Subsampling accuracy in commercial beam trawl catches
Van de Walle, S. (2012). Subsampling accuracy in commercial beam trawl catches. MSc Thesis. ILVO Visserij/Marine Biology, Ghent University: Oostende & Ghent. [32] pp.
An ecosystem approach to fisheries management enlarges the focus from single (commercial) species approach to a wider focus on ecosystem functioning. This approach requires data for the entire catch, including non-commercial species, like benthic invertebrates, in addition to the typically targeted fish species. Such data is rarely available since it is highly labour intensive and catches are often too large and difficult to manipulate. Hence, when collecting data on total catch composition, one typically has to perform subsampling in order to obtain (reliable?) estimates. This study focuses on the subsample size necessary to obtain reliable species abundance estimates in flatfish beam trawl fisheries in the southern North Sea. Note that we do not include commercial fish species in this study, since their abundances are usually fully counted and documented. The minimal size for a reasonably representative subsample depends on the area-specific diversity of the catch and its mixture during catch handling. In this study subsamples were manually mixed on board as best as possible. A variety of hauls with different total weights were randomly divided into 10 litre-buckets. Using an index of relative abundance, each species was classified into different categories in each haul: rare, common, abundant and dominant. Other categorisations systems were also used as comparison. This index identified two highly dominant species, namely Asterias rubens and Psammechinus miliaris. A large number of species were categorized as common and abundant (e.g. Liocarcinus holsatus, Pagurus bernhardus, etc.),while most were classified as rare (e.g. Callionymus lyra, Aphrodita aculeata, etc.). Subsequently a permutation approach was used to identify the number of 10 litre-buckets required to reasonably estimate species abundance within different ranges of accuracy. This number of required buckets then translates into an acceptable subsample size. As expected, the larger subsample (i.e. the more buckets combined to make a subsample), the higher the accuracy of the estimates. In addition to this, estimates for abundant species required a lower subsample size than estimations for rare species. This study quantifies these abundance specific relationships between subsample size and accuracy and allows for recommendations for sampling methodology in the case off flatfish beam trawl fishery in the southern North Sea. As further investigation, a similar approach was used for the number of species recorded in a subsample. Catch diversity is an important factor influencing error rates. Studying the relationship between mean error rates and catch size and diversity revealed a cautious trend towards lower error at higher evenness and larger catch sizes.
All data in the Integrated Marine Information System (IMIS) is subject to the VLIZ privacy policy
