Date Awarded


Document Type


Degree Name

Doctor of Philosophy (Ph.D.)


Virginia Institute of Marine Science


Carl Hershner


This research develops a model to identify indicators of potential suitable spawning habitat for river herring, Alosa pseudoharengus and A. aestivalis, using watershed and stream-reach metrics. The results of icthyoplankton samples collected from thirty-four streams feeding into the Rappahannock River below the Embree Dam at Fredericksburg indicate where river herring spawning occurred. Watershed and stream-reach metrics were either measured in the field or derived from digital data in a GIS. Benthic macroinvertebrate analysis was used to compare habitat quality among sites. Streams were classified as either absence or presence of herring eggs or larvae based on the results of the ichthyoplankton samples. Depending on the distributions of the metrics, T-tests or Mann-Whitney U tests determined which metrics were significantly different between the absence and presence groups. Variables from Principal Component Analyses were used in discriminant analyses to examine the relative importance of watershed and stream-reach metrics in predicting suitable spawning habitat. The results of the analyses show that river herring tend to spawn in larger, elongated watersheds with greater mean elevations and habitat complexity. River herring prefer watersheds with greater percentages of deciduous forest and less grassland areas and stream reaches with less organic matter and less fine sediment in the substrate, and more canopy cover and snags. The discriminant analysis using watershed metrics has a better predication ability, 88.2%, than other discriminant models using stream-reach metrics. Except for two metrics, the ratios of %Chironomous to %Chironomidae and %Chironomous to %Chironomini, most of the benthic macroinvertebrate metrics indicate that the presence group has a more degraded environment. This model could be used to target restoration efforts not only in the Rappahannock River but elsewhere in the Chesapeake Bay. Furthermore, the multi-scale model design could be used to target management efforts for other aquatic species.



© The Author