Improving the Assessment of the Lane Snapper (Lutjanus synagris) in a Small-Scale Fishery in Honduras

Abstract

Small-scale fisheries are difficult to assess and manage adequately due to uncertainties surrounding collected data. For example, data are often limited to fishery-dependent data, and fishing gear, such as a gillnet, is very size-selective. This makes the estimation of important parameters such as total mortality and natural mortality difficult, and thus fishery assessment is challenging. Despite these difficulties, management decisions must be made with the available information. The objectives of this thesis were to estimate the total and natural mortality of Lane Snapper (Lutjanus synagris) in the coastal waters of Honduras and determine if the length metric method can be utilized for assessment of this fishery. The total mortality was estimated using a regression catch-curve analysis applied to fishery-dependent catch per unit effort (CPUE) data. The method accounted for gear selectivity by fitting multiple gear-selectivity curves and determining the best curve using a statistical model selection method. The estimated mortality was compared with other estimates obtained without using a regression catch-curve analysis. The natural mortality was estimated using four different methods; these methods included the Peterson and Wroblewski, Lorenzen, and Sekharan estimators as well as using the “Fish Life” R package. Then, fishing mortality was estimated by subtracting natural mortality from the total mortality. Our results showed that the skewed normal selectivity curve fit the data best. The total mortality estimates were higher compared with other estimates when the CPUE was not corrected for gear selectivity. The natural mortality was consistent among three of the different methods employed, with the exception being the Peterson & Wroblewski estimator. When estimating the fishing mortality, more variation came from total mortality compared to natural mortality. This study also demonstrates the importance of using multiple estimators for total and natural mortality to assess uncertainties associated with models and the underlying data. The three indicators proposed by Froese were estimated with and without accounting for size-selectivity of the gill nets, and then the resulting indicators were compared. The estimation of the different indicators suggest that the fishery is experiencing recruitment overfishing and growth overfishing is also occurring. My results suggest that the indicators proposed by Froese can be over- or under-estimated when gear selectivity is not taken into account.