Detectability of Estuarine Fishes in a Beach Seine Survey of Tidal Tributaries of Lower Chesapeake Bay

Abstract

Detectability, the probability of encountering a species at a sampling site, is often overlooked in fisheries research despite its potential to obscure inferences on habitat use and lead to biased estimates of abundance. We used occupancy models to explore factors affecting detectability and occupancy (Ψ), the probability that a species inhabits a site, for three fishes frequently captured in Chesapeake Bay seine surveys: young-of-the-year (age-0) striped bass Morone saxatilis, yearling Atlantic croaker Micropogonias undulatus, and spottail shiner Notropis hudsonius. Twelve repeat-sampling events occurred during the summers of 2008 and 2009 at 20 sites in the Mattaponi and Pamunkey rivers, Virginia. Although the environmental factors that influenced detection probabilities varied with species, the detectability of all species was positively related to effective net length (i.e., the maximum distance from shore that the seine was deployed). The mean detectability of age-0 striped bass, which occupied nearly every site (Ψ = 0.99, SE = 0.01), was 0.62 (SE = 0.06) and positively related to mean water temperature during sampling. The detectability of yearling Atlantic croakers was negatively related to water temperature at time of capture and was significantly greater in early summer than in late summer. In early summer, Atlantic croakers occupied all sampling sites; occupancy decreased in late summer (Ψ = 0.86, SE = 0.08) and was positively related to the mean salinity at the site. Unlike with Atlantic croakers, the mean detectability of spottail shiners was significantly greater in late summer than in early summer; detectability was positively related to turbidity at the time of sampling and reflected the increased availability of recently recruited individuals. Spottail shiners occupied fewer sites than did the two other species (Ψ = 0.59, SE = 0.21). Seasonal changes in detectability appeared to reflect changes in catchability associated with fish growth. Determining factors that influence detection probabilities is vital to improving our understanding of habitat use by fish and reducing the variation in fisheries sampling.

Received August 20, 2010; accepted March 24, 2011

ACKNOWLEDGMENTS

Study design and analytical advice were provided by D. MacKenzie (Proteus Research & Consulting, Ltd.) and J. Hines (U.S. Geological Survey [USGS] Patuxent Wildlife Research Laboratory). We also thank D. Hewitt (USGS Western Fisheries Research Center) for discussions concerning occupancy models. We are grateful to the following Virginia Institute of Marine Science (VIMS) students and staff for assistance in the field: K. Capossela, T. Clardy, A. Deary, M. Henderson, B. Jennings, P. Link, L. Machut, C. Magel, P. McGrath, M. Miller, M. Norwood, and F. Ribeiro. We appreciate the comments and suggestions provided on an earlier draft of this manuscript by R. Dickhut, E. Hilton, R. Latour (VIMS), and two anonymous reviewers. This project was funded by the Hall–Bonner Minority Graduate Student Program, and the Virginia Marine Resources Commission. This paper is contribution 3151 of the Virginia Institute of Marine Science.