Abstracts
Résumé
Les auteurs, disposant d'un ensemble de données quantitatives concernant des populations de poissons issus d'un même contexte biogéographique, l'Ardenne belge, leur appliquent divers traitements mathématiques dans le but de connaître celui(ceux) qui perme(tten)t de préciser l'existence de groupements Ichthyens (hiérarchisation de 2° ordre).
Il en ressort que l'analyse des fréquences relatives des espèces en chaque station par la méthode des composantes principales, met en évidence les structures biotypologiques et leurs relations :
- le classement structurel des eaux torrentielles ne dépend que d'un paramètre lié fortement à la vitesse du courant (en relation avec la pente).
- le classement structurel des eaux modérées à lentes est étroitement lié à un paramètre de limnophilie, complété de 2 paramètres probables demeurant à déterminer.
Les auteurs proposent, pour le rhitron et le potamon supérieur des affluents ardennais de la Meuse et du Rhin belges :
• un essai de typologie basé sur l'existence de groupes socio-écologiques de poissons téléostéens, définis au mieux comme suit :
- Rhéophile supra truite, chabot
médio vairon, loche franche
infra ombre, goujon, chevesne
- Limnophile supra hotu, vandoise, barbeau, gardon
médio tanche (réserves pour échantillonnage trop faible) ;
• une méthodologie de détermination du niveau typologique pour le rhitron et l'épipotamon ;
• les limites d'extension des zones typologiques obtenues.
Mots-clés:
- Ichthyotypologie,
- Ardenne,
- Meuse,
- Rhin,
- composantes principales
Abstract
Quantitative data about fish populations from a homogeneous biogeographical environment, the Belgian Ardenne rivera, have been subjected to statistical analysis. This analysis was designed to determine which statistical method(s) could be used to identify ichthyological groups (second-ordre hierarchy).
Electro-fishing data were available for 140 sites ; elimination of polluted stations left 98 sites, and of these 84 were used for the statistical analysis. Four matrices were constructed frome the raw data : absolute and relative species frequencies ; absolute and relative species biomasses. Correlation coefficients were used together with a hierarchical ranking method, based on chaining and principal component analysis (PCA). Methodological variations involved the selected data basic, selection of distance, species position criteria and chaining method.
Application of PCA to the species frequency data yielded a primary axis (33 % explained variance) which separated suprarheophilic and limnophilic species. The second, third and fourth axes explained 14, 10 and 8 % of the variance, respectively. Similarly, PCA analysis of the biomass data yielded a primary axis (25 % explained variance) separating suprarheophilic and moderate-water species. The second, third and fourth axes explained 13, 10 and 8 % al the variance. Projection of the structure in the 1-2-3 numerical sub-space (57 % of the information) is very expressive.
Hierarchical analysis of the proximity of different species led to the définition of the following socio-ecological groups :
- Rheophilic supra brown trout, bullhead
medio minnow, stone-loach
infra grayling, gudgeon, chub
- Limnophilic supra common nose, dace, barbet, roach
medio tench (with reserves)
Hierarchical analysis of the proximity between stations confirms, in the 2-4 space of the first PC analysis, these same divisions.
In conclusion, the PC analysis of the relative occurrence of the various species at each station reveals the biogeographical structures and their relations. The structural ranking of the torrential waters depends on a single parameter, strongly related to the flow velocity (in relation to the bed slope). The structural ranking of the moderate to slow-moving waters is strongly related to a limnophilic parameter, likely associated with two other unknown parameters (to be determined). A typological level determination is proposed for the rhitron and epipotamon, as are the extension limits for the defined typological zones.
Keywords:
- Ichthyotypologie,
- Ardenne,
- Meuse,
- Rhin,
- composantes principales
Download the article in PDF to read it.
Download