模拟计算平行样对流域生物信息流估算的影响  

Simulating the impacts of parallel samples on the estimations of upstream⁃to⁃downstream watershed biological information flow

作　　者：杨海乐 张辉[1] 吴金明[1] 李君轶[1] 王成友[1] 杜浩[1] 危起伟[1] YANG Haile;ZHANG Hui;WU Jinming;LI Junyi;WANG Chengyou;DU Hao;WEI Qiwei(Key Laboratory of Freshwater Biodiversity Conservation,Ministry of Agriculture and Rural Affairs of China,Yangtze River Fisheries Research Institute,Chinese Academy of Fishery Sciences,Wuhan 430223,China)

机构地区：[1]中国水产科学研究院长江水产研究所,农业农村部淡水生物多样性保护重点实验室,武汉430223

出　　处：《生态学报》2022年第13期5211-5225,共15页Acta Ecologica Sinica

基　　金：农业农村部财政专项“长江渔业资源与环境调查”(CJDC⁃2017⁃14);中国水产科学研究院中央级公益性科研院所基本科研业务费专项(2020JBF01,2020TD08)。

摘　　要：流域生物信息流是流域生态学研究中的重要内容,是流域生态系统中的物质输移和能量输移过程的信息标记,是用eDNA技术调查评估河流水体中物种组成空间特征的基础。估算流域生物信息流是流域生态系统过程研究和eDNA技术调查评估河流水体中物种组成空间特征的关键。在有限的调查采样中,平行样的数量如何影响流域生物信息流的估算,尚待解答。基于随机抽样调查的基本原理,提出假设——采样数量不影响流域生物信息流估算结果的准确度,但会影响其精密度,然后通过问题简化转化和模拟计算,对该假设进行了检验。模拟计算结果显示,随着样点生物信息检出度(平行样数量)的增大,流域生物信息流估算结果会从偏小逐渐靠近流域生物信息流实际值,同时其99.9%置信区间也逐渐集中于流域生物信息流实际值。即样点生物信息检出度(平行样数量)对流域生物信息流估算的准确度和精密度均有影响。在实际调查研究过程中,建议先在所研究区域对平行样数量和样点生物信息检出度的关系进行预评估,然后基于流域生物信息流估算可信度目标在正式实施方案中经济有效地设置平行样,基于多平行样调查结果估算流域生物信息流,再根据各样点生物信息检出状况对流域生物信息流估算结果进行后验评估。Watershed biological information flow(WBIF)is defined as the path,processes and control of biological information transport,exchange,interaction and feedback among different spaces and systems along with watershed ecosystem processes,and could be partly described as the land⁃to⁃river and upstream⁃to⁃downstream bioinformation transportation(including organisms,nucleic acids,peptides and other biomarkers),which is driven by the hydrologic processes of watershed systems.The WBIF labels the transport of organic matter and energy.The WBIF integrates the ecological processes of environmental DNA(eDNA),including the origin,state,transport,and fate of eDNA,and makes it possible that the species composition in river system is monitored and assessed using eDNA.The WBIF estimation is the key for watershed ecosystem processes studying and riverine biodiversity monitoring.However,in practice,the parallel samples in each sampling site always are limited.And how parallel samples would impact WBIF estimation is unknown.Based on the principles of stochastic sampling survey,we hypothesized that parallel samples would not impact the accuracy of the WBIF estimation,but affect the precision of the WBIF estimation.Then,we transformed this hypothesis into a set of formulas and tested it with a series of analog computation.Results showed that the number of parallel samples(efficiency of detection)affected both the accuracy and precision of the WBIF estimation.The optimal WBIF estimation was less than the actual WBIF in any condition.Along with the increase of parallel samples(efficiency of detection),the optimal WBIF estimation gradually neared to the actual WBIF,the range of WBIF estimation gradually focused on the actual WBIF.In other words,more parallel samples(higher efficiency of detection)led higher accuracy and precision of the WBIF estimation.In addition,the actual WBIF affected both the accuracy and precision of the WBIF estimation too.Larger actual WBIF led higher accuracy and precision of the WBIF estimation.The relat

关 键 词：流域生物信息流 环境DNA 平行样 检出度 流域生态学 

分 类 号：Q811.4[生物学—生物工程]