Date of Award


Publication Type

Doctoral Thesis

Degree Name



Great Lakes Institute for Environmental Research

First Advisor

Hugh MacIsaac

Second Advisor

Douglas Haffner


early detection, environmental DNA, invasive species, risk assessment




Management of species invasions ideally requires early detection of species at low abundance, which is often challenging for traditional methods. Environmental DNA (eDNA) provides a promising tool with enhanced sensitivity relative to traditional methods. As an emerging method, however, detection of species at low abundance based on eDNA needs to be optimized to improve detection rate and reduce false negatives. I conducted a meta-analysis, the results of which suggested the significance of using the highly sensitive PCR method and extensive sampling (i.e., replicates sampling and large water volume) to improve detection rate in eDNA-based, low-abundance species detection programs. Needs for improved assay sensitivity screening, testing and reporting were also identified to reduce false negatives and to inform future uses. I developed and optimized an eDNA-based early detection method for the invasive bivalve Limnoperna fortunei (golden mussel) and applied it to investigate the spatial-temporal distribution of golden mussel DNA in the central route of South to North Water Diversion Project (SNWDP) in China. I found that improved detection could be achieved by optimizing sensitivity of the method used either through screening primer pairs or PCR methods. A primer pair with a lower limit of detection (LoD) achieved earlier and lower abundance detection of the target species relative to those with higher LoD. Water samples containing re-suspended matter from the bottom layer were better for detection than those exclusively collected from the surface layer, and only sampling the latter caused false negatives. Quantitative PCR yielded higher detection rates than conventional PCR, while the quantification efficiency was reduced in field water samples as compared to total genomic DNA. Replicate sampling was critical to reduce false negative detections. The majority of positive detections of golden mussel DNA in the main canal of SNWDP were concentrated in warm months, and the occurrence of positive detections was significantly related to minimum daily air temperature, consistent with the expected spawning season of the species. Golden mussel DNA was detected as far as ~1150 km from the putative source of the individuals, indicating long-distance transport of veligers during spawning season. Finally, I tested the functional response and size-selective clearance of the golden mussel to project their potential impacts. Results indicated that golden mussels have a type I functional response, with clearance rate inversely related to food concentration. Presence of golden mussels suppressed suspended matter concentration, the extent of which was dependent on animal abundance, particle size, and their interactions. Golden mussels packaged fine suspended particles into coarser ones, and capture efficiency was inversely related to particle size. Given the suitable habitat and continuous water flow in the main canal of the SNWDP, it seems inevitable that it will be colonized by golden mussels. Abundance mitigation should be considered for the main canal, while containment and dispersal limitation should be prioritized to prevent further spread and reduce overall impact.