Using capture-mark-recapture techniques to estimate detection probabilities & fidelity of expression for the critically endangered James spinymussel (Pleurobema collina).

Publication Year:
Usage 219
Downloads 149
Abstract Views 70
Repository URL:
Esposito, Alaina C.
JMU Scholarly Commons
Pleurobema collina; James Spinymussel; conservation; modeling; freshwater; mussel
thesis / dissertation description
The critically endangered James Spinymussel (Pleurobema collina) is a species of freshwater mussel endemic to Virginia’s James and Dan River basins. In the last 20 years, P. collina has experienced a substantial decline in numbers and currently occupies approximately 10% of its original habitat; however, little information is known about this species to assist in conservation. A 230-meter reach of transitional habitat in Swift Run was selected for repeat observations to estimate detection probabilities using a Capture-Mark-Recapture framework. In June 2014, visual scouting began to locate and tag P. collina (including other mussels in the community) with PIT tags. Repeat surveys were conducted bi-weekly for three months to relocate tagged individuals, record their position, visibility on the surface, and relevant habitat characteristics. On average, 76% of all tagged bivalves were detected with the PIT tag reader, and 7.5% of those encountered were visible on the surface. The best-fitting logistic regression model to estimate P. collina’s visual detection probability incorporates mussel length, substrate grain size, and sampling season (AIC=94.8), and predicts that 14.0% of all P. collina are visually detectable during baseline flow conditions in the summer. Temporal variation in surface expression rates of P. collina did not vary significantly (p-value=0.90) but did significantly vary for V. constricta (p-value=0.001), indicating that there is heterogeneous expression on the surface for V. constricta. Multistate analyses similarly estimated the PIT tag reader detection of P. collina at 78%-79% and survival at 100%, during any sampling period. Multistate analyses also predict 8% of P. collina will move from a subsurface to a surface state, while 70% will move from a surface to a subsurface state during any sampling interval. Movement analyses of P. collina indicated that individuals occupying sandier habitats do not move more than those in more stable pebble/gravel habitats (p-value=0.61). Raster-analysis of the surrounding watershed show no land use changes from 2001-2006, and that natural vegetation and agriculture dominate the landscape, covering 9.5 km2 and 7 km2, respectively. This information will assist in estimating population sizes, and understand the dynamics between P. collina and their habitat, to aid managers in furthering conservation efforts.