Inferring trends in a small-scale, data-limited tropical fishery based on fishery-independent data

Size-based indicators have emerged as useful tools to analyze the status of fisheries which lack fishery dependent data over long time series, such as many coral reef fisheries. In this study, we calculate a number of size-based indicators for the Haitian fishery at the remote Navassa Island, where a reef fish visual census (RVC) data set is available over an 8-year study period (2002-2009). We also calculate the slope of the size spectrum indicator within a Bayesian framework, which allows for potential biases inherent in the RVC method to be accounted for in credibility intervals around parameter estimates. Results of our analyses suggest that stocks targeted by traps declined from 2002 to 2004, followed by a period of increase from 2006 to 2009. The slope of the size spectrum declined from 2002 to 2004 and remained constant for the remainder of the study period, and this pattern was driven by a decrease in abundance of larger species targeted by hook-and-line. Analysis of the Lmax spectrum also indicated a decrease in the occurrence rates of larger species throughout the study period. Our methods can be applied to fisheries in other areas where limited fishery-independent data and no fishery-dependent data are available. (C) 2011 Elsevier B.V. All rights reserved.