河流碳循环模型研究进展  

Review of the Models for Riverine Carbon Cycling

作　　者：刘娇娇 刘军志 宋超[3] 张蔚珍 刘勇勤 LIU Jiaojiao;LIU Junzhi;SONG Chao;ZHANG Weizhen;LIU Yongqin(Center for the Pan-third Pole Environment,Lanzhou University,Lanzhou 730000,China;College of Atmospheric Sciences,Lanzhou University,Lanzhou 730000,China;Institute of Innovation Ecology,Lanzhou University,Lanzhou 730000,China;State Key Laboratory of Tibetan Plateau Earth System,Environment and Resources(TPESER),Chinese Academy of Sciences,Beijing 100101,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]兰州大学泛第三极环境中心,甘肃兰州730000 [2]兰州大学大气科学学院,甘肃兰州730000 [3]兰州大学生态学院,甘肃兰州730000 [4]中国科学院青藏高原环境变化与地表过程重点实验室,北京100101 [5]中国科学院大学,北京100049

出　　处：《地球科学进展》2024年第2期111-123,共13页Advances in Earth Science

基　　金：国家自然科学基金面上项目(编号:42171132);甘肃省自然科学基金重点项目(编号:23JRRA1033)资助。

摘　　要：河流是连接陆地和海洋两大碳库的“重要管道”和“生物反应器”,深入理解河流碳循环过程并建立河流碳循环模型是估算区域尺度河流碳通量的重要手段。当前,对河流碳循环模型构建与应用的探讨还较为缺乏。通过文献调研对河流碳循环机理和已有模型研究进行回顾和总结,首先归纳了河流中颗粒有机碳、溶解有机碳和溶解无机碳的主要来源及相关迁移转化过程,然后对经验统计和机理过程两大类河流碳循环模型的模拟方法、应用现状和优缺点进行了综述。经验统计模型采用统计或机器学习方法建立河流碳通量与环境因子的关系,建模较为简单,但普适性和外推性较差;机理过程模型在陆面模式或水文模型中耦合河流碳循环相关过程,物理性和可靠性较强,但较为复杂。不同模型的侧重点和对河流碳循环过程的表达各不相同,适用场景也不相同。河流碳循环模拟研究目前尚处于起步阶段,现有模型对陆地和水体碳循环过程以及人类活动影响的表达普遍不足,无法准确模拟和预测河流碳循环过程的长期变化。今后应加强对河流碳循环过程的观测,提高对陆地和水体碳循环机理的认识,进而完善其在模型中的表达,提高河流碳循环模拟的精度,为中国实现“双碳”目标提供科学支撑。Rivers connect the terrestrial landscape and oceans and are considered“bioreactors”of carbon.Understanding the carbon cycling processes in rivers and constructing numerical models for riverine carbon cycling is imperative to estimate regional and global carbon budgets.The summary and discussion of the development and application of riverine carbon cycling models remains inadequate.This study reviewed the mechanisms and models of riverine carbon cycling based on a comprehensive literature review.First,we briefly overview the critical processes in migrating and transforming various carbon components,including particulate organic carbon,dissolved inorganic carbon,and dissolved organic carbon.Riverine carbon cycling models are classified into two types:statistical and process-based.The representative models’simulation methods,applications,advantages,and disadvantages were compared.Based on statistical or machine learning methods,empirical statistical models establish the relationship between the riverine carbon flux and environmental factors.This type of model is simple but has poor extrapolation and universality.Process-based models are based on land surface or hydrological models coupled with river carbon cycling-related biogeochemical processes.This model simulates and predicts variations in different riverine carbon fluxes and is more reliable but complicated.Such models typically focus on different scientific problems,and the representations of riverine carbon cycling-related processes differ among these models.Simulation research on riverine carbon cycling is still in its early stages;however,many shortcomings remain.For example,the representations of terrestrial and aquatic carbon cycling and human activities in existing riverine carbon cycling models are insufficient;thus,they cannot accurately simulate and predict long-term changes in riverine carbon cycling.In the future,it will be necessary to strengthen observations of river carbon cycling processes and improve our understanding of terrestrial and a

关 键 词：陆—水—气交换 碳收支 河流系统 经验统计模型 过程机理模型 

分 类 号：P343.1[天文地球—水文科学]