数据中心，北京 100101
    4.中国科学院大学，资源与环境学院，北京 100190
  • 收稿日期:2025-02-13 出版日期:2026-02-20 发布日期:2026-02-02
  • 通讯作者: 聂宁明
  • 作者简介:万萌，中国科学院计算机网络信息中心，工程师，主要研究方向为人工智能应用及平台，时间序列预测等。
    本文主要承担工作为数据处理，实验设计，模型构建及学习训练。
    WAN Meng is an engineer at the Computer Network Information Center, Chinese Academy of Sciences. His primary research interests include artificial intelligence applications and platforms, as well as time series prediction.
    In this paper, he is responsible for data processing, experimental design, system construction and training.
    E-mail: wanmengdamon@cnic.cn|聂宁明，中国科学院计算机网络信息中心，副研究员，主要研究方向为人工智能算法与应用软件。
    本文主要承担工作为算法设计。
    NIE Ningming is an associate professor at the Computer Network Information Center, Chinese Academy of Sciences. Her research interests include artificial intelligence algorithm and application software.
    In this paper, she is responsible for algorithm design.
    E-mail: nienm@sccas.cn
  • 基金资助:
    国家重点研发计划“标准化生态台站监测数据产品体系构建与系统开发”(2021YFF0703902)

The Real-Time Assimilation and Prediction System for Terrestrial Ecosystem Carbon Cycling Based on Workflow

WAN Meng1(),HE Honglin2,3,4,REN Xiaoli2,3,NIE Ningming1,*(),CAO Rongqiang1,WANG Zongguo1,LI Kai1,WANG Xiaoguang1,WANG Yangang1,WANG Jue1,GAO Chao4   

  1. 1. Computer Network Information Center, Chinese Academy of Sciences, Beijing 100083, China
    2. Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences andNatural Resources Research, Chinese Academy of Sciences, Beijing 100101, China
    3. National Ecosystem Science Data Center, Beijing 100101, China
    4. College of Resources and Environment, University of Chinese Academy of Sciences, Beijing 100190, China
  • Received:2025-02-13 Online:2026-02-20 Published:2026-02-02
  • Contact: NIE Ningming

RichHTML

1

PDF

0

摘要：

【目的】陆地生态系统碳循环对气候变化的反馈作用是全球变化研究的核心。目前缺乏实时、高效的自动化碳源汇评估和预测系统，难以快速准确地对全国陆地生态系统的碳源汇大小、稳定性和可持续性等进行定量评价。【方法】本研究构建了一套陆地生态系统碳循环实时同化预测系统，包括数据采集、传输、分析、工作流、调度、预测和可视化等多个核心模块。通过结合深度学习气象模型、碳循环过程模型、数据同化算法和生态迭代预测方法，不断融合实时传输的站点观测数据，实现了台站碳汇的实时短期预测，为从观测到预测的野外站科研模式提供范例。【结果】自2023年2月部署以来，系统已成功接入了鼎湖山、千烟洲、会同站等4个站点，迄今已积累超过11万条数据。【结论】系统显著提升了碳循环预测的实时性和效率，为生态研究和环境管理决策提供了可靠的数据支持和可观测的实时检索服务。

关键词: 气候变化, 碳循环, 同化预测, 深度学习, 生态系统, 数据同化, 短期预测

Abstract:

[Objective] The feedback effects of terrestrial ecosystem carbon cycling on climate change are central to global change research. Currently, there is a lack of real-time, efficient, and automated carbon source-sink assessment and prediction systems, making it difficult to quickly and accurately quantify the carbon sink size, stability, and sustainability. This limitation affects the formulation of carbon sequestration strategies and the implementation of carbon neutrality initiatives. [Methods] This study develops a real-time assimilation and prediction system for terrestrial ecosystem carbon cycling, comprising multiple core modules such as data collection, transmission, analysis, workflow management, scheduling, prediction, and visualization. By integrating deep learning-based meteorological models, carbon cycle process models, data assimilation algorithms, and ecological iterative prediction methods, the system continuously assimilates real-time station observation data to enable short-term carbon sink predictions, which serves as a paradigm for transitioning from observation to prediction in field station research. [Results] Since its deployment in February 2023, the system has successfully integrated with four stations, including Dinghushan, Qianyanzhou, and Huitong, accumulating over 110,000 data records to date. [Conclusion] The system significantly improves the timeliness and efficiency of carbon cycle prediction, providing reliable data support and observable real-time retrieval services for ecological research and environmental management decision-making.

Key words: climate change, carbon cycle, assimilation prediction, deep learning, ecosystem, data assimilation, short-term prediction