Making sense of the noise: The effect of hydrology on silver carp eDNA detection in the Chicago area waterway system.
Making sense of the noise: The effect of hydrology on silver carp eDNA detection in the Chicago area waterway system.
复制标题
DOI：
10.1016/j.scitotenv.2017.06.255
复制
发表时间：
2017-12
期刊：
影响因子：
--
通讯作者：
Jeffery W. Song;M. Small;Elizabeth A. Casman
中科院分区：
文献类型：
--
作者：
Jeffery W. Song;M. Small;Elizabeth A. Casman
环境DNA （eDNA）采样是监测水生入侵物种传播的新兴工具。在解释eDNA采样证据时的一个混淆因素是，eDNA可能存在于没有活目标生物的水中，来自排泄物、死组织、船只或污水排放等。在芝加哥地区水道系统（CAWS）中，为了防止非本地亚洲鲤鱼入侵密歇根湖，建立了电子鱼类扩散屏障，但自2009年以来，亚洲鲤鱼的eDNA偶尔被检测到高于屏障。本文研究了2009 - 2012年河流流量特征对入侵亚洲鲤鱼eDNA检测概率的影响。在CAWS中，水流方向主要远离密歇根湖，尽管在干旱期间流向密歇根湖的水流方向很少发生逆转。研究发现，入湖逆流量与eDNA检测概率呈显著正相关，而水位高度、降水、季节、水温、溶解氧浓度、pH、叶绿素浓度等协变量与eDNA检测概率无显著正相关。这表明水流方向对CAWS中eDNA检测有很大影响，在解释eDNA证据时应考虑到这一点。我们还发现，与二项回归模型相比，β -二项回归模型对eDNA检测概率的拟合更强。本文提供了一个统计建模框架，用于解释eDNA采样证据和评估影响eDNA检测的协变量。
Environmental DNA (eDNA) sampling is an emerging tool for monitoring the spread of aquatic invasive species. One confounding factor when interpreting eDNA sampling evidence is that eDNA can be present in the water in the absence of living target organisms, originating from excreta, dead tissue, boats, or sewage effluent, etc. In the Chicago Area Waterway System (CAWS), electric fish dispersal barriers were built to prevent non-native Asian carp species from invading Lake Michigan, and yet Asian carp eDNA has been detected above the barriers sporadically since 2009. In this paper the influence of stream flow characteristics in the CAWS on the probability of invasive Asian carp eDNA detection in the CAWS from 2009 to 2012 was examined. In the CAWS, the direction of stream flow is mostly away from Lake Michigan, though there are infrequent reversals in flow direction towards Lake Michigan during dry spells. We find that the flow reversal volume into the Lake has a statistically significant positive relationship with eDNA detection probability, while other covariates, like gage height, precipitation, season, water temperature, dissolved oxygen concentration, pH and chlorophyll concentration do not. This suggests that stream flow direction is highly influential on eDNA detection in the CAWS and should be considered when interpreting eDNA evidence. We also find that the beta-binomial regression model provides a stronger fit for eDNA detection probability compared to a binomial regression model. This paper provides a statistical modeling framework for interpreting eDNA sampling evidence and for evaluating covariates influencing eDNA detection.