Comparison of fish size spectra obtained from hydroacoustics and gillnets across seven European natural lakes

We conducted a systematic evaluation of the correspondence in fish length data obtained from vertical hydroacoustics and gillnetting across seven European natural lakes differing in abiotic and biotic characteristics. Length data were analyzed as continuous size spectra characterized by their maximum-likelihood estimated exponents b. First, we examined the relationship between size spectra obtained from the two sampling methods. We then examined whether size spectra from the two methods were correlated with lake descriptors separately or in combination. The modeled relationship between the exponents b from the two methods showed that the exponent b from the hydroacoustics was, on average, the same as that from the gillnet sampling in the seven lakes. The exponents b from the hydroacoustics and gillnets, when averaged, were significantly correlated with lake depth, while their differences were significantly correlated with mean air temperature. To conclude, the overall good correspondence between the continuous size distributions obtained by both methods supports the application of vertical hydroacoustics in acquiring size structure of fish communities in lakes, but not in fully replacing the invasive gillnetting. Yet, some specific methodological details require further research. Key words: continuous size distribution, fish length, target tracking, maximum-likelihood estimate, lake depth, air temperature Nous avons realise Revaluation systematique de la correspondance de donnees de longueur de poissons obtenues par hydroacoustique verticale et par peche au filet maillant dans sept lacs naturels europeens presentant differentes caracteristiques abiotiques et biotiques. Les donnees de longueur sont analysees en tant que spectres de tailles continus caracterises par les valeurs estimees de leurs exposants b correspondant au maximum de vraisemblance. Nous examinons d'abord la relation entre les spectres de tailles obtenus des deux methodes d'echantillonnage. Nous evaluons ensuite si les spectres de tailles des deux methodes sont correles a des variables descriptives des lacs prises separement ou en combinaison. La relation modelisee entre les exposants b des deux methodes montre que l'exposant b de la methode hydroacoustique est, en moyenne, le meme que celui de l'echantillonnage au filet maillant pour les sept lacs. Les valeurs moyennes des exposants b des methodes acoustique et au filet maillant sont significativement correlees a la profondeur du lac, alors que leurs differences sont significativement correlees a la temperature moyenne de l'air. En conclusion, la bonne correspondance generale des distributions de tailles continues obtenues par les deux methodes appuie l'utilisation de l'hydroacoustique verticale pour etablir la structure de tailles de communautes de poissons dans les lacs, mais cette methode ne peut remplacer completement la methode invasive au filet maillant. Certains details methodologiques necessitent toutefois plus de recherche. [Traduit par la Redaction] Mots-cles : distribution continue des tailles, longueur de poissons, suivi de cibles, valeur estimee correspondant au maximum de vraisemblance, profondeur du lac, temperature de l'air
1. Introduction The size distribution of aquatic organisms is based on a few fundamental principles. First, large organisms eat small ones, with the maximum size ratio between predator and prey [...]