Page View
Barry, Terence P.; Malison, Jeffrey A. (ed.) / Proceedings of PERCIS III, the Third International Percid Fish Symposium, University of Wisconsin, Madison, Wisconsin, U.S.A., July 20-24, 2003
(2004)
Prchalová, Marie; Kubečka, Jan
(Management) Are percid fish overestimated by gillnet sampling?, pp. 121-122
PDF (1.1 MB)
Page 121
ARE PERCID FISH OVERESTIMATED BY GILLNET SAMPLING? Marie Prchalova' and Jan Kubecka, Faculty of Biological Sciences, University of South Bohemia, Branigovskd 31, 370 05 Ceske Budejovice, Czech Republic and Hydrobiological Institute, Czech Academy of Sciences, Na Sadkach 7, 370 05 Cesk6 Budejovice, Czech Republic, marie.prchalovaCa)bf.Icu.cz Introduction. Common percid fish - perch (Perca fluviatilis), pikeperch (Stizostedion lucioperca) and ruffe (Gymnocephalus cernuus) - possess a structured body surface and have a relatively firm body structure (Kipling, 1963). This, together with a potential difference in activity, may result in higher probability of capture and retention in gillnets comparing with smooth bodied fish like cyprinids or salmonids. Due to this, percid fish can be overestimated by gillnets. To verify potential bias, we compared fish species composition based on catches with gillnets and a different fishing gear. As a reference gear, shore seines at the same time and site was used. Methods. Suitable data for observing potential overestimation of percid fish were taken from three Dutch water-supply reservoirs (De Gijster, Honderd en Dertig and Petrusplaat) and two Czech reservoirs (Rimov and Staviste) during 1998 - 2002. Nordic multimesh benthic gillnets were used for the studies. Sets included 13 different mesh sizes - 6, 8, 10, 12.5, 16, 19.5, 24, 29, 35, 43, 55, 70 and 90 mm (knot to knot). Each panel was 1.5 m high and 25 m long. Benthic gillnets were set for 12 hours overnight in the littoral in the depth 2 - 3 m. Hauls with 50 m shore seines with mesh size 10 mm (knot to knot) were carried out during the same night and on the place as close to the gillnets as possible to not disturb fish. Absolute numbers of caught fish from both gears were converted to a ratio of number of percids to the sum of numbers of percids and cyprinids. Cyprinids created major part of abundance in selected reservoirs and they represent a close group considering the morphology. This ratio was tested with multifactor analysis of variance (MANOVA, p-level of 0.05) after arcsin transformation. An overestimation index of percids from gillnets was determined as a ratio of the proportion in gillnets / proportion in seine. Values above I mean overestimation and values below I signify underestimation of percids by gillnets. Results. Percids (perch, pikeperch, ruffe) and cyprinid fishes (mainly roach Rutilus rutilus, common bream Abramis brama and bleak Alburnus alburnus) comprised 90 - 100% of both gillnet and seine catches in all reservoirs (Fig. 1). The rest of catches were represented by a sporadic catch of pike (Esox lucius), trout (Salmo trutta), eel (Anguilla anguilla) and smelt (Osmerus eperlanus, in Dutch reservoirs only). We compared 26 couples of ratios of the number of percids to the sum of percids and cyprinids - 10 from Dutch reservoirs, 15 from the Rkimov reservoir and one from the Staviste reservoir. A higher proportion of percid fish in gillnets was found in 77% of cases. Statistical analysis showed significant differences between proportions of percids 100 -I 80 - 60- 40 . 20 Dutch Dutch Czech Czech gillnets seines gillnets seines * others m bream o roach o ruffe E perch from two compared fishing gears (p-level 0.0 142). Figure 1. Average species composition from gillnets and seines in Dutch and Czech reservoirs. In individual catches, the overestimation index varied between 0.7 and 16. Table I shows the weighted average of the overestimation index for total number of percids and for particular perciform species. With only one exception - ruffe in Dutch reservoirs perciform species were overestimated by gillnets. The highest overestimations were gained for ruffe and perch in Czech reservoirs and for pikeperch from Dutch reservoirs. Overestimation index of percids was found to be adversely related to the proportion of percids in the community from the reference catch (Fig 2). In high percid densities, the overestimation index cannot be high by definition. This may partly explain higher values of overestimation indexes in Czech reservoirs (Table 1). After log-transformation, the relationship in Figure 2 can be described by linear function (Log y= -0.74 Log x + 1.33; R2 0.75) and has a hyperbolic course similar for all three perciform species. Discussion. The catchability of fish by gillnets is affected by factors, which are related either to the characteristics of the net or of the fish (Reis & Pawson, 1999). Many works were oriented to the significance of nets parameters (Kurkilahti, 1999), but less attention was paid to the fish attributes. 121
Copyright January 2004 University of Wisconsin Board of Regents.| For information on re-use see: http://digital.library.wisc.edu/1711.dl/Copyright




