Deriving blue ling abundance indices from industry haul by haul data
26-08-2009. Lu 2807 fois.
Pascal Lorance, Lionel Pawlowski, and Verena M. Trenkel
ICES CM 2009/L:12
Theme Session: Bringing collaborative science – industry research data into stock assessment and fishery management: evaluating progress and future options
A database of tallybooks, from skippers' own logbooks, provided by the French industry involved in deep-water fishing to the west of the British Isles was used to standardise blue ling Landings per Unit of Effort (LPUEs). The data covered the years 1992-2008 with more extensive data for the period 2000-2007. For each haul, landings by species, tow duration, depth and location were reported. Compared to EU logbooks, this database is on a haul by haul basis instead of being aggregated by fishing sub-trips combining hauls from the same day, ICES rectangle and gear. Moreover, it includes depth, which is a major factor for catch rates in deepwater fisheries. LPUEs were estimated from Generalised Additive Models (GAMs) with depth, vessel, statistical rectangle and zone by year as explanatory variables. Owing to the statistical distribution of landings rates, landings were modelled by a Tweedie distribution, which is a compound Poisson distribution and allows to handle data with many zeros, as it is typical for catch data. In order to investigate how to reliably track stock trends, LPUEs were estimated in five regions for different subsets including or not the spawning season, when blue ling aggregates, or considering tows where blue ling was only a bycatch. The results based on the tallybook data indicated that blue ling LPUEs have been mainly stable over the past decade. This is consistent with stable mean length in the landings. Haul by haul data are suitable to derive abundance indices for deep-water fisheries assessment.