Abstract
In this work, a new stock assessment method, Direct Survival Analysis, is proposed and described. The parameter estimation of the Weibull survival model proposed by Ferrandis (2007) is obtained using trawl survey data. This estimation is used to establish a baseline survival function, which is in turn used to estimate the specific survival functions in the different cohorts considered through an adaptation of the separable model of the fishing mortality rates introduced by Pope and Shepherd (1982). It is thus possible to test hypotheses on the evolution of survival during the period studied and to identify trends in recruitment. A link is established between the preceding analysis of trawl survey data and the commercial catch-at-age data that are generally obtained to evaluate the population using analytical models. The estimated baseline survival, with the proposed versions of the stock and catch equations and the adaptation of the Separable Model, may be applied to commercial catch-at-age data. This makes it possible to estimate the survival corresponding to the landing data, the initial size of the cohort and finally, an effective age of first capture, in order to complete the parameter model estimation and consequently the estimation of the whole survival and mortality, along with the reference parameters that are useful for management purposes. Alternatively, this estimation of an effective age of first capture may be obtained by adapting the demographic structure of trawl survey data to that of the commercial fleet through suitable selectivity models of the commercial gears. The complete model provides the evaluation of the stock at any age. The coherence (and hence the mutual “calibration”) between the two kinds of information may be analysed and compared with results obtained by other methods, such as virtual population analysis (VPA), in order to improve the diagnosis of the state of exploitation of the population. The model may be drawn up in a deterministic format, but the main concepts may be interpreted as expectations if stock and catch are considered as stochastic processes.
Highlights
Results produced by trawl surveys have often been criticised on the grounds of their lack of robustness based on their high variability
Some initial steps of this promising method will be presented in future applications recommended by the recently created Permanent Working Group on Stock Assessment Methodology (PWGAM) of the General Fisheries Commission of the Mediterranean (GFCM)
We use the baseline survival already estimated from trawl surveys, with parameters Z0 and α, in order to estimate the survival functions that correspond to the cohorts given by commercial catch-at-age data, considering the same species in the same or neighbouring regions in overlapping periods
Summary
Results produced by trawl surveys have often been criticised on the grounds of their lack of robustness based on their high variability. The Weibull distribution here applied, is a natural extension of the generally accepted exponential model underlying the classical Virtual Population Analysis (Gulland, 1965) and its further modifications (Separable VPA by Pope and Shepherd, 1982, and XSA by Shepherd, 1991, 1999) It produces a natural adaptation of the stock and catch equations that makes it possible to “rewrite” the theory and computational tools of population dynamics in a relatively easy way, understandable to all the fisheries scientist who are familiar with widespread software such as FISAT (Gayanilo and Pauly, 1997), VIT (Lleonart and Salat 1997), VPA Suite (Darby and Flatman, 1994). Some initial steps of this promising method will be presented in future applications recommended by the recently created Permanent Working Group on Stock Assessment Methodology (PWGAM) of the General Fisheries Commission of the Mediterranean (GFCM)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
